sign in sign up
<<
Home , Aminohydrolase, Etirinotecan pegol, IFL (chemotherapy)

Chapter 20 Topoisomerase Interactive Agents 221

9 7 A 10 O O B – Indenoisoquinolines AB N OH N OH N 10 9 O 1 E + E O– O 11 20 O H 20 O 1 H CO 2 8 O O 3 7 HO HO (CH ) NH (CH ) OH NSC 706744 3 5 2 3 2 2 (lactone) Camptothecin (carboxylate) N MJ-III-65 ACTIVE INACTIVE H3CO 4 6 R O NSC 725776 R = (CH2)3 N N = LMP-776 N CH3 H2C = Indimitecan CH3 10 7 bis-piperidine N O H C CH AB 3 C 3 NSC 724998 H C H3C CH3 (CH2)3 N O 3 O N N O = LMP-400 N CH = Indotecan H C CH3 2 N Dibenzonaphthyridinones 2 H2C HO 9 Carboxylesterase 7 7 10 AB (CPT-11) N 2 O CH B B N Prodrug H C 3 H CO Genz-644282 2 N N 3 9 O HO 7 3 = SAR402674 Gimatecan 10 AB N SN-38 (CKD-602) H CO 8 5 3 (CH2)2–NHCH3 Active metabolite N O

H R2 H O C O OH O D E OH O HN N F O H C 9 OH 3 R O O N 14 O NH OH HN N O O HO O OCH3 O OH HO OH O O OH O HN N O CH3 ICRF-187 (Dexrazoxane) H OH O H2N O MeVa MeVa = Adriamycin: R = H O O Sar Sar : R = OH 3' O Pro Pro O O H3CO 4' OCH3 Val Val

D D CANCER THERAPEUTICS R OH 1 Thr Thr (4'-epidoxorubicin) OH R1 R2 CO CO ( -enantiomer of N NH3 doxorubicin O OH CH3

H3C O Etoposide OPO H CH O O phosphate 3 2 3 HO NH CH3 CH3 2 OH S Actinomycin D ()

Figure 20.3 Structure of topoisomerase inhibitors. (A) Camptothecin derivatives are instable at physiologic pH with the formation of a carboxylate derivative within minutes. Irinotecan is a prodrug and needs to be converted to SN-38 to trap Top1cc. (B) Non-camptothecin derivatives in clinical trials. (C) derivatives. (D) Demethyl epipodophyllotoxin derivatives. (E) Other intercalating Top2 inhibitors acting by trapping Top2cc. (F) Structure of dexrazoxane, which acts as a catalytic inhibitor of Top2.

Topotecan There are no formal guidelines for dose reductions in patients with hepatic dysfunction (defined as serum bilirubin> 1.5 mg/dL to Topotecan contains a basic side chain at position C-9 that <10 mg/dL). Topotecan additionally penetrates the blood-brain enhances its water solubility (see Fig. 20.3). Topotecan is approved barrier, achieving concentrations in cerebrospinal fluid that are for the treatment of ovarian cancer,45 small-cell lung cancer,46 approximately 30% that of plasma levels.48 and as a single agent and in combination with for cervi- cal cancer.47 Additionally, it is active in acute myeloid leukemia (AML) and myelodysplastic syndrome (see Table 20.2). Topotecan Camptothecin Conjugates and Analogs is administered intravenously as a single agent at a dose of 1.5 mg/ m2 as a 30-minute infusion daily for 5 days, followed by a 2-week New formulations of camptothecin conjugates and analogs are period of rest for the treatment of solid tumors or at a dose of 0.75 currently in clinical development in an effort to improve the mg/m2 as a 30-minute infusion daily for 3 days in combination therapeutic index (Table 20.3). The development of camptothecin with cisplatin on day 1, every 3 weeks, for the treatment of cervical conjugates is based on the notion that the addition of a bulky conju- cancer. gate would allow for a more consistent delivery system and extend Myelosuppression is the most common dose-limiting toxicity. the half-life of the molecule. Extensive prior radiation or previous bone marrow–suppressive CRLX101, formerly IT-101, a covalent cyclodextrin- increases the risk of topotecan-induced myelo- copolymer camptothecin conjugate, has suppression. Other toxicities include nausea, vomiting, diarrhea, plasma concentrations and area under the curve (AUC) that are fatigue, alopecia, and transient hepatic transaminitis. approximately 100-fold higher than camptothecin, with a half-life Topotecan and its metabolites are primarily cleared by the kid- in the range of 17 to 20 hours compared to 1.3 hours for camp- neys, requiring dose reduction in patients with renal dysfunction. tothecin.49 It has demonstrated antitumor activity in preclinical A 50% dose reduction is recommended for patients with moderate studies in irinotecan-resistant tumors with complete tumor regres- renal impairment (creatinine clearance 20 to 39 mL per minute). sion in human non–small-cell lung cancer, Ewing sarcoma, and 222 Cancer Therapeutics

TABLE 20.2 U.S. Food and Drug Administration–Approved Camptothecin Analogs

Irinotecan FDA approved for: (Camptosar) Metastatic colorectal cancer First-line therapy in combination with Diarrhea (dose reductions are 5-FU/LV recommended for patients who are homozygous for the UGT1A1*28 allele) Second-line therapy as a single agent Myelosuppression Category 2Aa recommendations: Pancreatic cancer First-line therapy in combination with , 5-FU/LV Extensive-stage small-cell lung First-line therapy in combination with cancer cisplatin or Category 2Bb recommendations: Esophageal and GEJ cancers, gastric cancer, cervical cancer, anaplastic gliomas and glioblastomas, non–small-cell lung cancer, ovarian cancer Topotecan FDA approved for: (Hycamtin) Cervical cancer Stage IVB, recurrent, or persistent Myelosuppression carcinoma of the cervix not amenable to curative treatment with surgery and/or radiation therapy Ovarian cancer After failure of initial therapy Small-cell lung cancer After failure of initial therapy Class 2B recommendations: AML, MDS a Category 2A: Recommendations are based upon lower-level evidence, there is uniform National Comprehensive Cancer Network consensus that the intervention is appropriate. b Category 2B: Recommendations are based upon lower-level evidence, there is National Comprehensive Cancer Network consensus that the intervention is appropriate. MDS, myelodysplastic syndrome. lymphoma xenograft models.50 Preliminary data from Phase 1 on the 21-day administration schedule were dehydration and di- studies indicate that CRLX101 is well tolerated at a dose of 15 mg/ arrhea. is currently being evaluated in several m2 administered in a biweekly administration schedule.51 It is cur- phase 2 studies in lung cancer, colorectal cancer, and high-grade rently being studied in Phase 2 studies as a single agent and in gliomas,57–60 with evidence of clinical activity in refractory solid combination with chemotherapeutic agents in lung, renal cell tumors. A Phase 3 trial (The BEACON Study) is underway evalu- cancer, and gynecologic malignancies.52–54 ating NKTR-102 against the physicians’ choice in refractory breast Etirinotecan pegol (NKTR-102), an irinotecan polymer conju- cancer.61 gate, has a longer plasma circulation time with a lower maximum As an alternative to macromolecular conjugates, attempts have concentration of SN-38 compared with irinotecan. It was evalu- also been made to alter the camptothecin pentacyclic ring struc- ated in a Phase 2 study in platinum-resistant refractory epithelial ture with modifications of the A and B ring (see Fig. 20.3A) in ovarian cancer at a dose of 145 mg/m2 administered on a schedule an effort to improve solubility and enhance antitumor activity. of every 21 days; a median progression-free survival of 5.3 months Structure–activity relationship studies have shown that substitu- and median overall survival of 11.7 months was observed.55 Two tions at the 7, 9, and 10 positions serve to enhance the antitu- schedules of administration, 145 mg/m2 administered every 14 days mor activity of camptothecin.62 Belotecan, a novel camptothecin versus every 21 days, have been tested in a Phase 2 study of NKTR- analog, has a water-solubilizing group at the 7 position of the B 102 in patients with previously treated metastatic .56 ring of camptothecin (see Fig. 20.3A). Several Phase 2 studies have Of the 70 patients evaluated in this study, 20 patients achieved evaluated belotecan in combination with carboplatin in recurrent an objective response (29%; 95% confidence interval [CI] 18.4 ovarian cancer63 and in combination with cisplatin in extensive- to 40.6). For both these studies, the most common adverse events stage small-cell lung cancer,64 demonstrating activity in these can- cers; however, these combinations were associated with prominent hematologic toxicities. Phase 2 studies evaluating belotecan as a TABLE 20.3 single agent in patients with recurrent or progressive carcinoma of the uterine cervix failed to show activity.65 Gimatecan is a Topoisomerase I Inhibitors in Development lipophilic oral camptothecin analog (see Fig. 20.3A). Pharmaco- kinetic studies demonstrate that gimatecan is primarily present in Camptothecin Camptothecin Noncamptothecin plasma as the lactone form (>85%), and has a long half-life of Conjugates Analogs Agents 77.1 +/− 29.6 hours, with an increase in maximum concentra- 66 CRLX101 Belotecan Indenoisoquinoline tion (Cmax) and AUC of three- to six-fold after multiple dosing. NKTR-102 Gimatecan Indotecan (LMP-400) Phase 2 studies show that gimatecan has demonstrated activity in previously treated ovarian cancer, with myelosuppression as the MM-398 Homocamptothecin Indimitecan (LMP-776) 67 Elomotecan Dibenzo naphthyridine main toxicity. Diflomotecan Genz-644282 Newer development of analogs have attempted to modify the E-ring through introduction of an electron-withdrawing group at

tahir99 - UnitedVRG Chapter 20 Topoisomerase Interactive Agents 223 the α position in an effort to overcome the instability of the E- advanced neuroblastoma, advanced non–small-cell lung cancer, ring while maintaining the binding capability of the camptothecin advanced ovarian cancer, advanced transitional cell bladder can- analog to the Top1-DNA cleavage complex. Collectively called cer, cervical cancer, and Langerhans cell tumors. Doxorubicin homocamptothecin analogs, two have been tested in clinical trials has activity in other malignancies as well, including soft tissue and include diflomotecan68 and elomotecan.69 The dose-limiting sarcoma, osteosarcoma, carcinoid, and liver cancer (Table 20.4). toxicity in the Phase I study of elomotecan was neutropenia. A Doxorubicin is typically administered at a recommended dose of 2 five-member E-ring derivative has also been developed and has 30 to 75 mg/m every 3 weeks intravenously. reached a Phase 1 clinical trial.70,71 Major acute toxicities of doxorubicin include myelosuppres- sion, mucositis, alopecia, nausea, and vomiting. Myelosuppression is the acute dose-limiting toxicity. Other toxicities, including di- Noncamptothecin Topoisomerase I Inhibitors arrhea, nausea, vomiting, mucositis, and alopecia, are dose and schedule related. Prophylactic antiemetics are routinely given Noncamptothecin Top1 inhibitors are in clinical development, with bolus doses of doxorubicin, and longer infusions are associ- and include indenoisoquinolines and dibenzonaphthyridines (see ated with less nausea and less cardiotoxicity. Patients should also be Fig. 20.3B). Two indenoisoquinoline derivatives are currently warned to expect their urine to redden after drug administration. in clinical development, indotecan (LMP400) and indimitecan 72,73 Doxorubicin is a potent vesicant, and extravasation can lead to se- (LMP776). Early in vitro studies show enhanced potency vere necrosis of skin and local tissues, requiring surgical debride- compared with , and persistence of Top1 cleavage 74 ment and skin grafts. Infusions via a central venous catheter are complexes. Genz-644282, a dibenzonaphthyridine derivative, recommended. Other toxicities of doxorubicin include radiation demonstrated enhanced antitumor activity in preclinical studies75 76 recall and the risk of developing secondary leukemia. Radiation and is currently being evaluated in Phase 1 clinical trials. recall is an inflammatory reaction at sites of previous radiation and can lead to pericarditis, pleural effusion, and skin rash. Secondary TOPOISOMERASE II INHIBITORS: leukemias are thought to be a result of balanced translocations that INTERCALATORS AND result from Top2 poisoning by the , albeit to lesser degree than other Top2 poisons, such as the epipodophyllotoxins NONINTERCALATORS (see the following).82 Anthracyclines are cleared mainly by metabolism to less active Topoisomerase II inhibitors can be classified in two main classes: forms and by biliary . Less than 10% of the administered DNA intercalators, which encompass different chemical classes dose is cleared by the kidneys. Dose reductions should be made (Fig. 20.3C, E), and nonintercalators represented by the epipodo- in patients with elevated plasma bilirubin. Doxorubicin should be phyllotoxin derivatives (see Fig. 20.3D). Although both act by dose reduced by 50% for plasma bilirubin concentrations ranging trapping Top2 cleavage complexes (Top2cc), DNA intercalators from 1.2 to 3.0 mg/dL, by 75% for values of 3.1 to 5.0 mg/dL, and CANCER THERAPEUTICS exhibit a second effect as drug concentrations increase above low withheld for values greater than 5 mg/dL. micromolar values: they block the formation of Top2cc by inter- calating into DNA and destabilizing the binding of Top2 to DNA. This explains why Top2α and Top2β are trapped over a relatively Liposomal Doxorubicin narrow concentration range by anthracyclines, and why intercala- tors have additional effects besides trapping Top2cc, namely in- Doxorubicin is also available in a polyethylene glycol (PEG)ylated hibition of a broad range of DNA processing including liposomal form, which allows for enhancement of drug delivery. helicases, polymerase, and even nucleosome destabilization. Use of liposomal doxorubicin has been associated with less cardio- toxicity even at doses exceeding 500 mg/m2.83 Additionally, liposo- Doxorubicin mal doxorubicin produces less nausea and vomiting and relatively mild myelosuppression compared to doxorubicin. Unique to the liposomal formulation is the risk of hand–foot syndrome and an Doxorubicin and daunorubicin were the first anthracyclines dis- acute infusion reaction manifested by flushing, dyspnea, edema, covered in the 1960s and remain among the most widely used an- fever, chills, rash, bronchospasm, and hypertension. These infu- ticancer agents over a broad spectrum of malignancies. Although sion reactions are related to the rate of infusion; therefore, the doxorubicin only differs by one hydroxyl substitution on position recommended administration schedule is set at an initial rate of 14 (see Fig. 20.3C), doxorubicin has a much broader anticancer 1 mg per minute for the first 10 to 15 minutes. The rate may be activity than daunorubicin. Anthracyclines are natural products slowly increased to complete infusion over 60 minutes if no reac- derived from Streptomyces peucetius variation caesius. They were tion occurs. Typical dosing schedules include 50 mg/m2 intrave- found to target Top2 well after their clinical approval.77 Subse- nous infusion every 4 weeks for four courses in ovarian cancer, quent searches for less toxic drugs and formulations led to the ap- 20 mg/m2 intravenous infusion every 3 weeks in AIDS-related Ka- proval of liposomal doxorubicin, , and epirubicin. posi sarcoma, and 30 mg/m2 intravenous infusion in combination Anthracyclines are flat, planar molecules that are relatively with to be given on days 1, 4, 8, and 11 every 3 weeks hydrophobic. The quinone structure of anthracyclines (see in multiple myeloma. Fig. 20.3C) enhances the catalysis of oxidation-reduction reac- tions, thereby promoting the generation of oxygen free radicals, which may be involved in antitumor effects as well as the cardio- Daunorubicin toxicity associated with these drugs.78,79 Anthracyclines are also substrates for P-glycoprotein and Mrp-1, and drug efflux is thought Despite its chemical similarity (see Fig. 20.3C), daunorubicin is to be a major drug resistance determinant.80,81 considerably less active in solid tumors compared to doxorubicin. Doxorubicin is available in a standard salt form and as a li- It is FDA approved for the treatment of ALL and AML. Dauno- posomal formulation. FDA-labeled indications for standard rubicin is typically administered via intravenous push over 3 to 5 doxorubicin include acute lymphocytic leukemia (ALL), AML, minutes at a dose of 30 to 45 mg/m2 per day on 3 consecutive days chronic lymphoid leukemia, Hodgkin lymphoma, non-Hodgkin in combination chemotherapy. For induction therapy for pediatric lymphoma, mantle cell lymphoma, multiple myeloma, myco- acute lymphoblastic leukemia, daunorubicin is dosed at 25 mg/m2 sis fungoides, Kaposi sarcoma, breast cancer (adjuvant therapy intravenously in combination with and prednisone. In and advanced), advanced prostate cancer, advanced gastric can- children less than 2 years of age or in those who have a body sur- cer, Ewing sarcoma, thyroid cancer, advanced nephroblastoma, face area less than 0.5 m2, current recommendations are based on 224 Cancer Therapeutics

TABLE 20.4 U.S. Food And Drug Administration–Approved Topoisomerase II Inhibitors in Clinical Use

Compound Tumor Type Clinical Indication Major Toxicities I. Anthracyclines Doxorubicin (Adriamycin) Breast carcinoma Adjuvant setting with axillary Dose-dependent cardiotoxicity LN involvement following Myelosuppression resection of primary breast cancer ALL In combination with other AML cytotoxic agents Wilms’ tumor Neuroblastoma Sarcomas Ovarian cancer Transitional cell bladder cancer Thyroid cancer Gastric cancer Hodgkin lymphoma Non-Hodgkin lymphoma Pegylated liposomal Ovarian cancer After failure of platinum-based Myelosuppression doxorubicin (Doxil) chemotherapy Stomatitis AIDS-related Kaposi sarcoma After failure of prior systemic Hand-foot syndrome chemotherapy Dosage reduction recommended with hepatic dysfunction Multiple myeloma In combination with bortezomib Daunorubicin (Cerubidine) ALL Induction therapy Dose-dependent cardiotoxicity AML Myelosuppression Epirubicin (Ellence) Breast cancer Adjuvant therapy in patients Dose-dependent cardiotoxicity with evidence of axillary node Myelosuppression tumor involvement following primary resection Idarubicin (Idamycin) AML Induction therapy Dose-dependent cardiotoxicity Myelosuppression II. Anthracenediones Mitoxantrone (Novantrone) Prostate cancer Hormone-refractory prostate Myelosuppression AML cancer Dactinomycin (Cosmegen) Wilms’ tumor Myelosuppression Rhabdomyosarcoma Ewing sarcoma Nonseminomatous testicular cancer Gestational trophoblastic neoplasia III. Epipodophyllotoxins Etoposide (VePesid) Small-cell lung cancer First-line in combination Myelosuppression Testicular cancer First-line in combination Teniposide (Vumon) Pediatric lymphoblastic leukemia Refractory setting Myelosuppression

LN, lymph node. body mass index (1 mg/kg) rather than body surface area. A higher Epirubicin dose of daunorubicin at 60 mg/m2 per day to 90 mg/m2 per day intravenously for 3 consecutive days is currently recommended as Epirubicin is an epimer of doxorubicin (see Fig. 20.3C) with in- part of the induction combination regimen for the treatment of creased lipophilicity. It is FDA approved for adjuvant therapy of breast acute myeloblastic leukemia. Daunorubicin has similar toxicities to cancer but is also used in combination for the treatment of a variety of doxorubicin, including myelosuppression, cardiac toxicity, nausea, malignancies. Epirubicin is administered intravenously at doses rang- vomiting, alopecia, and is also a vesicant. Daunorubicin is metab- ing from 60 to 120 mg/m2 every 3 to 4 weeks. Epirubicin has a similar olized by the liver and undergoes substantial elimination by the toxicity profile to doxorubicin but is overall better tolerated. kidneys, requiring dose reductions for both renal and hepatic dys- In addition to being converted to an enol by an aldose reduc- function. A 50% dose reduction is recommended for either serum tase, epirubicin has a unique steric orientation of the C-4 hydroxyl creatinine or bilirubin greater than 3 mg/dL, and a 25% reduction group that allows it to serve as a substrate for conjugation reactions in dose for bilirubin concentrations ranging from 1.2 to 3.0 mg/dL. mediated by liver glucuronosyltransferases and sulfatases. As such,

tahir99 - UnitedVRG Chapter 20 Topoisomerase Interactive Agents 225 dose adjustments are recommended in the setting of hepatic dys- less than 550 mg/m2 are considered safe. Additionally, liposomal function. For patients with serum bilirubin of 1.2 to 3 mg/dL or doxorubicin is associated with less cardiac toxicity. aspartate aminotransferase of 2 to 4 times the upper limit of normal, Cardiac function can be monitored during treatment with an- a 50% dose reduction is recommended. For patients with bilirubin thracyclines by electrocardiography, echocardiography, or radio- greater than 3 mg/dL or aspartate aminotransferase greater than 4 nuclide scans. Numerous studies have established the danger of times the upper limit of normal, a dose reduction of 75% is recom- embarking on anthracycline therapy in patients with underlying mended. Due to limited data, no specific dose recommendations cardiac disease (e.g., a baseline left ventricular ejection fraction of are currently available for patients with renal impairment, although less than 50%) and of continuing therapy after a documented de- current recommendations are for consideration of dose adjustments crease in the ejection fraction by more than 10% (if this decrease in patients with serum creatinine greater than 5 mg/dL. falls below the lower limit of normal). Because anthracycline- induced cardiotoxicity has been related to the generation of free Idarubicin radicals, efforts have been aimed at attenuating this effect through the targeting of redox response and reduction in oxidative stress. Dexrazoxane is a metal chelator that decreases the myocardial tox- Idarubicin is a synthetic derivative of daunorubicin, but lacks the icity of doxorubicin in breast cancer patients. In two multicenter, 4-methoxy group (see Fig. 20.3C). It is FDA approved as part of double-blind studies, advanced breast cancer patients were ran- combination for AML and is also active in domized to chemotherapy with dexrazoxane or a placebo; dexra- ALL. It is given intravenously at a dose of 12 mg/m2 for 3 consecu- zoxane was shown to have a cardioprotective effect based on serial, tive days, typically in combination with . Idarubicin has noninvasive cardiac testing during the course of the trial and is similar toxicities as daunorubicin. Its primary active metabolite is approved for that use by the FDA.90 Dexrazoxane chelates iron and idarubicinol, and elimination is mainly through the biliary system copper, thereby interfering with the redox reactions that generate and, to a lesser extent, through renal excretion. A 50% dose reduc- free radicals and damage myocardial lipids. Notably, dexrazoxane tion is recommended for serum bilirubin of 2.6 to 5 mg/dL and is also a Top2 catalytic inhibitor (see Fig. 20.3F), which potentially idarubicin should not be given if the bilirubin is greater than 5 mg/ might minimize the therapeutic activity of anthracyclines by inter- dL. Additionally, dose reductions in renal impairment are advised, fering with the trapping of Top2 cleavage complexes by anthracy- but specific guidelines are not available. clines.2,3,91 Other agents currently in use include β-blockers and statins. A recent meta-analysis of 12 randomized controlled trials Cardiac Toxicity of Anthracyclines and 2 observational studies involving the use of agents to prevent the cardiotoxicity associated with anthracyclines demonstrated Anthracyclines are responsible for cardiac toxicities, and special relatively similar efficacy regardless of which prophylactic treat- considerations are necessary to minimize this severe side effect. ment was used.92

Acute doxorubicin cardiotoxicity is reversible, and clinical signs CANCER THERAPEUTICS include tachycardia, hypotension, electrocardiogram changes, and Anthracenediones arrhythmias. It develops during or within days of anthracycline in- fusion, and its incidence can be significantly reduced by slowing Mitoxantrone (see Fig. 20.3E) is currently the only clinically ap- doxorubicin infusion rates. proved anthracenedione. Compared to anthracyclines, mitoxan- Chronic and delayed cardiotoxicity is more common and more trone is less cardiotoxic owing to a decreased ability to undergo severe because it is irreversible. Chronic cardiotoxicity with con- oxidation-reduction reactions and form free radicals. gestive heart failure peaks at 1 to 3 months but can occur even Mitoxantrone is FDA approved for the treatment of advanced years after therapy. Myocardial damage has been shown to occur hormone-refractory prostate cancer93 and AML.94 It is typically ad- by several mechanisms. The classical mechanism is by the direct ministered intravenously at a dose of 12 to 14 mg/m2 every 3 weeks generation of reactive oxygen species (ROS) during the electron in the treatment of prostate cancer, and at a dose of 12 mg/m2 in transfer from the semiquinone to quinone moieties of the anthra- combination with cytosine arabinoside for 3 days in the treatment cycline,84 which leads to myocardial damage. ROS can also be of AML. generated by mitochondrial damage resulting from drug-mediated Toxicities are generally less severe compared to doxorubi- inactivation of the oxidative phosphorylation chain because doxo- cin and include myelosuppression, nausea, vomiting, alopecia, rubicin accumulates not only in chromatin, but also in mitochon- and mucositis. Cardiac toxicity can be seen at cumulative doses dria.78,79 A recent study has also related doxorubicin cardiotoxicity greater than 160 mg/m2.95 Mitoxantrone is rapidly cleared from to the poisoning of Top2β cleavage complexes in myocardiocytes.85 the plasma and is highly concentrated in tissues. The majority of Endomyocardial biopsy is characterized by a predominant finding the drug is eliminated in the feces, with a small amount undergo- of multifocal areas of patchy and interstitial fibrosis (stellate scars) ing renal excretion. Dose adjustments for hepatic dysfunction are and occasional vacuolated myocardial cells (Adria cells). Myocyte recommended, but formal guidelines are currently not available. hypertrophy and degeneration, loss of cross-striations, and the absence of myocarditis are also characteristic of this diagnosis.86 The incidence of cardiomyopathy is related to both the cumula- Dactinomycin tive dose and the schedule of administration, and predisposition to cardiac damage includes a previous history of heart disease, hyper- Dactinomycin was the first antibiotic shown to have antitumor ac- tension, radiation to the mediastinum, age greater than 65 years or tivity96 and consists of a planar phenoxazone ring attached to two younger than 4 years, prior use of anthracyclines or other cardiac peptide side chains. This unique structure allows for tight intercala- toxins, and coadministration of other chemotherapy agents (e.g., tion into DNA between adjacent guanine–cytosine bases, leading , , or trastuzumab).87,88 Sequential to Top2 and Top1 poisoning and transcription inhibition.97 Dacti- administration of paclitaxel followed by doxorubicin in breast can- nomycin was one of the first drugs shown to be transported by P- cer patients is associated with cardiomyopathy at total doxorubicin glycoprotein, and represents the major mechanism of resistance.98 doses above 340 to 380 mg/m2, whereas the reverse sequence of Dactinomycin is FDA approved for Ewing sarcoma,99 gesta- drug administration did not yield the same systemic toxicities at tional trophoblastic neoplasm,100 metastatic nonseminomatous tes- these doses.89 When doxorubicin is given in a low-dose weekly regi- ticular cancer,101 nephroblastoma,102 and rhabdomyosarcoma.103 men (10 to 20 mg/m2 per week) or by slow continuous infusion Typically, it is administered intravenously at doses of 15 μg/kg for over 96 hours, cumulative doses of more than 500 mg/m2 can be 5 days in combination with other chemotherapeutic agents for given. Doses of epirubicin less than 1,000 mg/m2 and daunorubicin the treatment of nephroblastoma, rhabdomyosarcoma, and Ewing 226 Cancer Therapeutics sarcoma; at does of 12 μg/kg intravenously as a single agent in is FDA approved for refractory pediatric ALL.114,115 In pediatric the treatment of gestational trophoblastic neoplasias; and at doses ALL studies, doses ranged from 165 mg/m2 intravenously in com- of 1,000 μg/m2 intravenously on day 1 as part of a combination bination with cytarabine to 250 mg/m2 intravenously weekly in regimen with cyclophosphamide, , , and cis- combination with vincristine and prednisone. Similar to etopo- platin in the treatment of metastatic nonseminomatous testicular side, the dose-limiting toxicity of teniposide is myelosuppression. cancer. Toxicities include myelosuppression, veno-occlusive dis- Additional toxicities include mild-to-moderate nausea, vomiting, ease of the liver, nausea, vomiting, alopecia, erythema, and acne. diarrhea, alopecia, and secondary leukemia. Teniposide is associ- Additionally, similar to doxorubicin, dactinomycin can cause ra- ated with greater frequency of hypersensitivity reactions compared diation recall and severe tissue necrosis in cases of extravasation. to etoposide. Dactinomycin is largely excreted unchanged in the feces and Teniposide is 99% bound to albumin and, as compared to eto- urine. Guidelines for dosing in patients with impaired renal or poside, undergoes hepatic metabolism more extensively and renal liver function are currently not available. clearance less extensively. No specific guidelines are currently available on dose adjustments for renal or hepatic dysfunction. Epipodophyllotoxins THERAPY-RELATED SECONDARY ACUTE Epipodophyllotoxins are glycoside derivatives of podophyllotoxin, LEUKEMIA an antimicrotubule agent extracted from the mandrake plant. Two derivatives, demethylated on the pendant ring (see R1 in One of the major complications of Top2 inhibitor therapies, Fig. 20.3D), etoposide and teniposide were shown to primarily especially for etoposide and mitoxantrone, is acute secondary function as Top2 poisons rather than through antimicrotubule leukemia, which occurs in approximately 5% of patients. Therapy- mechanisms.104,105 Epipodophyllotoxins poison Top2 through a related AMLs (t-AML) are characterized by their relatively rapid mechanism distinct from that of anthracyclines and other DNA 106 onset (they can occur only a few months after therapy) and the intercalators. without intercalating into normal DNA in the ab- presence of recurrent balanced translocations involving the mixed sence of Top2. Therefore, they are “cleaner” Top2 inhibitors than lineage leukemia (MLL) locus on 11q23 and over 50 partner the anthracyclines, anthracenediones, and dactinomycin. How- genes.116 The molecular mechanism is likely from the disjoining ever, etoposide and teniposide trap Top2 cleavage complexes by of two drug-trapped Top2 cleavage complexes on different chro- base stacking in a ternary complex at the interface of the DNA and mosomes (see Fig. 20.2C) in relationship with transcription col- the Top2 homodimer. Mechanisms that have been implicated in lisions and illegitimate relegation.117 Top2β, rather than Top2α, resistance to etoposide include drug efflux, because epipodophyl- has been implicated in the generation of these disjoined cleavage lotoxins are substrates for P-glycoprotein107 ; altered localization of complexes.117,118 Top2α; decreased cellular expression of Top2α108; and impaired phosphorylation of Top2.109 FUTURE DIRECTIONS Etoposide Current challenges in the development of topoisomerase inhibitors Etoposide (see Fig. 20.3D) is available in intravenous and oral lie in the inherent chemical instability of current and established forms. It is FDA approved for the treatment of small-cell lung can- agents. In addition to recent developments designed to enhance cer110 and refractory testicular cancer.111 It also has activity in he- the stability with semisynthetic analogs and the development of matologic malignancies and various solid tumors. The intravenous novel delivery systems in an effort to achieve higher intratumoral form is generally administered at doses of 35 to 50 mg/m2 for 4 to concentrations, attention is also being focused on targeting other 5 days every 3 to 4 weeks in combination therapy for small-cell topoisomerase isoenzymes. Driving this trend has been the recent 2 elucidation of the role of Top2β inhibition in the development lung cancer, and 50 to 100 mg/m for 5 days every 3 to 4 weeks 86,117,118 in combination therapy for refractory testicular cancer. The dose of treatment-related cardiotoxicity and secondary AML. of oral etoposide is usually twice the intravenous dose. Oral bio- In addition to combination chemotherapy regimens already in availability is highly variable due to dependence on intestinal use, attempts have also been made for the sequential inhibition 112 of Top1 and Top2. Based on early preclinical models suggesting P-glycoprotein. 120 The dose-limiting toxicity for etoposide is myelosuppression, synergy with sequential inhibition of Top1 and Top2, phase 1 with white blood cell count nadirs typically occurring on days 10 studies have evaluated the sequential administration of topotecan to 14. Thrombocytopenia is less common than leukopenia. Ad- and etoposide in extensive-stage small-cell lung cancer and ovar- ian cancer, with significant myelosuppression as the dose-limiting ditionally, mild to moderate nausea, vomiting, diarrhea, mucositis, 121,122 and alopecia are associated with etoposide. Among topoisomerase toxicity. Future rational drug combinations include target- inhibitors, epipodophyllotoxins have the greatest association with ing DNA repair pathways in combination with Top1 inhibition, secondary malignancies, with etoposide having the highest risk, although further characterization is needed of the specific DNA with an estimated 4% 6-year cumulative risk.113 The majority of repair and stress response pathways invoked in response to DNA etoposide is cleared unchanged by the kidneys, and a 25% dose damage as a result of Top1 inhibition. However, one such attempt reduction is recommended in patients with a creatinine clearance of combining topotecan with , a small molecule inhibi- of 15 to 50 mL per minute. A 50% dose reduction is recommended tor of poly (ADP-ribose) polymerase, was poorly tolerated due to significant myelosuppression, thus limiting the doses of topotecan in patients with a creatinine clearance less than 15 mL per min- 123 ute. Because the unbound fraction of etoposide is dependent on that could be safely administered. albumin and bilirubin concentrations, dose adjustments for he- Molecular characterization of tumors to better define patient patic dysfunction are advised, but consensus guidelines are cur- selection and the development of pharmacodynamic biomarkers rently not available. to monitor the response to treatment and to optimize the com- bination dose and schedules is needed for the further clinical development of topoisomerase inhibitors. Validated assays have Teniposide been developed to evaluate topoisomerase 1 levels and levels of plosphorylated histone H2AX (gamma-H2AX) as a marker of DNA Teniposide contains a thiophene group in place of the methyl damage response to topoisomerase inhibition,124,125 and are being group on the glucose moiety of etoposide (Fig. 20.3D). Teniposide incorporated in current phase I studies of indenoisoquinolines.72,73

tahir99 - UnitedVRG Chapter 20 Topoisomerase Interactive Agents 227

SELECTED REFERENCES

The full reference list can be accessed at lwwhealthlibrary.com/oncology. 70. Takagi K, Dexheimer TS, Redon C, et al. Novel E-ring camptothecin keto analogues (S38809 and S39625) are stable, potent, and selective topoisom- 1. Nitiss JL. DNA topoisomerase II and its growing repertoire of biological func- erase I inhibitors without being substrates of drug efflux transporters. Mol tions. Nat Rev Cancer 2009;9(5):327–337. Cancer Ther 2007;6(12 Pt 1):3229–3238. 2. Nitiss JL. Targeting DNA topoisomerase II in cancer chemotherapy. Nat Rev 71. Lansiaux A, Léonce S, Kraus-Berthier L, et al. Novel stable camptothecin de- Cancer 2009;9(5):338–350. rivatives replacing the E-ring lactone by a ketone function are potent inhibitors 5. Wang JC. A journey in the world of DNA rings and beyond. Annu Rev of topoisomerase I and promising antitumor drugs. Mol Pharmacol 2007;72(2): Biochem 2009;78:31–54. 311–319. 6. Wang JC. Cellular roles of DNA topoisomerases: a molecular perspective. 74. Antony S, Agama KK, Miao ZH, et al. Novel indenoisoquinolines NSC Nat Rev Mol Cell Biol 2002;3(6):430–440. 725776 and NSC 724998 produce persistent topoisomerase I cleavage com- 16. Pommier Y, Marchand C. Interfacial inhibitors: targeting macromolecular plexes and overcome multidrug resistance. Cancer Res 2007;67:10397–10405. complexes. Nat Rev Drug Discov 2011;11(1):25–36. 75. Kurtzberg LS, Roth S, Krumbholz R, et al. Genz-644282, a novel non- 21. Hsiang YH, Lihou MG, Liu LF. Arrest of DNA replication by drug-stabilized camptothecin topoisomerase I inhibitor for cancer treatment. Clin Cancer topoisomerase I-DNA cleavable complexes as a mechanism of cell killing by Res 2011;17:2777–2787. camptothecin. Cancer Res 1989;49(18):5077–5082. 79. Doroshow JH, Davies KJ. Redox cycling of anthracyclines by cardiac 25. Koster DA, Palle K, Bot ES, et al. Antitumor drugs impede DNA uncoiling mitochondria. II. Formation of superoxide anion, hydrogen peroxide, and by topoisomerase I. Nature 2007;448(7150):213–217. hydroxyl radical. J Biol Chem 1986;261:3068–3074. 31. Pommier Y. Drugging topoisomerases: lessons and challenges. ACS Chem 82. Felix CA, Kolaris CP, Osheroff N. Topoisomerase II and the etiology of Biol 2013;8(1):82–95. chromosomal translocations. DNA Repair 2006;5:1093–1108. 33. Wall ME, Wani MC. Camptothecin and taxol: discovery to clinic—thirteenth 83. O’Brien ME, Wigler N, Inbar M, et al. Reduced cardiotoxicity and Bruce F. Cain Memorial Award Lecture. Cancer Res 1995;55:753–760. comparable efficacy in a phase III trial of pegylated liposomal doxorubicin 36. Pommier Y, Cushman M. The indenoisoquinoline noncamptothecin topoi- HCl (CAELYX/Doxil) versus conventional doxorubicin for first-line treat- somerase I inhibitors: update and perspectives. Mol Cancer Ther 2009;8(5): ment of metastatic breast cancer. Ann Oncol 2004;15(3):440–449. 1008–1014. 85. Zhang S, Liu X, Bawa-Khalfe T, et al. Identification of the molecular basis of 37. Douillard JY, Cunningham D, Roth AD, et al. Irinotecan combined with doxorubicin-induced cardiotoxicity. Nat Med 2012;18:1639–1642. compared with fluorouracil alone as first-line treatment for met- 90. Swain SM, Whaley FS, Gerber MC, et al. Cardioprotection with dexrazoxane astatic colorectal cancer: a multicentre randomised trial. Lancet 2000;355: for doxorubicin-containing therapy in advanced breast cancer. J Clin Oncol 1041–1047. 15(4):1318–1332. 38. Saltz LB, Cox JV, Blanke C, et al. Irinotecan plus fluorouracil and leucovorin for 92. Kalam K, Marwick TH. Role of cardioprotective therapy for prevention of metastatic colorectal cancer. Irinotecan Study Group. N Engl J Med 2000;343: cardiotoxicity with chemotherapy: a systematic review and meta-analysis. Eur 905–914. J Cancer 2013;49:2900–2909. 39. Conroy T, Desseigne F, Tchou M, et al. FOLFIRINOX versus 93. Tannock IF, Osoba D, Stockler MR, et al. Chemotherapy with mitoxantrone for metastatic pancreatic cancer. N Engl J Med 2011;364:1817–1825. plus prednisone or prednisone alone for symptomatic hormone-resistant 40. Hanna N, Bunn PA Jr, Langer C, et al. Randomized phase III trial comparing prostate cancer: a Canadian randomized trial with palliative end points. irinotecan/cisplatin with etoposide/cisplatin in patients with previously un- J Clin Oncol 1996;14:1756–1764. treated extensive-stage disease small-cell lung cancer. J Clin Oncol 2006;24: 95. Shenkenberg TD, Von Hoff DD. Mitoxantrone: a new anticancer drug with

2038–2043. significant clinical activity. Ann Intern Med 1986;105:67–81. CANCER THERAPEUTICS 43. Innocenti F, Undevia SD, Iyer L, et al. Genetic variants in the UDP- 104. Chen GL, Yang L, Rowe TC, et al. Nonintercalative antitumor drugs glucuronosyltransferase 1A1 gene predict the risk of severe neutropenia of interfere with the breakage-reunion reaction of mammalian DNA topoisom- irinotecan. J Clin Oncol 2004;22:1382–1388. erase. J Biol Chem 1984;259(21):13560–13566. 50. Young C, Schluep T, Hwang J, et al. CRLX101 (formerly IT-101)-A novel 106. Ross W, Rowe T, Glisson B, et al. Role of topoisomerase II in mediating epi- nanopharmaceutical of camptothecin in clinical development. Curr Bioact podophyllotoxin-induced DNA cleavage. Cancer Res 1984;44:5857–5860. Compd 2011;7:8–14. 110. Sundstrøm S, Bremnes RM, Kaasa S, et al. Cisplatin and etoposide regimen 55. Vergote IB, Garcia A, Micha J et al. Randomized multicentre phase II trial is superior to cyclophosphamide, epirubicin, and vincristine regimen in comparing two schedules of etirinotecan pegol (NKTR-102) in women small-cell lung cancer: results from a randomized phase III trial with 5 years’ with recurrent platinum-resistant/refractory epithelial ovarian cancer. J Clin follow-up. J Clin Oncol 2002;20:4665–4672. Oncol 2013;31(32):4060–4066. 111. Nichols CR, Catalano PJ, Crawford ED, et al. Randomized comparison of 62. Basili S, Moro S. Novel camptothecin derivatives as topoisomerase I inhibi- cisplatin and etoposide and either bleomycin or in treatment of tors. Expert Opin Ther Pat 2009;19:555–574. advanced disseminated germ cell tumors: an Eastern Cooperative Oncology 66. Frapolli R, Zucchetti M, Sessa C, et al. Clinical of Group, Southwest Oncology Group, and Cancer and Leukemia Group B the new oral camptothecin gimatecan: the inter-patient variability is re- Study. J Clin Oncol 1998;16:1287–1293. lated to alpha1-acid glycoprotein plasma levels. Eur J Cancer 2010;46: 113. Smith MA, Rubinstein L, Anderson JR, et al. Secondary leukemia or 505–516. myelodysplastic syndrome after treatment with epipodophyllotoxins. J Clin 69. Trocóniz IF, Cendrós JM, Soto E, et al. Population pharmacokinetic/ Oncol 1999;17:569–577. pharmacodynamics modeling of drug-induced adverse effects of a novel 116. Lovett BD, Lo Nigro L, Rappaport EF, et al. Near-precise interchromosomal homocamptothecin analog, elomotecan (BN80927), in a Phase I dose recombination and functional DNA topoisomerase II cleavage sites at MLL and finding study in patients with advanced solid tumors. Cancer Chemother AF-4 genomic breakpoints in treatment-related acute lymphoblastic leukemia Pharmacol 2012;70:239–250. with t(4;11) translocation. Proc Natl Acad Sci U S A 2001;98(17):9802–9807. 21 Antimicrotubule Agents

Christopher J. Hoimes and Lyndsay N. Harris

MICROTUBULES the inhibition of the 5′-triphosphate–dependent efflux pump of P-glycoprotein. are vital and dynamic cytoskeletal polymers that Paclitaxel initially received regulatory approval in the United play a critical role in cell division, signaling, vesicle transport, States in 1992 for the treatment of patients with ovarian can- shape, and polarity, which make them attractive targets in an- cer after failure of first-line or subsequent chemotherapy (Table ticancer regimens and drug design.1 Microtubules are com- 21.1).1,4 Subsequently, it has been approved for several other posed of 13 linear protofilaments of polymerized α/β-tubulin indications, including advanced breast cancer after anthracy- heterodimers arranged in parallel around a cylindrical axis and cline-based regimens6; combination chemotherapy of lymph associated with regulatory proteins such as -asso- node–positive breast cancer in the adjuvant setting7; advanced ciated proteins, tau, and motor proteins kinesin and dynein.2 ovarian cancer in combination with a platinum compound; sec- The specific biologic functions of microtubules are due to their ond-line treatment of AIDS-related Kaposi sarcoma; and first-line unique polymerization dynamics. Tubulin polymerization is treatment of non–small-cell lung cancer (NSCLC) in combina- mediated by a nucleation-elongation mechanism. One end of tion with cisplatin8 (see Table 21.1). In addition to the U.S. Food the microtubules, termed the plus end, is kinetically more dy- and Drug Administration (FDA) on-label indications, paclitaxel namic than the other end, termed the minus end (Fig. 21.1). is widely used for several other tumor types, such as cancers of Microtubule dynamics are governed by two principal processes unknown origin, bladder, esophagus, gastric, head and neck, and driven by guanosine 5′-triphosphate (GTP) hydrolysis: tread- cervical cancers. The U.S. patent for paclitaxel expired in 2002, milling or poleward flux is the net growth at one end of the and a generic form of paclitaxel is now available. microtubule and the net shortening at the opposite end, and dy- was first approved for use in the United States in namic instability, which is a process in which the microtubule 1996 for patients with metastatic breast cancer that progressed or ends switch spontaneously between states of slow sustained relapsed after anthracycline-based chemotherapy, which was later growth and rapid depolymerization.2 Antimicrotubule agents broadened to a general second-line indication (see Table 21.1).4,6 are tubulin-binding drugs that directly bind tubules, inhibitors Subsequently, it received regulatory approval in adjuvant chemo- of tubulin-associated scaffold kinases, or inhibitors of their asso- therapy of stage II breast cancer in combination with Adriamycin ciated mitotic motor proteins to, ultimately, disrupt microtubule and cyclophosphamide (TAC)9, and first-line treatment for locally dynamics. They are broadly classified as microtubule stabilizing advanced or metastatic breast cancer.10 In addition, docetaxel or microtubule destabilizing agents according to their effects on has indications in nonresectable, locally advanced, or metastatic tubulin polymerization. NSCLC after failure of or in combination with cisplatin therapy; metastatic castration-resistant prostate cancer in combination with prednisone11; first-line treatment of gastric adenocarcinoma, including gastroesophageal junction adenocarcinoma in combi- nation with cisplatin and 5-fluorouracil (5-FU)12; and inoperable locally advanced squamous cell cancer of the head and neck in Taxanes were the first-in-class microtubule stabilizing drugs. combination with cisplatin and 5-FU (see Table 21.1). Docetaxel Ancient medicinal attempts at cardiac pharmacotherapy using came off patent in 2010 and a generic form is available. material from the toxic coniferous yew tree, Taxus spp., were likely related to the plant’s alkaloid taxine effect on sodium and calcium channels. compounds are the result of a drug Mechanism of Action screening of 35,000 plant extracts in 1963 that led to the identi- fication of activity from the bark extract of the Pacific yew tree, The unique mechanism of action for paclitaxel was initially de- Taxus brevifolia. Paclitaxel was identified as the active constitu- fined by Schiff et al.13 in 1979, who showed that it bound to the in- ent with a report of its activity in carcinoma cell lines in 1971.3 terior surface of the microtubule lumen at binding sites completely Motivation to identify taxanes derived from the more abundant distinct from those of exchangeable GTP, colchicine, podophyllo- and available needles of Taxus baccata led to the development toxin, and the vinca alkaloids.14 The taxanes profoundly alter the of docetaxel, which is synthesized by the addition of a side chain tubulin dissociation rate constants at both ends of the microtubule, to 10-deacetylbaccatin III, an inactive taxane precursor.4 The suppressing treadmilling and dynamic instability. Dose-dependent taxane rings of paclitaxel and docetaxel are linked to an ester taxane β-tubular binding induces mitotic arrest at the G2/M tran- side chain attached to the C13 position of the ring, which is sition and induces cell death. By stabilizing microtubules, they essential for antimicrotubule and antitumor activity. Nanopar- also can stall ligand-dependent intracellular trafficking, as shown ticle albumin-bound paclitaxel (nab-paclitaxel) is a formulation in sequestration of the androgen receptor to the cytosol in meta- that avoids the solvent related side effects of non–water-soluble static prostate cancer patients treated with docetaxel, and is asso- paclitaxel and docetaxel. Overcoming docetaxel and paclitaxel’s ciated with decreased androgen-regulated gene expression, such susceptibility to the P-glycoprotein efflux pump led to the devel- as prostate-specific antigen (PSA).15,16 Peripheral neuropathy is a opment of .5 Cabazitaxel is synthesized by adding two common dose-limiting toxicity across the antimicrotubule agents methoxy groups to the 10-deacetylbaccatin III, which results in and likely is a result of their direct effect on microtubules. Studies 228

tahir99 - UnitedVRG Chapter 21 Antimicrotubule Agents 229

GTP Microtubule Microtubule stabilizers destabilizers Plus end

MAP α Kinesin Colchicine Taxane α binding site Tubulin polymer β β Estramustine Dynein binding site binding site Vinca- 8nm maytansinoid-auristatin binding site Minus end 24nm

Figure 21.1 Antimicrotubule agents bind tubulin directly or inhibit its associated proteins. Taxanes and have distinct binding pockets within the same site on the interior surface of the tubule. Estramustine has a distinct site on β-tubulin, although it also directly binds microtubule-associated proteins (MAP). (Adapted from Lieberman M, Marks A. Mark’s Basic Medical Biochemistry: A Clinical Approach. 3rd ed. Philadelphia: Lippincott Williams & Wilkins; 2009.) have shown that they inhibit anterograde and/or retrograde fast Nanoparticle Albumin-Bound Paclitaxel axonal transport and can explain the demyelinating “dying back” pattern seen and the vulnerability of sensory neurons with the lon- Nab-paclitaxel is a solvent-free colloidal suspension made by ho- gest axonal projections.17 mogenizing paclitaxel with 3% to 4% albumin under high pres- sure to form nanoparticles of ∼130 nm that disperse in plasma to Recent evidence suggests that microtubule inhibitors have ∼ 25 collateral effects during interphase that lead to cell death. For in- 10 nm (see Table 21.1). It received regulatory approval in the stance, paclitaxel-stabilized microtubules serve as a scaffold for the United States in 2005 based on results in patients with metastatic CANCER THERAPEUTICS binding of the death-effector domain of pro-caspase-8, and thereby breast cancer, and is now also approved in combination with car- enabling a caspase-8 downstream proteolytic cascade.18,19 This cas- boplatin for first-line treatment of locally advanced or metastatic NSCLC, and in combination with gemcitabine for first-line treat- pase-8–dependent mechanism also serves as an important basis for 26–28 the understanding of the loss of function and/or low expression of ment of metastatic pancreatic adenocarcinoma. The improved the breast cancer 1, early onset gene (BRCA1) association with responses seen with nab-paclitaxel, when compared to solvent- resistance to taxane therapy.20 based paclitaxel, are not fully understood. Nab-paclitaxel likely Another mechanism of the anticancer effect of taxanes is capitalizes on several mechanisms, which include an improved currently being elaborated and is tied to the B-cell lymphoma-2 pharmacokinetic profile with a larger volume of distribution and a (Bcl-2) antiapoptosis family of proteins. Paclitaxel has been shown higher maximal concentration of circulating, unbound, free drug; to cause the phosphorylation of Bcl-2 and the sequestration of Bak improved tumor accumulation by the enhanced permeability and and Bim; however, this seemingly cancer-protective phosphoryla- retention (EPR) effect; and receptor-mediated transcytosis via an tion needs to be reconciled and likely correlates with Bcl-2–ex- albumin-specific receptor (gp60) for endothelial transcytosis and 21–23 binding of secreted protein acidic and rich in cysteine (SPARC) pression levels. Interestingly, neutralizing Bcl-2 homology 3 29,30 (BH3) domains with compounds such as ABT-737 is synergistic in the tumor interstitium. In contrast to cremophor/ethanol with docetaxel.24 (CrEL) solvent-based paclitaxel, nab-paclitaxel exhibits an exten- sive extravascular volume of distribution exceeding that of water, indicating extensive tissue and extravascular protein distribution. Clinical Pharmacology Some studies show that nab-paclitaxel achieves 33% higher drug concentration over CrEL-paclitaxel.31 Additionally, the maximum Paclitaxel concentration (Cmax), the mean plasma half-life of 15 to 18 hours, the area under curve (AUC), and the dose-independent plasma With prolonged infusion schedules (6 and 24 hours), drug disposi- clearance correspond to linear pharmacokinetics over 80 to 300 tion is a biphasic process with values for alpha and beta half-lives mg/m2.29,32 The improved deposition of a nanoparticle, such as averaging approximately 20 minutes and 6 hours, respectively.4 nab-paclitaxel in a tumor tissue, can occur passively through an When administered as a 3-hour infusion, the pharmacokinetics are EPR effect in areas of leaky vasculature, sufficient vascular pore nonlinear and may lead to unexpected toxicity with a small dose size, and decreased lymphatic flow.25,33 Once in the tissue, the escalation, or a disproportionate decrease in drug exposure and nab-paclitaxel nanovehicle can deliver the drug locally or benefit loss of tumor response with a dose reduction. Approximately 71% from further receptor-mediated targeting to SPARC, which has of an administered dose of paclitaxel is excreted in the stool via been shown to be overexpressed, and correlates with disease pro- the enterohepatic circulation over 5 days as either the parent com- gression in many tumor types.34–38 Although preclinical models, as pound or metabolites in humans. Renal clearance of paclitaxel and well as one clinical trial, have shown how nanoparticle therapy can metabolites is minimal, accounting for 14% of the administered benefit from this targeted approach,39,40 correlative data for nab- dose. In humans, the bulk of drug disposition is metabolized by paclitaxel is limited. The high stromal SPARC level was associated cytochrome P-450 mixed-function oxidases—specifically, the iso- with longer survival in patients treated with nab-paclitaxel in the enzymes CYP2C8 and CYP3A4, which metabolize paclitaxel to hy- phase I/II study of patients with pancreatic cancer; however, this droxylated 3′p-hydroxypaclitaxel (minor) and 6α-hydroxypaclitaxel correlative analysis was not included in the phase III trial report (major), as well as dihydroxylated metabolites. and requires validation.28,41 230 Cancer Therapeutics

TABLE 21.1 Antimicrotubule Agents: Dosages and Toxicities

Chemotherapeutic Agent Dosage Indications Common Toxicities Paclitaxel 135–200 mg/m2 IV over 3 h or 135 Adjuvant therapy of node-positive Myelosuppression, mg/m2 IV over 24 h every 3 wk; or breast cancer; metastatic breast, hypersensitivity, nausea and 80 mg/m2 IV over 1 h weekly ovarian, non–small-cell lung, bladder, vomiting, alopecia, arthralgia, esophagus, cervical, gastric, and head myalgia, peripheral neuropathy and neck cancer; AIDS-related Kaposi sarcoma; cancer of unknown origin Docetaxel 60–100 mg/m2 IV over 1 h every Adjuvant therapy of node-positive Myelosuppression, 3 wk breast cancer; metastatic breast, hypersensitivity, edema, gastric, head and neck, prostate, alopecia, nail damage, rash, non–small-cell lung, and diarrhea, nausea, vomiting, ovarian cancer asthenia, neuropathy Cabazitaxel 25 mg/m2 IV every 3 wk over 1 h Docetaxel-refractory metastatic Neutropenia, infections, castration resistant prostate cancer myelosuppression, diarrhea, nausea, vomiting, constipation, abdominal pain, asthenia Nab-paclitaxel 260 mg/m2 IV over 30 min every Metastatic breast cancer, non–small- Myelosuppression, nausea, 3 wk; or 125 mg/m2 IV weekly on cell lung cancer, pancreatic cancer vomiting, alopecia, myalgia, days 1, 8, and 15 every 28 d peripheral neuropathy 40 mg/m2 IV over 3 h every 3 wk Metastatic and locally advanced Myelosuppression, fatigue/ breast cancer asthenia, myalgia/arthralgia, alopecia, nausea, vomiting, stomatitis/mucositis, diarrhea, musculoskeletal pain Vincristine 0.5–1.4 mg/m2/wk IV (maximum Lymphoma, acute leukemia, Constipation, nausea, 2 mg per dose); or 0.4 mg/d neuroblastoma, rhabdomyosarcoma, vomiting, alopecia, diplopia, continuous infusion for 4 d AIDS-related Kaposi sarcoma, myelosuppression multiple myeloma, testicular cancer Vinblastine 6 mg/m2 IV on days 1 and 15 as Hodgkin and non-Hodgkin lymphoma; Myelosuppression, part of the ABVD regimen; 0.15 mg/ Kaposi sarcoma; breast, testicular, constipation, alopecia, kg IV on days 1 and 2 as part of the bladder, prostate, and renal cell malaise, bone pain PVB regimen; 3 mg/m2 IV as part of cancer days 2, 15, 22 MVAC regimen 25–30 mg/m2 IV weekly Non–small-cell lung, breast, cervical, Alopecia, diarrhea, nausea, and ovarian cancer vomiting, asthenia, neuromyopathy Estramustine 14 mg/kg PO daily in 3 or 4 divided Metastatic prostate cancer Nausea, vomiting, doses gynecomastia, fluid retention Ado-trastuzumab 3.6 mg/kg IV every 3 wk Metastatic breast cancer Thrombocytopenia, nausea, emtansine constipation or diarrhea, peripheral neuropathy, fatigue, increased AST/ALT Brentuximab vedotin 1.8 mg/kg every 3 wk, maximum Refractory Hodgkin lymphoma, Neutropenia, anemia, dose 180 mg refractory systemic anaplastic large thrombocytopenia, fatigue, cell lymphoma fever, peripheral neuropathy

ABVD, doxorubicin (Adriamycin), bleomycin, vinblastine, ; PVB, cisplatin, vinblastine, bleomycin; MVAC, , vinblastine, doxorubicin (Adriamycin), cisplatin; IV, intravenous; PO, by mouth; AST/ALT, aspartate amniotransferase–alanine amniotransferase.

Docetaxel P-450 mixed-function oxidases, particularly isoforms CYP3A4 and CYP3A5, are principally involved in biotransformation. The The pharmacokinetics of docetaxel on a 1-hour schedule is tri- 2 4 principal pharmacokinetic determinants of toxicity, particularly exponential and linear at doses of 115 mg/m or less. Terminal neutropenia, are drug exposure and the time that plasma con- half-lives ranging from 11.1 to 18.5 hours has been reported. centrations exceed biologically relevant concentrations. The The most important determinants of docetaxel clearance were baseline level of α -acid glycoprotein may be elevated as an α 1 the body surface area (BSA), hepatic function, and plasma 1- acute phase reactant in advanced disease and is an independent acid glycoprotein concentration. is high predictor of response and a major objective prognostic factor of (greater than 80%), and binding is primarily to α1-acid glyco- survival in patients with non–small-cell lung cancer treated with protein, albumin, and lipoproteins. The hepatic cytochrome docetaxel chemotherapy.

tahir99 - UnitedVRG Chapter 21 Antimicrotubule Agents 231

Cabazitaxel Neuropathy is the principal toxicity of paclitaxel. Paclitaxel in- duces a peripheral neuropathy that presents in a symmetric stocking Cabazitaxel is a semisynthetic derivative of the natural taxoid 48 β glove distribution, at first transient and then persistent. A neuro- 10-deacetylbaccatin III. It binds to and stabilizes the -tubulin logic examination reveals sensory loss, and neurophysiologic studies subunit, resulting in the inhibition of microtubule depolymeriza- reveal axonal degeneration and demyelination.48 Compared with cis- tion and cell division, arrest in the G2/M phase, and the 5 platin, a loss of deep tendon reflexes occurs less commonly; however, inhibition of tumor cell proliferation. It is active against diverse autonomic and motor changes can occur. Severe neurotoxicity is un- cancer cell lines and tumor models that are sensitive and resistant common when paclitaxel is given alone at doses below 200 mg/m2 to docetaxel, including prostate, mammary, melanoma, kidney, 2 5 on a 3- or 24-hour schedule every 3 weeks, or below 100 mg/m on a colon, pancreas, lung, gastric, and head and neck. Cabazitaxel continuous weekly schedule. There is no convincing evidence that is a poor substrate for the membrane-associated, multidrug resis- any specific measure is effective at ameliorating existing manifesta- tance P-glycoprotein efflux pump; therefore, is useful for treating tions or preventing the development or worsening of neurotoxicity.48 docetaxel-refractory prostate cancer for which it gained FDA ap- 5 42 Neutropenia is also frequent with paclitaxel. The onset is usu- proval in 2010. In addition, it penetrates the blood–brain barrier. ally on days 8 to 11, and recovery is generally complete by days 15 Pharmacokinetics of cabazitaxel is similar to docetaxel; however, to 21 with an every 3 weeks dosing regimen. Neutropenia is noncu- cabazitaxel has a larger volume of distribution and a longer termi- 43,44 mulative, and the duration of severe neutropenia—even in heav- nal half-life (mean 77.3 hours versus 11.2 hours for docetaxel). ily pretreated patients—is usually brief. Severity of neutropenia is related to the duration of exposure above the biologically relevant levels of 0.05 to 0.10 μM/L, and paclitaxel’s nonlinear pharmaco- Tesetaxel (DJ-927, XRP6258) is a semisynthetic, orally bioavail- kinetics should be considered whenever adjusting dose.49 able taxane currently in clinical trials in breast, gastric, and pros- The most common cardiac rhythm disturbance, a transient tate cancer. Administration in phase I and II trials has been once sinus bradycardia, can be observed in up to 30% of patients. Rou- per week or every 3 weeks and not associated with hypersensi- tine cardiac monitoring during paclitaxel therapy is not necessary tivity and possibly less neurotoxicity compared to other taxanes. but is advisable for patients who may not be able to tolerate brady- Dose-limiting toxicity has been neutropenia. Overall responses in arrhythmias. Drug-related gastrointestinal effects, such as vomiting phase II studies have been 50% and 38% in patients treated for and diarrhea, are uncommon. Severe hepatotoxicity and pancre- first- and second-line breast cancer, respectively. A phase I/II study atitis have also been noted rarely. Pulmonary toxicities, including in advanced NSCLC showed an overall response rate of 5.6%. acute bilateral pneumonitis, have been reported. Extravasation of Tesetaxel activity is independent of P-glycoprotein expression.45 large volumes can cause moderate soft tissue injury. Paclitaxel also Pharmacokinetics on a schedule of every 3 weeks have an AUC induces reversible alopecia of the scalp in a dose-related fashion. of ∼1,750 ng/mL per hour, a half life of ∼170 hours, and no drug Nail disorders have also been reported with paclitaxel use and interactions that have been noted.46 include ridging, nail bed pigmentation, onychorrhexis, and ony- CANCER THERAPEUTICS cholysis. These side effects have been reported more commonly with dose-intensified paclitaxel regimens. Drug Interactions Recent studies have suggested a role for the adenosine triphos- phatase (ATP)-binding cassette (ABC) transporter polymorphisms Sequence-dependent pharmacokinetic and toxicologic interactions in the development of neuropathy and neutropenia. Sissung et al.50 between paclitaxel and several other chemotherapy agents have reported that patients carrying two reference alleles for the ABCB1 been noted. The sequence of cisplatin followed by paclitaxel (on (P-glycoprotein, MDR1) 3435C greater than T polymorphism had a 24-hour schedule) induces more profound neutropenia than the a reduced risk to develop neuropathy as compared to patients car- reverse sequence, which is explained by a 33% reduction in the = 47 rying at least one variant allele (P .09). Data from a large con- clearance of paclitaxel after cisplatin. Treatment with paclitaxel trolled trial to evaluate these and other candidate polymorphisms on either a 3- or 24-hour schedule followed by carboplatin has been failed to detect a significant association between genotype and demonstrated to produce equivalent neutropenia and less thrombo- outcome or toxicity for any of the genes analyzed, although the cytopenia as compared to carboplatin as a single agent, which is not correlative studies were retrospective and the sample size was in- explained by pharmacokinetic interactions. Neutropenia and mu- adequate to rule out smaller differences.51 A large randomized trial cositis are more severe when paclitaxel is administered on a 24-hour of the CALGB 40101 using an integrated genomewide associate schedule before doxorubicin, compared to the reverse sequence, study found two polymorphisms associated with paclitaxel-in- which is most likely due to an approximately 32% reduction in duced polyneuropathy.52 Both are involved in nerve development the clearance rates of doxorubicin and doxorubicinol when doxo- and maintenance, including the hereditary peripheral neuropathy rubicin is administered after paclitaxel. Several agents that inhibit Charcot-Marie-Tooth disease gene, FGD4. Further studies are re- cytochrome P-450 mixed-function oxidases interfere with the me- quired to adequately assess the role of these variants in predicting tabolism of paclitaxel and docetaxel in human microsomes in vitro; toxicity from taxane therapy. however, the clinical relevance of these findings is not known.47 Nab-paclitaxel Toxicity Hypersensitivity reactions have not been observed during the infu- Paclitaxel sion period and, therefore, steroid premedications are not neces- sary. The main dose-limiting toxicities are neutropenia and sensory The micelle-forming CrEL vehicle, which is required for suspen- neuropathy. In a trial comparing weekly paclitaxel 90 mg/m2 to sion and intravenous delivery of paclitaxel, causes its nonlinear nab-paclitaxel 150 mg/m2 to ixabepilone in patients with meta- pharmacokinetics and thereby impacts its therapeutic index. static breast cancer, there was more hematologic toxicity and pe- CrEL causes hypersensitivity reactions, with major reactions usu- ripheral neuropathy in the nab-paclitaxel arm compared to the ally occurring within the first 10 minutes after the first treatment paclitaxel arm, although median progression-free survival was not and resolving completely after stopping the treatment. All patients significantly different at the 12-month follow-up.53 This led to dose should be premedicated with steroids, diphenhydramine, and an reductions in 45% of patients in the nab-paclitaxel arm compared H2 antagonist, although up to 3% will still have reactions. Those with 15% for the paclitaxel arm.53 Other toxicities include alope- who have major reactions have been rechallenged successfully cia, diarrhea, nausea and vomiting, elevations in liver enzymes, after receiving high doses of corticosteroids. arthralgia, myalgia, and asthenia. 232 Cancer Therapeutics

Docetaxel malignancies (see Table 21.1).1 Their clinical efficacy in several 4 combination therapies has led to the development of various novel Neutropenia is the main toxicity of docetaxel. When docetaxel is semisynthetic analogs, including vinorelbine (VRL), administered on an every 3 weeks schedule, the onset of neutrope- (VDS), and (VFL). nia is usually noted on day 8, with complete resolution by days 15 to 21. Neutropenia is significantly less when low doses are admin- istered weekly. FDA black box warnings include increased toxicity Mechanism of Action in patients with abnormal liver function and, in select NSCLC patients that received prior platinum, severe hypersensitivity reac- In contrast to the taxanes, the vinca alkaloids depolymerize micro- tions and severe fluid retention despite dexamethasone at-home tubules and destroy mitotic spindles.1 At low but clinically relevant premedication. concentrations, VBL does not depolymerize spindle microtubules, Hypersensitivity reactions were noted in approximately 31% of yet it powerfully blocks mitosis. This has been suggested to occur patients who received the drug without premedications in early as a result of the suppression of microtubule dynamics rather than studies.4 Symptoms include flushing, rash, chest tightness, back microtubule depolymerization. This group of compounds binds pain, dyspnea, and fever or chills. Severe hypotension, broncho- to the β subunit of tubulin dimers at a distinct region called the spasm, generalized rash, and erythema may also occur.54 Major vinca-binding domain. Importantly, VBL binding induces a con- reactions usually occur during the first two courses and within min- formational change in tubulin in connection with tubulin self- utes after the start of treatment. Signs and symptoms generally re- association. In mitotic spindles, the slowing of the growth and solve within 15 minutes after cessation of treatment, and docetaxel shortening or treadmilling dynamics of the microtubules block can usually be reinstituted without sequelae after treatment with mitotic progression. Disruption of the normal mitotic spindle diphenhydramine and an H2-. Docetaxel in- assembly leads to delayed cell cycle progress with chromosomes duces a unique fluid retention syndrome characterized by edema, stuck at the spindle poles and unable to pass from metaphase into weight gain, and third-space fluid collection. Fluid retention is cu- anaphase, which eventually induces to apoptosis. The naturally mulative and is due to increased capillary permeability. Prophylac- occurring vinca alkaloids VCR and VBL, the semisynthetic analog tic treatment with corticosteroids has been demonstrated to reduce VRL, and a novel bifluorinated analog VFL have similar mecha- the incidence of fluid retention. Aggressive and early treatment nisms of action. with diuretics has been successfully used to manage fluid reten- Tissue and tumor sensitivities to the vinca alkaloids, which, in tion. Skin toxicity may occur in as many as 50% to 75% of patients; part, relate to differences in drug transport and accumulation, also however, premedication may reduce the overall incidence of this vary. Intracellular or extracellular concentration ratios range from effect.4 Other cutaneous effects include palmar–plantar erythro- five- to 500-fold depending on the individual cell type, lipophi- dysesthesia and onychodystrophy. Docetaxel produces neurotox- licity, tissue-specific factors such as tubulin isotype composition, icity, which is qualitatively similar to that of paclitaxel; however, and tissue-specific microtubule-associated proteins (MAP).59–61 neurosensory and neuromuscular effects are generally less fre- Although the vinca alkaloids are retained in cells for long periods quent and less severe than with paclitaxel. Mild-to-moderate pe- of time and thus may have prolonged cellular effects, intracellular ripheral neurotoxicity occurs in approximately 40% of untreated retention is markedly different among the various vinca alkaloids. patients.55 Asthenia has been a prominent complaint in patients For instance, VBL appears to be retained in lipophilic tissue much who have been treated with large cumulative doses. Stomatitis ap- more than either VCR or VDS.59 Newer theories of antimicrotu- pears to occur more frequently with docetaxel than with paclitaxel. bule agents’ mechanism of action have emerged, suggesting that Other reported toxicities of note include necrotizing enterocolitis, the more important target of these drugs may be the tumor vascu- interstitial pneumonitis, and organizing pneumonia.56,57 lature, as reviewed in the next section. Cabazitaxel Clinical Pharmacology A phase III multi-institutional study of men with metastatic castra- tion-resistant prostate cancer who had failed docetaxel improved The vinca alkaloids are usually administered intravenously as a overall median survival on cabazitaxel compared to mitoxan- 58 brief infusion, and their pharmacokinetic behavior in plasma has trone. Cabazitaxel was approved by the FDA in June 2010 to generally been explained by a three-compartment model. The treat metastatic castration-resistant prostate cancer in those who vinca alkaloids share many pharmacokinetic properties, including had received prior chemotherapy. This was despite a higher rate large volumes of distribution, high clearance rates, and long termi- of adverse deaths (4.9%), a third of which were due to neutro- nal half-lives that reflect the high magnitude and avidity of drug penic sepsis. Cabazitaxel was associated with more grade 3 or 4 binding in peripheral tissues. VCR has the longest terminal half- neutropenia (82%) than mitoxantrone (58%). Side effects reported life and the lowest clearance rate; VBL has the shortest terminal in more than 20% of patients treated with cabazitaxel included half-life and the highest clearance rate; and VDS has intermediate myelosuppression, diarrhea, nausea, vomiting, constipation, ab- characteristics. Although prolonged infusion schedules may avoid dominal pain, or asthenia. FDA black box warnings are similar to excessively toxic peak concentrations and increase the duration those for docetaxel. of drug exposure in plasma above biologically relevant threshold concentrations, there is little evidence to support the notion that prolonged infusions are more effective than bolus schedules. The VINCA ALKALOIDS longest half-life and lowest clearance rate of VCR may account for its greater propensity to induce neurotoxicity, but there are many The vinca alkaloids have been some of the most active agents in other nonpharmacokinetic determinants of tissue sensitivity, as dis- cancer chemotherapy since their introduction 40 years ago. The cussed in the previous section. naturally occurring members of the family, vinblastine (VBL) and vincristine (VCR), were isolated from the leaves of the periwinkle plant Catharanthus roseus G. Don. In the late 1950s, their antimi- Vincristine totic and, therefore, cancer chemotherapeutic potential was dis- covered by groups both at Eli Lilly Research Laboratories and at After conventional doses of VCR (1.4 mg/m2) given as brief infu- the University of Western Ontario, and they came into widespread sions, peak plasma levels approach 0.4 μmol. Plasma clearance is use for the single-agent treatment of childhood hematologic slow, and terminal half-lives that range from 23 to 85 hours have and solid malignancies and, shortly after, for adult hematologic been reported. VCR is metabolized and excreted primarily by the

tahir99 - UnitedVRG Chapter 21 Antimicrotubule Agents 233 hepatobiliary system. The nature of the VCR metabolites identi- of the vinca alkaloids, which may result in increased vinca-related fied to date, as well as the results of metabolic studies in vitro, toxicity. To minimize the possibility of this interaction, the vinca indicate that VCR metabolism is mediated principally by hepatic alkaloids should be given 12 to 24 hours before L-. cytochrome P-450 CYP3A5. The combined use of and the vinca alkaloids has been associated with acute dyspnea and bronchospasm. The onset Vinblastine of these pulmonary toxicities has ranged from within minutes to hours after treatment with the vinca alkaloids, or up to 2 weeks after mitomycin C. The clinical pharmacology of VBL is similar to that of VCR. VBL Treatment with the vinca alkaloids has precipitated seizures as- binding to plasma proteins and formed elements of blood is exten- sociated with subtherapeutic plasma phenytoin concentrations.66 sive.62,63 Peak plasma drug concentrations are approximately 0.4 Reduced plasma phenytoin levels have been noted from 24 hours μm after rapid intravenous injections of VBL at standard doses. to 10 days after treatment with VCR and VBL. Because of the Distribution is rapid, and terminal half-lives range from 20 to 24 importance of the cytochrome P-450 CYP3A isoenzyme in vinca hours. Like VCR, VBL disposition is principally through the hepa- alkaloid metabolism, administration of the vinca alkaloids with tobiliary system with excretion in feces (approximately 95%); how- erythromycin and other inhibitors of CYP3A may lead to severe ever, fecal excretion of the parent compound is low, indicating that toxicity.67 Concomitantly administered drugs, such as pentobarbi- hepatic metabolism is extensive.59 tal and H2-receptor antagonists, may also influence VCR clearance by modulating hepatic cytochrome P-450 metabolic processes.66 Vinorelbine

The pharmacologic behavior of VRL is similar to that of the other Toxicity vinca alkaloids, and plasma concentrations after rapid intravenous administration have been reported to decline in either a biexpo- Despite close similarities in structure, the vinca alkaloids differ in nential or triexponential manner.64 After intravenous administra- their safety profiles. Neutropenia is the principal dose-limiting tox- tion, there is a rapid decay of VRL concentrations followed by a icity of VBL and VRL. Thrombocytopenia and anemia occur less much slower elimination phase (terminal half-life, 18 to 49 hours). commonly. The onset of neutropenia is usually day 7 to 11, with Plasma protein binding, principally to α1-acid glycoprotein, albu- recovery by day 14 to 21, and can be potentiated by hepatic dys- min, and lipoproteins, has been reported to range from 80% to function. Gastrointestinal autonomic dysfunction, as manifested 91%, and drug binding to platelets is extensive.64 VRL is widely dis- by bloating, constipation, ileus, and abdominal pain, occur most tributed, and high concentrations are found in virtually all tissues, commonly with VCR or high doses of the other vinca alkaloids. except the central nervous system.64 The wide distribution of VRL Mucositis occurs more frequently with VBL than with VRL and reflects its lipophilicity, which is among the highest of the vinca is least common with VCR. Nausea, vomiting, diarrhea,31,43,45 and CANCER THERAPEUTICS alkaloids. As with other vinca alkaloids, the liver is the principal pancreatitis53,54 also occur to a lesser extent. excretory organ, and up to 80% of VRL is excreted in the feces, VCR principally induces neurotoxicity characterized by a pe- whereas urinary excretion represents only 16% to 30% of total drug ripheral, symmetric mixed sensory motor and autonomic polyneu- disposition, the bulk of which is unmetabolized VRL. Studies in ropathy.68,69 Toxic manifestations include constipation, abdominal humans indicate that 4-O-deacetyl-VRL and 3,6-epoxy-VRL are cramps, paralytic ileus, urinary retention, orthostatic hypotension, the principal metabolites, and several minor hydroxy-VRL isomer and hypertension. Its primary neuropathologic effects are due to metabolites have been identified. Although most metabolites are interference with axonal microtubule function. Early symmetric inactive, the deacetyl-VRL metabolite may be as active as VRL. sensory impairment and paresthesias can progress to neuritic pain The cytochrome P-450 CYP3A isoenzyme appears to be princi- and loss of deep tendon reflexes with continued treatment, which pally involved in biotransformation. may be followed by foot drop, wrist drop, motor dysfunction, ataxia, and paralysis. Cranial nerves are rarely affected because Vinflunine the uptake of VCR into the central nervous system is low. Severe neurotoxicity occurs infrequently with VBL and VDS. VRL has been shown to have a lower affinity for axonal microtubules than VFL is a novel semisynthetic microtubule inhibitor with a fluori- either VCR or VBL, which seems to be confirmed by clinical ob- nated catharanthine moiety, which translates into lower affinity for servations.70 Mild-to-moderate peripheral neuropathy, principally the vinca binding site on tubulin and, therefore, different quantita- characterized by sensory effects, occurs in 7% to 31% of patients, tive effects on microtubule dynamics.65 The low affinity for tubulin and constipation and other autonomic effects are noted in 30% of may be responsible for its reduced clinical neurotoxicity. Despite patients, whereas severe toxicity occurs in 2% to 3%. this lower affinity, it is more active in vivo than other vinca alka- In adults, neurotoxicity may occur after treatment with cumu- loids, and resistance develops more slowly. VFL is a new vinca lative doses as little as 5 to 6 mg, and manifestations may be pro- and still under clinical development. Its volume of distribution is found after cumulative doses of 15 to 20 mg. Patients with delayed large, and has a terminal half-life of nearly 40 hours.65 The only biliary excretion or hepatic dysfunction, and those with antecedent active metabolite is 4-O-deacetylvinflunine, which has a terminal neurologic disorders, such as Charcot-Marie-Tooth disease, hered- half-life approximately 5 days longer than that of the parent com- itary and sensory neuropathy type 1, and Guillain-Barré syndrome, pound.65 are predisposed to neurotoxicity. The vinca alkaloids are potent vesicants. To decrease the risk of Drug Interactions phlebitis, the vein should be adequately flushed after treatment. If extravasation is suspected, treatment should be discontinued, as- Methotrexate accumulation in tumor cells is enhanced in vitro piration of any residual drug remaining in the tissues should be by the presence of VCR or VBL, an effect mediated by a vinca attempted, and prompt application of heat (not ice) for 1 hour four alkaloid–induced blockade of drug efflux; however, the minimal times daily for 3 to 5 days can limit tissue damage.71 Hyaluroni- concentrations of VCR required to achieve this effect occur only dase, 150 to 1,500 U (15 U/mL in 6 mL 0.9% sodium chloride transiently in vivo.66 The vinca alkaloids also inhibit the cellular in- solution) subcutaneously, through six clockwise injections in a flux of the epipodophyllotoxins in vitro, resulting in less cytotoxic- circumferential manner using a 25-gauge needle (changing the ity. However, the clinical implications of this potential interaction needle with each new injection) into the surrounding tissues may are unknown. L-asparaginase may reduce the hepatic clearance minimize discomfort and latent cellulitis. A surgical consultation 234 Cancer Therapeutics to consider early debridement is also recommended. Mild and in addition to fatigue, nausea, emesis, and diarrhea.55,74 It also has reversible alopecia occurs in approximately 10% and 20% of pa- been evaluated in other solid tumors such as ovarian, prostate, and tients treated with VLR and VCR, respectively. Acute cardiac is- renal cell carcinomas.75 Epothilones are still undergoing evalua- chemia, chest pains without evidence of ischemia, fever, Raynaud tions in several clinical trials. Pharmacokinetic studies based on syndrome, hand–foot syndrome, and pulmonary and liver toxicity patupilone have shown large volume of distribution (41-fold the (transaminitis and hyperbilirubinemia) have also been reported total body water) and low body clearance (13% of hepatic blood with use of the vinca alkaloids. All of the vinca alkaloids can cause flow).76 There do not appear to be active metabolites once the par- a syndrome of inappropriate secretion of antidiuretic hormone ent drug is hydrolyzed, which is the main elimination pathway.76 (SIADH), and patients who are receiving intensive hydration are particularly prone to severe hyponatremia secondary to SIADH. Maytansinoids and Auristatins: DM1, MMAE

MICROTUBULE ANTAGONISTS Antibody drug conjugates (ADC) were first attempted with delivery of doxorubicin. Although tissue localization seemed promising, it became clear that the delivery of more potent chemotherapeutics was necessary.77,78 One of the major advances for the promise of ADC came with the discovery and development of highly potent Estramustine is a conjugate of nor- linked to anticancer compounds such as calicheamicins, maytansinoids, 17β-estradiol by a carbamate ester bridge. Estramustine phosphate and auristatins.78 The next necessary advance was a linker that received regulatory approval in the United States in 1981 for treat- released the drug only when intended, and avoiding, or in some ing patients with castration-resistant prostate cancer (CRPC). Al- cases capitalizing on, in vivo proteases, oxidizing, or reducing en- though the recommended daily dose of estramustine phosphate is vironments. Gemtuzumab ozogamicin was the first ADC using 14 mg/kg per day, patients are usually treated in the daily dosing calicheamicin, a potent DNA minor groove binder (and not a mi- range of 10 to 16 mg/kg in three to four divided daily doses (see crotubule agent), approved in 2000 although withdrawn from the Table 21.1). Estramustine has significant activity in CRPC and market in 2013 due to failed confirmatory studies. Maytansinoids had been used in combination with VBL or docetaxel. However, and auristatins are unrelated, although are both tubulin-binding phase III trials in patients with CRPC showed that when com- agents of the vinca binding site and inhibit tubulin polymeriza- bined with docetaxel, there is no added benefit to overall survival tion.78 They are 100- to 1,000-fold more cytotoxic that most cancer compared to docetaxel alone.72,73 chemotherapeutics.79 Estramustine binds to β-tubulin at a site distinct from the col- Drug maytansinoid-1 (DM1) is the chemotherapeutic de- chicine and vinca alkaloid binding sites. This agent depolymerizes livered using a thioether linker in the ADC ado-trastuzumab microtubules and microfilaments, binds to and disrupts MAPs, and emtansine (T-DM1) that was FDA approved for patients with inhibits cell growth at high concentrations, resulting in mitotic ar- HER2- positive metastatic breast cancer previously treated with rest and apoptosis in tumor cells. The selective accumulation and trastuzumab and taxane chemotherapy.80,81 In the international actions of estramustine phosphate and its metabolite, estromus- phase III study, there was a 3.2-month improved progression-free tine, in specific tissues appear to be dependent on the expression survival among patients that received T-DM1 compared to those of the estramustine-binding protein (EMBP). The disposition of receiving standard treatment with and lapatinib.81 estramustine is principally by rapid oxidative metabolism of the Despite a potent chemotherapeutic, the tolerability was much bet- parent compound to estromustine. Estromustine concentrations in ter in the experimental arm, which was dosed at 3.6 mg/kg intra- plasma are maximal within 2 to 4 hours after oral administration, venously every 21 days. The most common side effects in the trial and the mean elimination half-life of estromustine is 14 hours. Es- were thrombocytopenia (12.8%), transient transaminitis (4.3%), as tromustine and estramustine are principally excreted in the feces, well as nausea, fatigue, myalgias, and arthralgias.81 with only small amounts of conjugated estrone and estradiol de- Monomethyl auristatin E (MMAE) is linked to a monoclonal tected in the urine (less than 1%). antibody against CD30 as an ADC (brentuximab vedotin, SGN35) In general, this agent has a manageable safety profile. Nausea and approved for refractory Hodgkin lymphoma or anaplastic large and vomiting are the principal toxicities encountered. In contrast cell lymphoma. The linker is a peptide-based substrate for cathep- to the taxanes and the vinca alkaloids, myelosuppression is rarely sin-B and thereby designed to detect the lysosome/endosome clinically relevant. Common estrogenic side effects include gy- compartment for drug release.82,83 Dose-limiting toxicities include necomastia, nipple tenderness, and fluid retention. Thromboem- thrombocytopenia, hyperglycemia, diarrhea, and vomiting, and bolic complications may occur in up to 10% of patients. the most common side effects in this heavily pretreated population (including autologous stem cell transplant) includes peripheral 84 Epothilones neuropathy (42%), nausea (35%), and fatigue (34%). The FDA black box warning includes contraindicated use with bleomycin due to increased pulmonary toxicity and the risk of John Cunning- The epothilones are macrolide compounds that were initially iso- ham (JC) virus–induced progressive multifocal leukoencephalop- lated from the mycobacterium Sorangium cellulosum. They exert athy. Reports of severe pancreatitis are also emerging.85 their cytotoxic effects by promoting tubulin polymerization and inducing mitotic arrest.74 In general, the epothilones are more po- tent than the taxanes. In contrast to the taxanes and vinca alkaloids, MITOTIC MOTOR PROTEIN INHIBITORS overexpression of the efflux protein P-glycoprotein minimally af- fects the cytotoxicity of epothilones. Epothilones include the natu- ral epothilone B (patupilone; EPO906) and several semisynthetic Aurora Kinase and Pololike Kinase Inhibitors epothilone compounds such as aza-epothilone B (ixabepilone; BMS-247550), epothilone D (deoxyepothilone B, KOS-862), and Aurora kinases are serine/threonine kinases crucial for mitosis in a fully synthetic analog, sagopilone (ZK-EPO).75 their recruitment of mitotic motor proteins for spindle formation. Ixabepilone has been evaluated in several schedules using a They are particularly overexpressed in high growth rate tumors. cremophor-based formulation and is FDA approved for the treat- Aurora A and B kinases are expressed globally throughout all tis- ment of patients with breast cancer.75 It is active in breast cancer sues, and Aurora C kinase is expressed in testes and participates in previously treated with paclitaxel or docetaxel. The principal tox- meiosis. Aurora A kinase is expressed and frequently amplified in icities observed include neutropenia and peripheral neuropathy, many epithelial tumors and implicated in the microtubule-targeted

tahir99 - UnitedVRG Chapter 21 Antimicrotubule Agents 235 agent-resistant phenotype.86 Aurora A kinase interacts with p53, antimicrotubule drugs and may confer drug resistance.93,97 Inher- and there is evidence that p53 wild-type tumors are more sensitive ent differences in microtubule dynamics and drug interactions to aurora A kinase inhibitors than p53 mutant tumors.87 MLN-8237 have been observed with some isotypes in vitro and in vivo.98 Sev- has an IC50 of 1 nm for aurora A kinase and >200 nm for aurora B eral taxane-resistant mutant cell lines that have structurally altered kinase and is in clinical development for treatment-related neuro- α- and β-tubulin proteins and an impaired ability to polymerize endocrine prostate cancer.86,88 The main dose-limiting toxicity of into microtubules have also been identified.99 Mutations of tubu- these agents is neutropenia. Pololike kinases (PLKs) are serine or lin isotype genes, gene amplifications, and isotype switching have threonine kinases crucial for cell cycle process. Overexpression of also been reported in taxane-resistant cell lines.99 In patients, levels PLKs has been shown to be related to histologic grading and poor of class III β-tubulin have been shown to correlate with response— prognosis in several types of cancer. BI-2536 and ON01910 are those with high RNA levels have poor response—and immunohis- PLK inhibitors in early clinical development.89 tochemical stains can correlate and may be predictive.96,100,101 As opposed to taxanes, resistance to vinca alkaloids has been associ- Kinesin Spindle Protein Inhibitor ated with decreased class II β-tubulin expression.97,98 MAPs are important structural and regulatory components Ispinesib of microtubules that act in concert to remodel the microtubule network by stabilizing or destabilizing microtubules during mi- Kinesin spindle protein (KSP; also known as EG5) is a kinesin tosis or cytokinesis. Alterations in the activity and/or balance of motor protein required to establish mitotic-spindle bipolarity.90 stabilizing or destabilizing MAPs can profoundly affect microtu- Several KSP inhibitors have been evaluated in early phase clini- bule function.99,102 The overexpression of stathmin, a destabiliz- cal trials. SB-715992 (ispinesib) is a small-molecule inhibitor of ing protein, has been reported to decrease sensitivity to paclitaxel KSP ATPase and has been evaluated in two different schedules.89 and vinblastine.1 An analysis of predictive or prognostic factors in a The dose-limiting toxicity is neutropenia. Ispinesib was found to large phase 3 study (National Surgical Adjuvant Breast and Bowel be inactive in phase 2 studies evaluating efficacy in patients with Project NSABP-B 28) in patients with node-positive breast can- castration-resistant and largely docetaxel-resistant prostate cancer, cer showed that MAP-tau, a stabilizing protein, was a prognostic advanced renal cancer, and head and neck cancer.90–92 factor; however, it was not predictive for benefit from paclitaxel- based chemotherapy.1,93 In a separate randomized controlled trial in breast cancer (TAX 307), where the only variable was docetaxel, MECHANISMS OF RESISTANCE TO MAP-tau was also shown to be prognostic, but not predictive of MICROTUBULE INHIBITORS taxane benefit.103 Additional studies have shown a correlation with BRCA1 loss Drug resistance is often complex and multifaceted and can involve measured by gene or protein expression, or gene signatures, with diverse mechanisms such as (1) factors that reduce the ability of resistance to taxane and sensitivity to DNA-damaging agents CANCER THERAPEUTICS drugs to reach their cellular target (e.g., activation of detoxification (such as cisplatin and anthracyclines).104–107 BRCA1 is a tumor- pathways and decreased drug accumulation); (2) modifications in suppressor gene with DNA damage response and repair, as well as the drug target; and (3) events downstream of the target (e.g., de- cell cycle checkpoint activation, which explains why its loss leads creased sensitivity to, or defective, apoptotic signals). Many tubu- to enhanced cisplatin sensitivity.20 BRCA1 also indirectly regulates lin binding agents are substrates for multidrug transporters such microtubule dynamics and stability and can favorably control how as P-glycoprotein and the multidrug resistance gene (MDR1).93,94 microtubules respond to paclitaxel treatment via their association The MDR1-encoded gene product MDR1 (ABC subfamily with pro-caspase-8. The loss of BRCA1 can lead to impaired tax- B1; ABCB1) and MDR2 (ABC subfamily ABCB4) are the best- ane-induced activation of apoptosis due to microtubules that are characterized ABC transporters thought to confer drug resistance more dynamic and less susceptible to taxane-induced stabilization to taxanes.94,95 MDR-related taxane resistance can be reversed by and proximity-induced activation of caspase-8 signaling.20 many classes of drugs, including the calcium channel blockers, In addition to resistance, certain tumor subtypes may be sen- cyclosporin A, and antiarrhythmic agents.94,95 However, the clini- sitive to the taxane dosing schedule. In two randomized trials of cal utility of this approach has never been proven, despite several low-dose, weekly paclitaxel, the luminal breast cancer subtype was clinical trials. The role of ABC transporters in resistance to micro- found to have a better outcome compared with the control arm. tubule inhibitors remains to be determined.96 This suggests that not only the drug, but also the schedule may An increasing number of studies suggest that the expression influence the response to therapy and that genomic approaches of individual tubulin isotypes are altered in cells resistant to may reveal these insights.108

SELECTED REFERENCES

The full reference list can be accessed at lwwhealthlibrary.com/oncology. 13. Schiff PB, Fant J, Horwitz SB. Promotion of microtubule assembly in vitro by taxol. Nature 1979;277:665–667. 1. Kavallaris M. Microtubules and resistance to tubulin-binding agents. Nat 19. Komlodi-Pasztor E, Sackett D, Wilkerson J, et al. Mitosis is not a key tar- Rev Cancer 2010;10:194–204. get of microtubule agents in patient tumors. Nat Rev Clin Oncol 2011;8: 3. Wani MC, Taylor HL, Wall ME, et al. Plant antitumor agents. VI. Isolation 244–250. and structure of taxol, a novel antileukemic and antitumor agent from Taxus 20. Sung M, Giannakakou P. BRCA1 regulates microtubule dynamics and tax- brevifolia. J Am Chem Soc 1971;93:2325–2327. ane-induced apoptotic cell signaling. Oncogene 2014;33(11):1418–1428. 5. Vrignaud P, Sémiond D, Lejeune P, et al. Preclinical antitumor activity of 23. Dai H, Ding H, Meng XW, et al. Contribution of Bcl-2 phosphorylation to cabazitaxel, a semisynthetic taxane active in taxane-resistant tumors. Clin Bak binding and drug resistance. Cancer Res 2013;73(23)6998–7008. Cancer Res 2013;19:2973–2983. 24. Oakes SR, Vaillant F, Lim E, et al. Sensitization of BCL-2–expressing breast 9. Martin M, Pienkowski T, Mackey J, et al. Adjuvant docetaxel for node-posi- tumors to chemotherapy by the BH3 mimetic ABT-737. Proc Natl Acad Sci tive breast cancer. N Engl J Med 2005;352:2302–2313. 2012;109:2766–2771. 10. Jones SE, Erban J, Overmoyer B, et al. Randomized phase III study of 25. Chauhan VP, Stylianopoulos T, Martin JD, et al. Normalization of tumour docetaxel compared with paclitaxel in metastatic breast cancer. J Clin Oncol blood vessels improves the delivery of nanomedicines in a size-dependent 2005;23:5542–5551. manner. Nat Nanotechnol 2012;7:383–388. 11. Tannock IF, de Wit R, Berry WR, et al. Docetaxel plus prednisone or mi- 26. Gradishar W, Tjulandin S, Davidson N, et al. Phase III trial of nanoparticle toxantrone plus prednisone for advanced prostate cancer. N Engl J Med albumin-bound paclitaxel compared with polyethylated castor oil-based pa- 2004;351:1502–1512. clitaxel in women with breast cancer. J Clin Oncol 2005;23:7794–7803. 236 Cancer Therapeutics

27. Socinski MA, Bondarenko I, Karaseva NA, et al. Weekly nab-paclitaxel in 64. Rowinsky EK, Noe DA, Trump DL, et al. Pharmacokinetic, bioavailability, combination with carboplatin versus solvent-based paclitaxel plus carbopla- and feasibility study of oral vinorelbine in patients with solid tumors. J Clin tin as first-line therapy in patients with advanced non–small-cell lung cancer: Oncol 1994;12:1754–1763. final results of a Phase III trial.J Clin Oncol 2012;30:2055–2062. 72. Petrylak D, Hussain MHA, Tangen C, et al. Docetaxel and estramustine 28. Von Hoff DD, Ervin T, Arena FP, et al. Increased survival in pancreatic compared with mitoxantrone and prednisone for advanced refractory pros- cancer with nab-paclitaxel plus gemcitabine. N Engl J Med 2013;369: tate cancer. N Engl J Med 2004;351:1513–1520. 1691–1703. 73. Tannock IF, de Wit R, Berry WR, et al. Docetaxel plus prednisone or mi- 32. Nyman DW, Campbell KJ, Hersh E, et al. Phase I and pharmacokinetics trial toxantrone plus prednisone for advanced prostate cancer. N Engl J Med of ABI-007, a novel nanoparticle formulation of paclitaxel in patients with 2004;351:1502–1512. advanced nonhematologic malignancies. J Clin Oncol 2005;23:7785–7793. 74. Lee JJ, Kelly WK. Epothilones: tubulin polymerization as a novel target for 39. Cheng CJ, Saltzman WM. Enhanced siRNA delivery into cells by exploiting prostate cancer therapy. Nat Clin Pract Oncol 2009;6:85–92. the synergy between targeting ligands and cell-penetrating peptides. Bioma- 76. Kelly K, Zollinger M, Lozac’h F, et al. Metabolism of patupilone in patients terials 2011;32:6194–6203. with advanced solid tumor malignancies. Invest New Drugs 2013;31:605–615. 40. Davis ME, Zuckerman JE, Choi CHJ, et al. Evidence of RNAi in humans 79. Doronina SO, Toki BE, Torgov MY, et al. Development of potent mono- from systemically administered siRNA via targeted nanoparticles. Nature clonal antibody auristatin conjugates for cancer therapy. Nat Biotechnol 2010;464:1067–1070. 2003;21:778–784. 43. Diéras V, Lortholary A, Laurence V, et al. Cabazitaxel in patients with ad- 80. Lewis Phillips GD, Li G, Dugger DL, et al. Targeting HER2-positive breast vanced solid tumours: results of a Phase I and pharmacokinetic study. Eur J cancer with trastuzumab-DM1, an antibody–cytotoxic drug conjugate. Can- Cancer 2013;49:25–34. cer Res 2008;68:9280–9290. 44. Mita AC, Denis LJ, Rowinsky EK, et al. Phase I and pharmacokinetic study 81. Verma S, Miles D, Gianni L, et al. Trastuzumab emtansine for HER2-posi- of XRP6258 (RPR 116258A), a novel taxane, administered as a 1-hour infu- tive advanced breast cancer. N Engl J Med 2012;367:1783–1791. sion every 3 weeks in patients with advanced solid tumors. Clin Cancer Res 83. Younes A, Bartlett NL, Leonard JP, et al. Brentuximab vedotin (SGN-35) for 2009;15:723–730. relapsed CD30-positive lymphomas. N Engl J Med 2010;363:1812–1821. 45. Yared JA, Tkaczuk KH. Update on taxane development: new analogs and 86. Mosquera JM, Beltran H, Park K, et al. Concurrent AURKA and MYCN new formulations. Drug Des Devel Ther 2012;6:371–384. gene amplifications are harbingers of lethal treatment-related neuroendo- 48. Kudlowitz D, Muggia F. Defining risks of taxane neuropathy: insights from crine prostate cancer. Neoplasia 2013;15:1–10. randomized clinical trials. Clin Cancer Res 2013;19:4570–4577. 94. Gottesman MM, Fojo T, Bates SE. Multidrug resistance in cancer: role of 49. Henningsson A, Karlsson MO, Viganò L, et al. Mechanism-based pharmaco- ATP-dependent transporters. Nat Rev Cancer 2002;2:48–58. kinetic model for paclitaxel. J Clin Oncol 2001;19:4065–4073. 99. Orr GA, Verdier-Pinard P, McDaid H, et al. Mechanisms of Taxol resistance 50. Sissung T, Mross K, Steinberg S, et al. Association of ABCB1 genotypes with related to microtubules. Oncogene 2003;22:7280–7295. paclitaxel-mediated peripheral neuropathy and neutropenia. Eur J Cancer 100. Monzó M, Rosell R, Sánchez JJ, et al. Paclitaxel resistance in non-small- 2006;42:2893–2896. cell lung cancer associated with beta-tubulin gene mutations. J Clin Oncol 53. Rugo H, Barry W, Moreno Aspitia A, et al. CALGB 40502/NCCTG N063H: 1999;17:1786–1793. Randomized phase III trial of weekly paclitaxel (P) compared to weekly 102. Baquero MT, Hanna JA, Neumeister V, et al. Stathmin expression and its nanoparticle albumin bound nab-paclitaxel (NP) or ixabepilone (Ix) with or relationship to microtubule-associated protein tau and outcome in breast without bevacizumab (B) as first-line therapy for locally recurrent or meta- cancer. Cancer 2012;118:4660–4669. static breast cancer (MBC). J Clin Oncol 2012;30. 103. Baquero MT, Lostritto K, Gustavson MD, et al. Evaluation of prognostic and 56. Alsamarai S, Charpidou AG, Matthay RA, et al. Pneumonitis related to predictive value of microtubule associated protein tau in two independent docetaxel: case report and review of the literature. In Vivo 2009;23:635–637. cohorts. Breast Cancer Res 2011;13(5):R85. 58. de Bono JS, Oudard S, Ozguroglu M, et al. Prednisone plus cabazitaxel 106. Font A, Taron M, Gago JL, et al. BRCA1 mRNA expression and outcome or mitoxantrone for metastatic castration-resistant prostate cancer pro- to neoadjuvant cisplatin-based chemotherapy in bladder cancer. Ann Oncol gressing after docetaxel treatment: a randomised open-label trial. Lancet 2011;22:139–144. 2010;376:1147–1154. 108. Martin M, Prat A, Rodriguez-Lescure A, et al. PAM50 proliferation score as 62. Bender RA, Castle MC, Margileth DA, et al. The pharmacokinetics of a predictor of weekly paclitaxel benefit in breast cancer. Breast Cancer Res [3H]-vincristine in man. Clin Pharmacol Ther 1977;22:430–435. Treat 2013;138:457–466.

tahir99 - UnitedVRG Kinase Inhibitors as 22 Anticancer Drugs

Charles L. Sawyers

INTRODUCTION EARLY SUCCESSES: TARGETING CANCERS WITH WELL-KNOWN KINASE MUTATIONS In 2001, the first tyrosine-kinase inhibitor imatinib was approved (BCR-ABL, KIT, HER2) for clinical use in chronic myeloid leukemia. The spectacular suc- cess of this first-in-class agent ushered in a transformation in can- From the beginning, clinical trials of imatinib were restricted to cer drug discovery from efforts that were largely based on novel patients with Philadelphia chromosome–positive chronic myeloid cytotoxic chemotherapy agents to an almost exclusive focus on leukemia. For what seem like obvious reasons, there was never any molecularly targeted agents across the pharmaceutical and bio- serious discussion about treating patients with Philadelphia chro- technology industry and academia. This chapter summarizes this ∼ mosome–negative leukemia because the assumption was that only remarkable progress in this field over 15 years, with the focus patients with the BCR-ABL fusion gene would have a chance of on the concepts underlying this paradigm shift as well as the con- responding. This was clearly a wise decision because hematologic siderable challenges that remain (Table 22.1). Readers in search response rates approached 90% and cytogenetic remissions were of more specific details on individual drugs and their indications seen in nearly half of the patients in the early phase studies.3 It was should consult the relevant disease-specific chapters elsewhere in obvious that the drug worked, and imatinib was approved in record this volume as well as references cited within this chapter. Readers time. Unwittingly, the power of genome-based patient selection was should also note that the epidermal growth factor receptor (EGFR) demonstrated in the clinical development of the very first kinase and human epidermal growth factor receptor 2 (HER2) receptor inhibitor. As we will see, it took nearly a decade for this lesson to be tyrosine kinases covered here have also been successfully targeted fully learned. Today, the much larger clinical experience, with an CANCER THERAPEUTICS by monoclonal antibodies that engage these proteins at the cell array of different kinase inhibitors across many tumor types, has led surface. These drugs, referred to as biologics rather than small to a much better understanding of the principles that dictate onco- molecule inhibitors, are covered in other chapters. The chapter gene addiction that, in retrospect, were staring us in the face. Fore- is organized around kinase targets rather than diseases and, inten- most among them is the notion that tumors with a somatic mutation tionally, has a historical flow to make certain thematic points and or amplification of a kinase drug target are much more likely to be to illustrate the broad lessons that have been and continue to be dependent on that target for survival. Hence, a patient whose tumor learned through the clinical development of these exciting agents. has such a mutation is much more likely to respond to treatment Perhaps the most stunning discovery from the clinical trials of with the appropriate inhibitor. This has also led to a new paradigm the Abelson murine leukemia (ABL) kinase inhibitor imatinib was at the regulatory level of drug approval requiring codevelopment the recognition that tumor cells acquire exquisite dependence on of a companion diagnostic (a molecularly based diagnostic test that the breakpoint cluster region protein BCR-ABL fusion oncogene, 1 reliably identifies patients with the mutation) with the new drug. created by the Philadelphia chromosome translocation. Although After chronic myeloid leukemia, the next example to illustrate this may seem intuitive at first glance, consider the fact that the this principle was gastrointestinal stromal tumor (GIST), which is translocation arises in an otherwise normal hematopoietic stem associated with mutations in the KIT tyrosine-kinase receptor or, cell, the survival of which is regulated by a complex array of growth more rarely, in the platelet-derived growth factor (PDGF) recep- factors and interactions with the bone marrow microenvironment. tor.4,5 Serendipitously, imatinib inhibits both KIT and the PDGF Although BCR-ABL clearly gives this cell a growth advantage that, receptor; therefore, the clinical test of KIT inhibition in GIST fol- over years, results in the clinical phenotype of chronic myeloid lowed quickly on the heels of the success in CML.6 In retrospect, leukemia, there was no reason to expect that these cells would the rapid progress made in these two diseases was based, in part, on depend on BCR-ABL for their survival when confronted with an the fact that the driver molecular lesion (BCR-ABL or KIT muta- inhibitor. In the absence of BCR-ABL, these tumor cells could tion, respectively) is present in nearly all patients who are diagnosed presumably rely on the marrow microenvironment, just like their with these two diseases. The molecular analysis merely confirmed normal, nontransformed neighbors. Thus, it seemed more likely the diagnosis that was made using standard clinical and histologic that, by shutting down the driver oncogene, BCR-ABL inhibitors criteria. Consequently, clinicians could identify the patients most might halt the progression of chronic myeloid leukemia but not likely to respond based on clinical criteria rather than rely on an eliminate the preexisting tumor cells. In fact, chronic myeloid leu- elaborate molecular profiling infrastructure to prescreen patients. kemia (CML) progenitors are eliminated after just a few months Consequently, clinical trials evaluating kinase inhibitors in CML of anti–BCR-ABL therapy, indicating they are dependent on the and GIST accrued quickly, and the therapeutic benefit became driver oncogene for their survival and have “forgotten” how to re- clear almost immediately. turn to normal. This phenomenon, subsequently documented in The notion that molecular alteration of a driver kinase deter- a variety of human malignancies, is colloquially termed oncogene 2 mines sensitivity to a cognate kinase inhibitor was further validated addiction. Although the molecular basis for this addiction still re- during the development of the dual EGFR/HER2 kinase inhibitor mains to be defined, the notion of finding an Achilles’ heel for lapatinib. Clinical trials of this kinase inhibitor were conducted in each cancer continues to captivate the cancer research community women with advanced HER2-positive breast cancer based on ear- and has spawned a broad array of efforts to elucidate the molecular lier success in these same patients with the monoclonal antibody identity of these targets and discover relevant inhibitors. trastuzumab, which targets the extracellular domain of the HER2 237 238 Cancer Therapeutics

TABLE 22.1 Kinase Inhibitors: Approved or Anticipated Approval In 2014

Target Drug Approved Indications Anticipated Future Indications ALK Crizotinib ALK mutant lung cancer ALK mutant neuroblastoma, Ceritinib anaplastic lymphoma BCR-ABL Imatinib Chronic myeloid leukemia Dasatinib Philadelphia chromosome–positive acute lymphoid leukemia Nilotinib T315 mutation only (ponatinib) Bosutinib Ponatinib BRAF Vemurafenib BRAF mutant melanoma Other BRAF mutant tumors Dabrafenib BTK Ibrutinib Chronic lymphocytic leukemia Mantle cell lymphoma EGFR Gefitinib Lung adenocarcinoma with EGFR mutation Erlotinib Afatinib HER2 Lapatinib Her2+ breast cancer JAK2 Ruxolitinib JAK2 mutant myelofibrosis KIT Imatinib Gastrointestinal stromal tumor Sunitinib MEK Trametinib BRAF mutant melanoma PI3K deltaa Chronic lymphocytic leukemia Indolent non-Hodgkin lymphoma PDGFR- α/β Imatinib Chronic myelomonocytic leukemia (with TEL-PDGFR-β fusion) hypereosinophilic syndrome (with PDGFR-β fusion) Dermatofibrosarcoma protuberans RET Vandetanib Medullary thyroid cancer Sorafenib Cabozantinib TORC1 Sirolimus (rapamycin) Kidney cancer (mTOR) Everolimus Breast cancer Temsirolimus Tuberous sclerosis VEGF Sorafenib Kidney cancer Receptor Sunitinib Hepatocellular carcinoma (sorafenib only) Axitinib Pancreatic neuroendocrine tumors (sunitinib) Pazopanib a Approval is anticipated based on positive phase 3 data and announcement of accepted Food and Drug Administration submission by the sponsor. kinase. Lapatinib was initially approved in combination with the of such a focused clinical development plan. Although consider- cytotoxic agent capecitabine for women with resistance to trastu- able preclinical data implicated EGFR as a cancer drug target, zumab,7 and then was subsequently approved for frontline use in there was little insight into which patients were most likely to metastatic breast cancer in combination with chemotherapy or benefit. The first clue that EGFR inhibitors would have a role in hormonal therapy, depending on estrogen receptor status. A key lung cancer came from the recognition by several astute clinicians ingredient that enabled the clinical development of lapatinib was of remarkable responses in a small fraction of patients with lung the routine use of HER2 gene amplification testing in the diagnosis adenocarcinoma.8 Further studies revealed the curious clinical of breast cancer, pioneered during the development of trastuzumab circumstance that those patients most likely to benefit tended to several years earlier. This widespread clinical practice allowed for be those who never smoked, women, and those of Asian ethnic- the rapid identification of those patients most likely to benefit. If ity.9 Clearly, there was a strong clinical signal in a subgroup of lapatinib trials had been conducted in unselected patients, the clin- patients, who could perhaps be enriched based on these clinical ical signal in breast cancer would likely have been missed. features, but it seemed that a unifying molecular lesion must be present. Three academic groups simultaneously converged on the answer. Mutations in the EGFR gene were detected in the 10% to The Serendipity of Unexpected Clinical 15% of patients with lung adenocarcinoma who had radiographic Responses: EGFR in Lung Cancer responses.10–12 It may seem surprising that mutations in a gene as highly visible as EGFR and in such a prevalent cancer had not In contrast to the logical development of imatinib and lapatinib in been detected earlier. But the motivation to search aggressively molecularly defined patient populations, the EGFR kinase inhibi- for EGFR mutations was not there until the clinical responses tors gefitinib and erlotinib entered the clinic without the benefit were seen. Perhaps even more surprising was the failure of the

tahir99 - UnitedVRG Chapter 22 Kinase Inhibitors as Anticancer Drugs 239

pharmaceutical company sponsors of the two most advanced com- analysis of tumors from patients treated on the BR.21 trial, which pounds, gefitinib and erlotinib, to embrace this important discov- concluded that EGFR mutations did not predict for a survival ad- ery and refocus future clinical development plans on patients with vantage.19 (EGFR gene amplificationwas associated with survival, EGFR mutant lung adenocarcinoma. but only in a univariate analysis.) This conclusion was concerning But that was 2004, when the prevailing approach to cancer drug because less than 30% of patients on the trial had tissue available development was an empiric one originally developed (with great for EGFR mutation analysis, raising questions about the adequacy success) for cytotoxic agents. Typically, small numbers of patients of the sample size. Furthermore, the EGFR mutation assay used with different cancers were treated in all comer phase I studies (no by the authors was subsequently criticized because a significant enrichment for subgroups) with the goal of eliciting a clinical sig- number of the EGFR mutations reported in these patients were nal in at least one tumor type. A single-agent response rate of 20% in residues not previously found by others, who had sequenced to 30% in a disease-specific phase II trial would justify a random- thousands of tumors. Many of these mutations were suspected to ized phase III registration trial, where the typical endpoint for drug be an artifact of working from formalin-fixed biopsies. Fortunately, approval is time to progression or survival. Cytotoxics were also typ- recent advances in DNA mutation detection, using massively par- ically evaluated in combination with existing standard of care treat- allel next-generation sequencing technology, have largely elimi- ment (typically approved chemotherapy agents) with the goal of nated this concern. These new platforms are now being used in increasing the response rate or enhancing the duration of response. the clinical setting. (Note: The use of the past tense here is intentional. As we will see Clinical investigators in Asia, where a greater fraction of lung later in this chapter, nearly all cancer drug development today is cancers (roughly 30%) are positive for EGFR mutations, addressed based on selecting patients with a certain molecular profile.) the question of whether mutations predict for clinical benefit in a The clinical development of gefitinib and erlotinib followed prospective trial. In this study known as IPASS, gefitinib was clearly the cytotoxic model. Both drugs had similarly low but convinc- superior to standard doublet chemotherapy as frontline therapy for ing single-agent response rates (10% to 15%) in chemotherapy- patients with advanced EGFR mutation–positive lung adenocar- refractory, advanced lung cancer. Indeed, gefitinib was originally cinoma.20 Conversely, EGFR mutation–negative patients fared granted accelerated approval by the U.S. Food and Drug Adminis- much worse with gefitinib and benefited from chemotherapy. In tration (FDA) in 2003 based on the impressive nature of these re- addition, EGFR mutation–positive patients had a more favorable sponses, contingent on the completion of formal phase III studies overall prognosis regardless of treatment, indicating that EGFR with survival endpoints.13 The sponsors of both drugs, therefore, mutation is also a prognostic biomarker. The IPASS trial serves as conducted phase III registration studies in patients with chemo- a compelling example of a properly designed (and executed) bio- therapy-refractory, advanced stage lung cancer but without pre- marker-driven clinical trial. Although the rationale for this clinical screening patients for EGFR mutation status. (In fairness, these development strategy had been demonstrated years earlier with trials were initiated prior to the discovery of EGFR mutations in BCR-ABL in leukemia, KIT in GIST, and HER2 in breast cancer, lung cancer but study amendments could have been considered.) it was difficult to derail the empiric approach that had been used CANCER THERAPEUTICS Erlotinib was approved in 2004 on the basis of a modest survival for decades in developing cytotoxic agents. advantage over placebo (the BR.21 trial); however, gefitinib failed to demonstrate a survival advantage in essentially the same patient population.14,15 This difference in outcome was surprising because A Mix of Science and Serendipity: PDGF the two drugs have highly similar chemical structures and biologic Receptor–Driven Leukemias and Sarcoma properties. Perhaps the most important difference was drug dose. Erlotinib was given at the maximum tolerated dose, which pro- The discovery of EGFR mutations in lung cancer (motivated by duces a high frequency of rash and diarrhea. Both side effects are dramatic clinical responses in a subset of patients treated with presumed on target consequences of EGFR inhibition because EGFR kinase inhibitors) is the most visible example of the power EGFR is highly expressed in skin and gastrointestinal epithelial of bedside-to-bench science, but it is not the only (or the first) such cells. In contrast, gefitinib was dosed slightly lower to mitigate example from the kinase inhibitor era. Shortly after the approval of these toxicities, with the rationale that responses were clearly docu- imatinib for CML in 2001, two case reports documented dramatic mented at lower doses. remissions in patients with hypereosinophilic syndrome (HES), a In parallel with the single-agent phase III trials in chemother- blood disorder characterized by prolonged elevation of eosinophil apy-refractory patients, both gefitinib and erlotinib were studied counts and subsequent organ dysfunction from eosinophil infil- as an upfront therapy for advanced lung cancer to determine if tration, when treated with imatinib.21,22 Although HES resembles either would improve the efficacy of standard doublet (carboplatin/ myeloproliferative diseases such as CML, the molecular pathogen- paclitaxel or gemcitabine/cisplatin) chemotherapy when all three esis of HES was completely unknown at the time. Reasoning that drugs were given in combination. These trials, termed INTACT-1 these clinical responses must be explained by inhibition of a driver and INTACT-2 (gefitinib with either gemcitabine/cisplatin or with kinase, a team of laboratory-based physician/scientists quickly carboplatin/paclitaxel) and TRIBUTE (erlotinib with carboplatin/ searched for mutations in the three kinases known to be inhibited paclitaxel), collectively enrolled over 3,000 patients.16–18 Excite- by imatinib (ABL, KIT, and PDGF receptor). ABL and KIT were ment in the oncology community was high based on the clear quickly excluded, but the PDGF receptor α (PDGFR-α) gene was single-agent activity of both EGFR inhibitors. But, both trials were targeted by an interstitial deletion that fused the upstream FIP1L1 spectacular failures; neither drug showed any benefit over chemo- gene to PDGFR-α23 FIP1L1-PDGFR-α is a constitutively active therapy alone. The fact that EGFR mutations are present in only tyrosine kinase, analogous to BCR-ABL, and is also inhibited 10% to 15% of patients (i.e., those likely to benefit) provided a logi- by imatinib. As with EGFR-mutant lung cancer, the molecular cal explanation. The clinical signal from those whose tumors had pathophysiology of HES was discovered by dissecting the mecha- EGFR mutations was likely diluted out by all the patients whose nism of response to the drug used to treat it. tumors had no EGFR alterations, many of whom benefited from The HES/FIP1L1-PDGFR-α story serves as a nice bookend chemotherapy. to an earlier discovery that the t(5,12) chromosome translocation, The convergence of the EGFR mutation discovery with these found rarely in patients with chronic myelomonocytic leukemia, clinical trial results will be remembered as a remarkable time in creates the TEL-PDGFR-β fusion tyrosine kinase.24 Similar to the history of targeted cancer therapies, not just for the important HES, treatment of patients with t(5,12) translocation-positive leu- role of these agents as lung cancer therapies, but also for missteps kemias with imatinib has also proven successful.25 A third example in deciding that the EGFR genotype should drive treatment selec- comes from dermatofibrosarcoma protuberans, a sarcoma charac- tion. Perhaps the most egregious error came from a retrospective terized by a t(17,22) translocation that fuses the COL1A gene to 240 Cancer Therapeutics the PDGFB ligand (not the receptor). COL1A-PDGFB is onco- community can learn from the data. This policy enabled pan can- genic through autocrine stimulation of the normal PDGF receptor cer mutational analyses that give an overall view of the genomic in these tumor cells. Patients with dermatofibrosarcoma protuber- landscape of cancer, serving as a blueprint for the community of ans respond to imatinib therapy because it targets the PDGF re- cancer researchers and drug developers.33,34 ceptor, just one step downstream from the oncogenic lesion.26 Rounding Out the Treatment of Exploiting the New Paradigm: Searching for Myeloproliferative Disorders: JAK2 and Other Kinase-Driven Cancers Myelofibrosis

The benefits of serendipity notwithstanding, the growing number Taken together with the BCR-ABL translocation in CML and of examples of successful kinase inhibitor therapy in tumors with FIP1L1-PDGFR-α in HES, the discovery of JAK2 mutations in a mutation or amplification of the drug target begged for a more polycythemia, essential thrombocytosis, and myelofibrosis pro- rational approach to drug discovery and development. In 2002, the vided a unifying understanding of myeloproliferative disorders as list of human tumors known to have mutations in kinases was quite diseases of abnormal kinase activation. The JAK family kinases are small. Due to advances in automated gene sequencing, it became the primary effectors of signaling through inflammatory cytokine possible to ask whether a much larger fraction of human cancers receptors and, therefore, had been considered compelling targets might also have such mutations through a brute force approach. for anti-inflammatory drugs. But the JAK2 mutation discovery im- To address this question comprehensively, one would have to se- mediately shifted these efforts toward developing JAK2 inhibitors quence all of the kinases in the genome in hundreds of samples for myeloproliferative disorders. Because most patients have a com- of each tumor type. Several early pilot studies demonstrated the mon JAK2 V617F mutation, these efforts could rapidly focus on potential of this approach by revealing important new targets for screening for activity against a single genotype. Progress has been drug development. Perhaps the most spectacular was the discovery rapid. Myelofibrosis was selected as the initial indication (instead of mutations in the BRAF kinase in over half of patients with mela- of essential thrombocytosis or polycythemia vera) because the time noma, as well as in a smaller fraction of colon and thyroid can- to registration is expected to be the shortest. Currently, ruxolitinib cers.27 Another was the discovery of mutations in the JAK2 kinase is approved for myelofibrosis based on shrinkage in spleen size as in nearly all patients with polycythemia vera, as well as a significant the primary endpoint. Clinical trials in essential thrombocytosis fraction of patients with myelofibrosis and essential thrombocyto- and polycythemia vera (versus hydroxyurea) are ongoing. Other sis.28–30 A third example was the identification of PIK3CA muta- JAK2 inhibitors are also in clinical development. tions in a variety of tumors, with the greatest frequencies in breast, endometrial, and colorectal cancers.31 PIK3CA encodes a lipid kinase that generates the second messenger phosphatidyl inositol BRAF Mutant Melanoma: Several Missteps 3-phosphate (PIP3). PIP3 activates growth and survival signaling Before Finding the Right Inhibitor through the AKT family of kinases as well as other downstream ef- fectors. Coupled with the well-established role of the phosphatase As with JAK2 mutations in myeloproliferative disorders, the discov- and tensin homolog (PTEN) lipid phosphatase in dephosphorylat- ery of BRAF mutations in patients with melanoma launched wide- ing PIP3, the discovery of PIK3CA mutations focused tremendous spread efforts to find potent BRAF inhibitors. One early candidate attention on developing inhibitors at multiple levels of this path- was the drug sorafenib, which had been optimized during drug way, as discussed further in the follow paragraphs. discovery to inhibit RAF kinases. (Sorafenib also inhibits vascular Each of these important discoveries—BRAF, JAK2, and endothelial growth factor (VEGF) receptors, which led to its ap- PIK3CA—came from relatively small efforts (less than 100 tu- proval in kidney cancer, as discussed later in this chapter.) Despite mors) and generally focused on resequencing only those exons that the compelling molecular rationale for targeting BRAF, clinical coded for regions of kinases where mutations had been found in results of sorafenib in melanoma were extremely disappointing other kinases (typically, the juxtamembrane and kinase domains). and reduced enthusiasm for pursuing BRAF as a drug target.35 These restricted searches were largely driven by the high cost of In hindsight, this concern was completely misguided. Sorafenib DNA sequencing using the Sanger method. In 2006, a compre- dosing is limited by toxicities that preclude achieving serum lev- hensive effort to sequence all of the exons in all kinases in 100 els in patients that potently inhibit RAF, but are sufficient to in- tumors could easily exceed several million dollars. Financial sup- hibit VEGF receptors. In addition, patients were enrolled without port for such projects could not be obtained easily through tra- screening for BRAF mutations in their tumors. Although the fre- ditional funding agencies because the risk/reward was considered quency of BRAF mutations in melanoma is high, the inclusion of too high. Furthermore, substantial infrastructure for sample acqui- patients without the BRAF mutation diluted the chance of seeing sition, microdissection of the tumors from normal tissue, nucleic any clinical signal. In short, the clinical evaluation of sorafenib in acid preparation, high throughput automated sequencing, and melanoma was poorly designed to test the hypothesis that BRAF computational analysis of the resulting data was essential. Few in- is a therapeutic target. The danger is that negative data from such stitutions were equipped to address these challenges. In response, clinical experiments can slow subsequent progress. It is critical to the National Cancer Institute in the United States (in partnership know the pharmacodynamic properties of the drug and the mo- with the National Human Genome Research Institute) and an lecular phenotype of the patients being studied when interpreting international group known as the International Cancer Genome the results of a negative study. Consortium (ICGC) launched large-scale efforts to sequence The fact that RAF kinases are intermediate components of the the complete genomes of thousands of cancers. In parallel, next- well-characterized RAS/ mitogen-activated protein (MAP) kinase generation sequencing technologies resulted in massive reduc- pathway (transducing signals from RAS to RAF to MEK to ERK) tions in cost, allowing a more comprehensive analysis of much raised the possibility that tumors with BRAF mutations might re- larger numbers of tumors. At the time of this writing, the US effort spond to inhibitors of one of these downstream kinases (Fig. 22.1). (called The Cancer Genome Atlas [TCGA]) had reported data on Preclinical studies revealed that tumor cell lines with BRAF mu- 29 different tumor types (https://tcga-data.nci.nih.gov/tcga/). The tation were exquisitely sensitive to inhibitors of the downstream international consortium has committed to sequencing 25,000 kinase MEK.36 (Sorafenib, in contrast, does not show this profile of tumors representing 50 different cancer subtypes.32 Both groups activity.37 Thus, proper preclinical screening would have revealed have enforced immediate release of all sequence information to the shortcomings of sorafenib as a BRAF inhibitor.) Curiously, cell the research community free of charge so that the entire scientific lines with a mutation or amplification of EGFR or HER2, which

tahir99 - UnitedVRG Chapter 22 Kinase Inhibitors as Anticancer Drugs 241 PDGF HRG EGF FGF

FGFR PDGFR EGFR HER2

GRB2 SOS

RAS NF1

A-RAF B-RAF C-RAF

Figure 22.1 The RAS–RAF–MEK–ERK signaling pathway. The classical mitogen-activated protein kinase (MAPK) pathway is activated in human MEK1/2 tumors by several mechanisms, including the binding of ligand to receptor tyrosine kinases (RTK), the mutational activation of an RTK, by loss of the tumor suppressor NF1, or by mutations in RAS, BRAF, and MEK1. Phosphorylation and, thus, activation of ERK regulates the transcription of target genes that promote cell cycle progression ERK1/2 and tumor survival. The ERK pathway contains a classical feedback loop in which the expression RSK1/2/3 of feedback elements such as SPRY and DUSP family proteins are regulated by the level of ERK activity. Loss of expression of SPRY and DUSP family members due to promoter methylation or deletion is thus permissive for persistently

elevated pathway output. In the case of tumors CANCER THERAPEUTICS with mutant BRAF, pathway output is enhanced Cyclin D by impaired upstream feedback regulation. FGF, SPRY Elk 1 Ets fibroblast growth factor; HRG, heregulin; NF1, Myc Fos DUSP neurofibromatosis 1. (From Bernt KM, Zhu N, Sinha AU, et al. MLL-rearranged leukemia is Negative Regulation Regulation dependent on aberrant H3K79 methylation by feedback of cell cycle of transcription DOT1L. Cancer Cell 2011;20(1):66–78, with permission.) function upstream in the pathway, were insensitive to MEK inhi- of PD325901 was discontinued because of safety concerns about bition. Even tumor lines with RAS mutations were variably sensi- ocular and neurologic toxicity.) Disappointingly, patients receiv- tive. In short, the preclinical data made a strong case that MEK ing AZD6244 had no benefit in progression-free survival when inhibitors should be effective in BRAF mutant melanoma, but not compared to -treated patients, raising further con- in other subtypes. The reason that HER2, EGFR, and RAS mutant cerns about the viability of BRAF as a drug target.41 A closer exami- tumors were not sensitive to MEK inhibitors is explained, at least nation of the data revealed that clinical responses were, indeed, in part, by the existence of negative feedback loops that modulate seen in patients receiving AZD6244. The fact that BRAF mutation the flux of signal transduction through MEK.38 status was not required for study entry likely diminished the clini- In parallel with the generation of these preclinical findings, cal signal in the AZD6244 arm, a lesson learned from the EGFR clinical trials of several MEK inhibitors were initiated. Patients inhibitor trials in lung cancer. Indeed, a different MEK inhibitor, with various cancers were enrolled in the early studies, but there trametinib, received FDA approval in 2013 based on activity in was a strong bias to include melanoma patients. Significant ef- melanoma patients with the BRAF mutation.42 forts were made to demonstrate MEK inhibition in tumor cells All doubts about BRAF as a target vanished in 2009 to 2010 by measuring the phosphorylation status of the direct downstream when dramatic clinical responses were observed with a novel substrate ERK using an immunohistochemical analysis of biop- BRAF inhibitor vemurafenib (PLX4032). Like sorafenib, this com- sies from patients with metastatic disease. Phase I studies of the pound was optimized to inhibit RAF, but with an additional focus two earliest compounds in clinical development (PD325901 and on mutant BRAF. Vemurafenib differs dramatically from sorafenib AZD6244) documented reduced phospho-ERK staining at mul- because it potently inhibits BRAF without the additional broad tiple dose levels in several patients for whom baseline and treat- range of activities that sorafenib has against other kinases like the ment biopsies were obtained.39,40 (In the following, we will learn VEGF receptor.43 The greater selectivity of vemurafenib relative that these pharmacodynamic studies, while well intentioned, were to sorafenib resulted in a much greater tolerability, such that it not quantitative enough to document the magnitude of MEK could be given at high doses while avoiding significant toxicity. inhibition in these patients.) Furthermore, clinical responses The early days of vemurafenib clinical development were plagued were observed in a few patients with BRAF mutant melanoma. by challenges in maximizing the oral bioavailability of the drug.44 Armed with this confidence, a randomized phase II clinical trial Consequently, the initial phase I clinical trial was temporarily of AZD6244 was conducted in advanced melanoma, with the halted to develop a novel formulation (i.e., the coingredients in chemotherapeutic agent temozolomide (which is approved for the drug capsule or tablet that improve solubility and absorption glioblastoma) as the comparator arm. (The clinical development through the gastrointestinal tract). Much higher serum levels were 242 Cancer Therapeutics obtained in patients who received the new vemurafenib formation 2007, a different ALK fusion gene called EML4-ALK was discov- and, shortly thereafter, complete and partial responses were ob- ered in a small fraction of patients with lung adenocarcinoma, served in about 80% of the melanoma patients with B-RAF mutant with an estimated frequency of 1% to 5%.50 This discovery did not tumors. Strikingly, no activity was observed in patients whose tu- immediately capture the attention of drug developers, but several mors were wild type for BRAF.45,46 The data were so compelling academic groups who had already begun testing lung cancer pa- that vemurafenib was immediately advanced to a phase III registra- tients seen at their institutions for EGFR mutations simply added tion trial. Similarly impressive responses in BRAF mutant mela- an EML4-ALK fusion test to the screening panel. Astute clinical noma patients were observed with a second potent RAF inhibitor investigators participating in the phase I trial of crizotinib, which dabrafenib,47 providing further proof that BRAF is a important was designed to include patients with a broad array of advanced cancer target. cancers, were aware of the off-target ALK activity and enrolled sev- The vemurafenib and dabrafenib data also provide insight into eral lung cancer patients with EML4-ALK fusions in the study. why sorafenib and the early MEK inhibitor trials failed to demon- These patients had remarkably dramatic responses.51 This seren- strate activity. One lesson is the critical importance of achieving dipitous finding in a few ALK-positive patients was confirmed in a adequate target inhibition. Clinical responses with vemurafenib larger cohort, resulting in a strongly positive pivotal phase III study were observed only after the drug was reformulated to achieve in ALK-positive lung cancer, just 2 years after the discovery of the substantially higher serum levels. Reductions in phospho-ERK EML4-ALK fusion.52 Crizotinib is also being evaluated in other staining (as documented by immunohistochemistry) were docu- diseases associated with genomic alterations in ALK, including mented in the earlier trials but, in retrospect, the assays were not large-cell anaplastic lymphoma, neuroblastoma,53 and inflamma- sensitive enough to distinguish between modest (∼50%) kinase tory myofibroblastic sarcoma.54 inhibition versus more complete BRAF or MEK inhibition. Ef- ficacy in preclinical models is significantly improved using doses that give >80% inhibition, and the human trial data suggest that Extending the Model to RET Mutations in this degree of pathway blockade is also required for a high clini- Thyroid Cancer: Clinical Responses, But Why? cal response rate.46 Collectively, these experiences illustrate the critical need for quantitative pharmacodynamic assays to measure Subsets of patients with papillary or medullary thyroid cancer have target inhibition early in clinical development. A second lesson is activating mutations or translocations targeting the RET tyrosine- the importance of genotyping all patients for mutation or ampli- kinase receptor, raising the question of whether RET inhibitors fication of the relevant drug target. Not only does this ensure that might have a role in this disease.55 Although no drugs specifically a sufficient number of patients with the biomarker of interest are designed to inhibit RET have entered the clinic, four compounds included in the study, but also that the results provide compelling with off-target activity against RET (vandetanib, sorafenib, mote- evidence early in clinical development in support (or not) of the sanib, and cabozantinib) have all shown single-agent activity in preclinical hypothesis. thyroid cancer studies.56–60 Vandetanib and cabozantinib are cur- rently approved in medullary thyroid cancer based on improved progression-free survival in phase III registration trials.61,62 Because Getting It Right: ALK and Lung Cancer all four compounds also inhibit VEGF receptor, it is unclear whether the clinical benefit observed in these studies is explained The development of the ALK inhibitor crizotinib (PF-02341066) by inhibition of RET, VEGF receptor, or both. Unlike the crizo- illustrates how an unexpected signal obtained in a small number tinib trials in ALK-positive lung cancer, enrollment in these regis- of patients can quickly shift a program in an entirely new direction tration studies was not restricted to patients with RET mutations. with a high probability of success. The key ingredient is this story is In addition to the fact that thyroid cancer patients are not routinely a familiar one—a strong molecular hypothesis backed up by clini- screened for these mutations, the primary reason for including all cal response data in a small number of carefully selected patients. comers in these studies is that clinical responses are observed in Crizotinib emerged from a drug discovery program at Pfizer that a larger fraction of patients than can be accounted for based on was focused on finding inhibitors of the MET receptor tyrosine ki- the suspected frequency of an RET mutation. Responses in pa- nase and entered the clinic with this target as its lead indication.48 tients without RET mutation (if they occur) might be explained As we previously learned with imatinib, essentially all kinase inhib- by mutations in other genes in the RAS-MAP kinase pathway such itors have activity against other targets (so called off-target activi- as BRAF or HRAS, which are found in a substantial fraction of pa- ties), which can sometimes prove to be advantageous. Off-target tients and typically do not overlap with RET alterations.55 Clearly, activities are typically discovered by screening compounds against detailed genotype/response correlations, as demonstrated in lung a large panel of kinases to establish profiles of relative selectivity cancer and melanoma, will clarify the role of these mutations against the intended target. Off-target activity, potency, and phar- in predicting the response to these drugs. Thyroid cancer is also maceutical properties (bioavailability, half-life) are all factors that a compelling indication for the BRAF and MEK inhibitors dis- influence the decision of which compound to advance to clinical cussed previously in melanoma. development. The primary off-target activity of crizotinib is against the ALK tyrosine kinase. ALK was first identified as a candidate driver oncogene in 1994 FLT3 Inhibitors in Acute Myeloid Leukemia: through the cloning of the t(2,5) chromosomal translocation as- Did the Genomics Mislead Us? sociated with anaplastic large cell lymphoma, which creates the nucleophosmin/anaplastic lymphoma kinase (NPM-ALK) fusion Shortly after the success of imatinib, the receptor tyrosine–kinase gene.49 This discovery, together with the demonstration that NPM- FLT3 emerged as a compelling drug candidate based on the pres- ALK causes lymphoma in mice, made a compelling case for ALK ence of activating mutations in about one-third of patients with as a drug target in this disease. But there was limited interest in acute myeloid leukemia.63 Laboratory studies documented that developing ALK inhibitors because this particular lymphoma sub- FLT3 alleles bearing these mutations, which occur as internal type is rare and most commonly found in children. (Companies are tandem duplications (ITD) of the juxtamembrane domain or a generally reluctant to develop drugs solely for pediatric indications point mutation in the kinase domain, function as driver oncogenes because of complexities related to dose selection and additional in mouse models, giving phenotypes analogous to BCR-ABL.64 regulatory guidelines. Efforts to streamline this development pro- As with RET in thyroid cancer, no compounds had been specifi- cess are underway, such as the Creating Hope Act, which provides cally optimized to target FLT3, but several drugs with off-target new incentives for companies to pursue pediatric indications.) In FLT3 activity were redirected to acute myeloid leukemia (AML).

tahir99 - UnitedVRG Chapter 22 Kinase Inhibitors as Anticancer Drugs 243

Disappointingly, the first three of the compounds tested (mi- VEGF receptor) have single-agent clinical activity in clear cell car- dostaurin, lestaurtinib, and sunitinib) showed only marginal cinoma.73–75 The high specificity of bevacizumab for VEGF leaves single-agent activity in relapsed AML patients, even in those with little doubt that the activity of this drug is explained by antiangio- FLT3 mutations.65–67 Despite the strong molecular rationale for genic effects. In contrast, the off-target activities of sorafenib and FLT3 as a driver lesion, questions were raised about the viability of sunitinib include several kinases expressed in kidney tumor cells, FLT3 as a drug target. Pharmacodynamic studies showed evidence stroma, and inflammatory cells (PDGFR, RAF, RET, FLT3, and of FLT3 kinase inhibition in tumor cells, but the magnitude and others). Interestingly, the primary effect of bevacizumab in kid- duration of these effects were difficult to quantify, raising the pos- ney cancer is disease stabilization, whereas sorafenib and sunitinib sibility of inadequate target inhibition.65 Indeed, the dose of all have substantial partial response rates. This raises the question of three compounds was limited by toxicities believed to be indepen- whether the superior antitumor activity of the VEGF receptor ki- dent of FLT3. A more pessimistic interpretation was that FLT3, nase inhibitors is due to the concurrent inhibition of other kinases. although presumably important for the initiation of AML, was no However, partial responses rates with next-generation VEGF re- longer required for tumor maintenance due to the accumulation ceptor inhibitors (axitinib, pazopanib, and tivozanib), all of which of additional driver genomic alterations. If true, even a complete have greater potency and selectivity for the VEGF receptor, are FLT3 blockade with a highly selective inhibitor would be expected similarly high, and reinforce the importance of the VEGF recep- to fail. But this view was not supported by the fact that clinical tor as the critical target in kidney cancer.76–78 Pazopanib is ap- responses were observed in the somewhat analogous situation of proved for advanced kidney cancer, whereas axitinib is approved single-agent ABL kinase inhibitor treatment of CML in blast crisis, as second-line therapy. where BCR-ABL is just one of many additional genomic altera- Two inhibitors of the mammalian target of rapamycin (mTOR) tions that contribute to disease progression, yet complete remis- kinase (temsirolimus and everolimus) are also approved for ad- sions are observed in many patients. vanced renal cell carcinoma.79,80 Both temsirolimus and everoli- Despite this pessimism about FLT3 as a viable drug target, sev- mus are known as rapalogs because both are chemical derivatives of eral drugs are now advancing toward drug registration trials. Mi- the natural product sirolimus (rapamycin). Sirolimus was approved dostaurin, one of the early compounds that showed disappointing more than 10 years ago to prevent graft rejection in transplant re- single-agent activity in relapsed AML, is being evaluated in a ran- cipients based on its immunosuppressive properties against T cells. domized phase III trial in newly diagnosed AML combined with Sirolimus also has potent antiproliferative effects against vascular standard induction chemotherapy. A single-arm phase II study endothelial cells and, on that basis, is incorporated into drug-elut- showed higher and more durable remission rates in FLT3 mutant ing cardiac stents to prevent coronary artery restenosis following patients when compared to historical controls.68 The second com- angioplasty.81 Rapalogs differ from all the other kinase inhibitors pound, quizartinib (AC220), is a next-generation FLT3 inhibitor discussed in this chapter in that they inhibit the kinase through an with greater potency and specificity and with single-agent activity allosteric mechanism rather than by targeting the mTOR kinase in FLT3 mutant relapsed AML—precisely the population where domain. Because rapalogs also inhibit the growth of cancer cell CANCER THERAPEUTICS midostaurin and others failed.69,70 The fact that some responder lines from different tissues of origin, clinical trials were initiated to patients have relapsed with drug-resistant gatekeeper mutations in study their potential role as anticancer agents in a broad range of the FLT3 kinase domain provides formal proof that FLT3 is the tumor types. Based on responses in a few phase I patients with dif- relevant target.71 Assuming these compounds prove successful in ferent tumor types (including kidney cancer), exploratory phase II AML, it will be important to examine their activity in the rare cases studies were conducted in several diseases. Single-agent activity of of pediatric acute lymphoid leukemia associated with FLT3 muta- temsirolimus was observed in a phase II kidney cancer study,82 then tion. Although the jury is still out on FLT3 inhibitors, the failure confirmed in a phase III registration trial.79 The phase III everoli- of early compounds in AML is reminiscent of the failures of early mus trial, which was initiated after temsirolimus, was noteworthy RAF and MEK inhibitors in melanoma. Collectively, these exam- because clinical benefit was demonstrated in patients who had pro- ples emphasize the importance of using optimized compounds to gressed on the VEGF receptor inhibitors sorafenib or sunitinib.80 test a molecularly based hypothesis in patients and to focus enroll- In parallel with the empirical clinical development of rapalogs, ment on those patients with the relevant molecular lesion. various laboratories explored the molecular basis for mTOR de- pendence in cancer cells. mTOR functions at the center of a com- Kidney Cancer: Targeting the Tumor and the plex network that integrates signals from growth factor receptors and nutrient sensors to regulate cell growth and size (Fig. 22.2). It Host With Mammalian Target of Rapamycin does so, in part, by controlling the translation of various mRNAs and VEGF Receptor Inhibitors with complex 5′ untranslated regions into protein. mTOR exists in two distinct complexes known as TOR complex 1 (TORC1) A recurring theme in this chapter is the critical role of driver ki- and TORC2. Rapalogs only inhibit the TORC1 complex, which is nase mutations in guiding the development of kinase inhibitors. largely responsible for downstream phosphorylation of targets such Ironically, several kinase inhibitors have been approved for kidney as S6K1/2 and 4EBP1/2 that regulate protein translation.83 The cancer over the past 5 years in a tumor type with no known ki- TORC2 complex contributes to the activation of AKT by phos- nase mutations. The most common molecular alteration in kid- phorylating the important regulatory serine residue S473 and is ney cancer is a loss of function in the Von Hippel-Lindau (VHL) unaffected by rapalogs. tumor suppressor gene, resulting in the activation of the hypoxia Two hypotheses have emerged to explain the clinical activity inducible factor68 pathway.72 As a consequence of VHL loss, which of rapalogs in kidney cancer. The antiproliferative activity of these normally targets hypoxia-inducible factor (HIF) proteins for prote- compounds against endothelial cells suggests an antiangiogenic asomal degradation, HIF-1α and HIF-2α are constitutively active mechanism, which is consistent with the clinical activity of the transcription factors that function as oncogenes through activa- VEGF receptor inhibitors. But rapalogs also inhibit the growth of tion of an array of downstream target genes. Among these is the kidney cancer cell lines in laboratory models where the effects on angiogenesis factor VEGF, which is secreted by HIF-expressing tumor angiogenesis have been eliminated. Interestingly, mRNAs cells and promotes the development and maintenance of tumor for HIF1/2 are among those whose translation is impaired by ra- neovasculature. HIF-mediated secretion of VEGF by tumor cells palogs, and this effect has been implicated as the primary mecha- likely explains the highly vascular histopathology of clear cell renal nism of rapalog activity in kidney cancer xenograft models.84 As carcinoma. All three currently approved angiogenesis inhibitors with the VEGF receptor inhibitors, a detailed molecular annota- (the monoclonal antibody bevacizumab targeting VEGF and the tion of tumors from responders and nonresponders will shed light kinase inhibitors sorafenib and sunitinib targeting or its receptor on these issues. 244 Cancer Therapeutics

Ligand (insulin, growth factors) RTK

PIP2 PTEN PIP3 PIP3 Ras PH PI3K P P P PI3K PH IRS1 AKT P P P P P PDK1

P mSIN1 mTOR PH Rictor mLST8 AKT P P mTORC2 Feedback inhibition LKB1 P AMPK P TSC2

GSK3 P TSC1 P

p90S6K P Rheb

P PRAS40 PH Rapamycin PDK1 mTOR Raptor mLST8

mTORC1

P P P p70S6K 4E-BP1

Translation

Figure 22.2 Feedback inhibition of the phosphatidylinositol 3-kinase (PI3K) pathway. Activated AKT regulates cellular growth through mammalian target of rapamycin (mTOR), a key player in protein synthesis and translation. mTOR forms part of two distinct complexes known as mTORC1, which contains mTOR, Raptor, mLST8, and PRAS40, and mTORC2, which contains mTOR, Rictor, mLST8, and mSIN1. mTORC1 is sensitive to rapamycin and controls protein synthesis and translation, at least in part, through p70S6K and eukaryotic translation initiation factor 4E–binding protein 1 (4E-BP1). AKT phosphorylates and inhibits tuberous sclerosis complex 2 (TSC2), resulting in increased mTORC1 activity. AKT also phosphorylates PRAS40, thus relieving the PRAS40 inhibitory effect on mTOR and the mTORC1 complex. mTORC2 and 3-phosphoinositide– dependent kinase (PDK1) phosphorylate AKT on Ser473 and Thr308, respectively, rendering it fully active. mTORC1-activated p70S6K can phosphorylate insulin receptor substrate 1 (IRS1), resulting in inhibition of PI3K activity. In addition, PDK1 phosphorylates and activates p70S6K and p90S6K. The latter has been shown to inhibit TSC2 activity through direct phosphorylation. Conversely, LKB1-activated AMP-activated protein kinase (AMPK) and glycogen synthase kinase 3 (GSK3) activate the TSC1/TSC2 complex through direct phosphorylation of TSC2. Thus, signals through PI3K as well as through LKB1 and AMPK converge on mTORC1. Inhibition of mTORC1 can lead to increased insulin receptor– mediated signaling, and inhibition of PDK1 may lead to activation of mTORC1 and may, paradoxically, promote tumor growth. (From Daigle SR, Olhava EJ, Therkelsen CA, et al. Selective killing of mixed lineage leukemia cells by a potent small-molecule DOT1L inhibitor. Cancer Cell 2011;20(1):53–65, with permission.)

tahir99 - UnitedVRG Chapter 22 Kinase Inhibitors as Anticancer Drugs 245

Other Indications for mTOR Inhibitors: Breast the most common abnormalities in cancer. Consequently, Cancer and Tuberous Sclerosis Complex intensive efforts at many pharmaceutical companies have been devoted to the discovery of small-molecule inhibitors targeting Mutant Cancers kinases in the PI3K pathway. Inhibitors of PI3K, AKT, and ATP- competitive (rather than allosteric) inhibitors of mTOR that target Two other indications for mTOR have emerged, both based on both the TORC1 and TORC2 complex are all in clinical develop- fundamental insights from laboratory studies but from quite differ- ment. Phase I clinical trials have, in general, established that the ent angles. Preclinical studies of estrogen receptor (ER) therapy in pathway can be efficiently targeted without serious toxicity other breast cancer suggested that phosphatidylinositol 3-kinase (PI3K) than easily manageable effects on glucose metabolism (which is pathway activation may be a mechanism of resistance and that this anticipated based on the importance of PI3K signaling in insulin resistance could be prevented or overcome by combined treatment signaling). Unfortunately, there has been no evidence to date of with ER-based drugs and rapalogs such as everolimus. Based on dramatic single-agent clinical activity with any of these agents, al- evidence that some women with progressive disease while receiv- though early results with PI3K alpha selective inhibitor BYL719 in ing the aromatase inhibitor letrozole have clinical benefit from the PIK3CA mutant breast cancer appear promising.91 addition of everolimus, randomized trials were initiated compar- However, the first approval of a direct PI3K inhibitor in can- + ing everolimus exemestane to exemestane alone (called BO- cer is likely to come in chronic lymphocytic leukemia and in + LERO-2), or everolimus tamoxifen to tamoxifen alone (called lymphoma, but not on the basis of tumor genomics. Normal and TAMRAD). Both studies demonstrated substantial improvements malignant B cells are dependent on PI3K delta as well as Bruton in time to progression in women with metastatic breast cancer 85,86 tyrosine kinase (BTK) for proliferation and survival, raising the who had already failed one aromatase inhibitor, resulting in possibility that inhibitors of these kinases might be broadly active FDA approval of the everolimus/exemestane combination. Evi- in B-cell malignancies. Concerns about toxicity on normal B cells dence of cross-talk between the PI3K pathway and hormone recep- were alleviated, in part, by the earlier clinical success of the CD20 tor signaling (ER in breast cancer, androgen receptor in prostate antibody rituximab in lymphoma, which also eliminates normal cancer) provides a molecular rationale for the clinical benefit of circulating B cells, but without significant clinical sequelae. The combination therapy and is currently under investigation in meta- 87 first such PI3K delta inhibitor, idelalisib, has shown impressive ac- static prostate cancer. tivity in indolent non-Hodgkin lymphoma as a single agent and Yet another indication for rapalog therapy emerged from the in relapsed chronic lymphocytic leukemia when given in combi- genetics of children with tuberous sclerosis caused by a loss of nation with rituximab. The BTK inhibitor ibrutinib, following a function mutations in tuberous sclerosis complex 1 (TSC1) or similar clinical development path, was recently approved as sec- TSC2, which encode the proteins hamartin and tuberin that func- ond-line therapy for chronic lymphocytic leukemia and for mantle tion in the PI3K signaling pathway just upstream of mTOR. Based cell lymphoma.92,93 on laboratory studies showing that TSC1- or TSC2-deficient cells CANCER THERAPEUTICS are exquisitely sensitive to rapalogs, a clinical trial was conducted in tuberous sclerosis patients with benign subependymal giant-cell astrocytomas (SEGA) that showed tumor shrinkage in 21 of 28 pa- COMBINATIONS OF KINASE INHIBITORS tients.88 This genetic dependence on mTOR in tumors with tuber- TO INDUCT RESPONSE AND PREVENT ous sclerosis complex (TSC) loss has also been observed in bladder RESISTANCE cancer. In a remarkable example of the power of comprehensive DNA sequencing to provide insight into rare clinical phenotypes, Preclinical studies indicate that combinations of kinase inhibitors investigators examined the tumor genome of the single complete are required to realize their full potential as anticancer agents. The responder patient on a phase II trial of everolimus in bladder can- most common rationale is to address the problem of concurrent cer and discovered somatic mutations in TSC2 as well as a second mutations in different pathways that alleviate dependence on a gene, NF2, that also controls mTOR activity.89 This plus other ex- single-driver oncogene. The best examples are cancers with muta- amples of how a retrospective genomic analysis of extraordinary tions in both the RAS/MAP kinase pathway (RAS or BRAF) and responders has led to a national effort to capture these cases, as well the PI3K pathway (PIK3CA or PTEN). In mouse models, such as prospective clinical trials of patients with the relevant tumor doubly mutant tumors fail to respond to single-agent treatment genotype regardless of histology (called basket trials). with either an AKT inhibitor or a MEK inhibitor. However, com- It is unclear why rapalogs have failed in other tumor types. bination treatment can give dramatic regressions.94 Similarly, ge- One explanation is the concurrence of PI3K pathway mutations netically engineered mice that develop KRAS-driven lung cancer with alterations in other pathways that mitigate sensitivity to ra- respond only to combination therapy with a PI3K inhibitor and a palogs. Another possibility is the disruption of negative feedback MEK inhibitor.95 To date, clinical trials combining different PI3K loops regulated by mTOR that inhibit signaling from upstream pathway and RAS/MAP kinase pathway inhibitors have been chal- receptor tyrosine kinases. Rapalogs paradoxically increase signal- lenging due to toxicities associated with continuous, concurrent ing through PI3K due to loss of this negative feedback. A primary PI3K and RAS/MAP kinase pathway inhibition. consequence is increased AKT activation, which signals to an array Many of the tumor types discussed in this chapter do respond to of downstream substrates that can enhance cell proliferation and treatment with a single-agent kinase, but relapse despite continued survival (other than TORC1, which remains inhibited by rapalog) inhibitor therapy. Research into the causes of “acquired” kinase in- (see Fig. 22.2). This problem might be overcome by combining hibitor resistance has revealed two primary mechanisms: (1) novel rapalogs with an inhibitor of an upstream kinase in the feedback mutations in the kinase domain of the drug target that preclude loop, such as HER kinases or the insulinlike growth factor recep- inhibition, or (2) bypass of the driver kinase signal by activation of tor (IGFR), to block this undesired effect of rapalogs on PI3K a parallel kinase pathway. In both cases, the solution is combina- activation.90 tion therapy to prevent the emergence of resistance. An elegant demonstration of this approach comes from CML where resis- tance to imatinib is primarily caused by mutations in the BCR- DIRECTLY TARGETING THE PI3K PATHWAY ABL kinase domain.96,97 The second-generation ABL inhibitors dasatinib and nilotinib are effective against most imatinib-resistant Mutations or copy number alterations (e.g., amplification or dele- BCR-ABL mutants and were initially approved as single-agent tion of oncogenes or tumor suppressor genes) in PI3K pathway therapy for imatinib-resistant CML.98,99 Very recently, both drugs genes (PIK3CA, PIK3R1, PTEN, AKT1, and others) are among have proven superior to imatinib in the upfront treatment of CML 246 Cancer Therapeutics due to increased potency and fewer mechanisms of acquired a treatment decision supported by clinical trial data) remains low, resistance.100–102 However, one BCR-ABL mutation called T315I but this number will undoubtedly grow. In addition, it is becoming is resistant to all three drugs. The third-generation ABL kinase in- apparent that many patients have rare mutations (defined as rare in hibitor ponatinib blocks T315I and showed activity in a phase II that histologic tumor type) but are, in theory, actionable. Because clinical trial that included CML patients with the T315I muta- these examples are unlikely to be formally evaluated in clinical tion,103 resulting in FDA approval. However, subsequent reports trials, many centers have opened basket studies (with eligibility of severe vascular occlusive events, such as stroke and heart fail- based solely on mutation profile) to capture these cases with some ure, led to withdrawal from the market, followed by approval for reports of remarkable success. restricted use in T315I-mutant patients. Analogous approaches More effort must be devoted to manipulating the dose and are ongoing in other diseases such as EGFR-mutant lung cancer, schedule of kinase inhibitor therapy to maximize efficacy and min- where acquired resistance to the frontline kinase inhibitor is also imize toxicity. To date, all kinase inhibitors have been developed associated with mutations in the target kinase.104,105 Promising based on the assumption that a 24/7 coverage of the target is re- clinical results have been reported with irreversible EGFR inhibi- quired for efficacy. Consequently, most compounds are optimized tors such as CO-1686 and AZD9291. to have a long serum half-life (12 to 24 hours). Phase II doses are The clinical development of kinase inhibitor combinations to then selected based on the maximum tolerated dose determined prevent acquired resistance is relatively straightforward. Because with daily administration. But a recent clinical of the ABL inhibi- the frontline drug is already approved, success would be deter- tor dasatinib in CML indicates that equivalent antitumor activity mined by an improvement in response duration using the com- can be achieved with intermittent therapy.106 By giving larger doses bination. The situation is more complex when two experimental intermittently, higher peak drug concentrations were achieved compounds (e.g., a PI3K pathway inhibitor and a MEK inhibi- that resulted in equivalent and possibly superior efficacy.107 Simi- tor) are combined, neither of which shows significant single- lar results were observed in laboratory studies of EGFR inhibitors agent activity. Older regulatory guidelines required a four-arm in EGFR-mutant lung cancer. Clinically robust, quantitative as- study that compared each single agent to the combination and to says of target inhibition are needed to hasten progress in this area. a control group in order to obtain approval of the combination. Although the focus of this chapter is kinase inhibitors, the themes Recognizing that this design could discourage drug developers as developed here should apply broadly to inhibitors of other cancer well as patients from moving forward because it requires a large targets. Inhibitors of the G-protein coupled receptor smoothened sample size, the FDA has issued new guidelines for the devel- (SMO) in patients with metastatic basal cell carcinoma or medul- opment of novel combinations that require a two-arm registra- loblastoma establish that the driver mutation hypothesis extends tion study comparing the combination to standard of care http:// beyond kinase inhibitors. SMO is a component in the Hedgehog www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatory pathway, which is constitutively activated in subsets of patients Information/Guidances/UCM236669.pdf. A more challenging with basal cell carcinoma and medulloblastoma due to mutations issue may be dose optimization and dose schedule that is needed in the Hedgehog ligand-binding receptor Patched-1. Treatment to safely combine two investigational drugs. Much like the devel- with the SMO inhibitor vismodegib led to impressive responses in opment of combination chemotherapy several decades ago, it may basal cell carcinoma and medulloblastoma patients whose tumors be important to select compounds with nonoverlapping toxicities had Patched-1 mutations,108,109 resulting in FDA approval. Other to allow for sufficient doses of each drug to be achieved. novel cancer targets are emerging from cancer genome sequencing projects. Somatic mutations in the Krebs cycle isocitrate dehydrogenase (IDH1/2) were found in subsets of patients with glio- SPECULATIONS ON THE FUTURE ROLE OF blastoma, AML, chondrosarcoma, and cholangiocarcinoma,110–112 KINASE INHIBITORS IN CANCER MEDICINE and the first IDH2 inhibitor has entered clinical trials in leukemia. Mutations in enzymes involved in chromatin remodeling, such as The role of genomics in predicting a response to kinase inhibitor the histone methyltransferase EZH2, have been reported in lym- therapy is now irrefutable. As the number of kinase driver muta- phoma and have spurred the ongoing development of EZH2 inhib- tions continues to grow, the field is likely to move away from the itors.113,114 Inhibitors of another histone methyltransferase DOT1L, current strategy of a companion diagnostic for each drug. Rather, which is required for the maintenance of mixed lineage leukemia comprehensive mutational profiling platforms that query each (MLL) fusion leukemias, are also in clinical development.115,116 tumor for hundreds of potential cancer mutations are more likely Kinase inhibitors are just the first wave of molecularly targeted to emerge as the diagnostic platform. The number of directly drugs ushered in by our understanding of the molecular underpin- actionable mutations (meaning the presence of a mutation defines nings of cancer cells. There is much more to follow.

SELECTED REFERENCES

The full reference list can be accessed at lwwhealthlibrary.com/oncology. atic patients with non-small cell lung cancer: a randomized trial. JAMA 2003;290(16):2149–2158. 1. Sawyers CL. Shifting paradigms: the seeds of oncogene addiction. Nat Med 10. Paez JG, Jänne PA, Lee JC, et al. EGFR mutations in lung cancer: corre- 2009;15(10):1158–1161. lation with clinical response to gefitinib therapy. Science 2004;304(5676): 2. Weinstein IB. Cancer. Addiction to oncogenes—the Achilles heal of cancer. 1497–1500. Science 2002;297(5578):63–64. 11. Lynch TJ, Bell DW, Sordella R, et al. Activating mutations in the epidermal 3. Druker BJ, Talpaz M, Resta DJ, et al., Efficacy and safety of a specific inhibi- growth factor receptor underlying responsiveness of non-small-cell lung can- tor of the BCR-ABL tyrosine kinase in chronic myeloid leukemia. N Engl J cer to gefitinib. N Engl J Med 2004;350(21):2129–2139. Med 2001;344(14):1031–1037. 12. Pao W, Miller V, Zakowski M, et al. EGF receptor gene mutations are 4. Hirota S, Isozaki K, Moriyama Y, et al. Gain-of-function mutations of c-kit in common in lung cancers from “never smokers” and are associated with human gastrointestinal stromal tumors. Science 1998;279(5350):577–580. sensitivity of tumors to gefitinib and erlotinib. Proc Natl Acad Sci U S A 5. Heinrich MC, Corless CL, Duensing A, et al. PDGFRA activating mutations 2004;101(36):13306–13311. in gastrointestinal stromal tumors. Science 2003;299(5607):708–710. 14. Shepherd FA, Rodrigues Pereira J, Ciuleanu T, et al. Erlotinib in previously 6. Demetri GD, von Mehren M, Blanke CD, et al. Efficacy and safety of ima- treated non-small-cell lung cancer. N Engl J Med 2005;353(2):123–132. tinib mesylate in advanced gastrointestinal stromal tumors. N Engl J Med 19. Tsao MS, Sakurada A, Cutz JC, et al. Erlotinib in lung cancer - molecular 2002;347(7):472–480. and clinical predictors of outcome. N Engl J Med 2005;353(2):133–144. 8. Kris MG, Natale RB, Herbst RS, et al. Efficacy of gefitinib, an inhibi- 20. Mok TS, Wu YL, Thongprasert S, et al. Gefitinib or carboplatin-paclitaxel in tor of the epidermal growth factor receptor tyrosine kinase, in symptom- pulmonary adenocarcinoma. N Engl J Med 2009;361(10):947–957.

tahir99 - UnitedVRG Chapter 22 Kinase Inhibitors as Anticancer Drugs 247

23. Cools J, DeAngelo DJ, Gotlib J, et al. A tyrosine kinase created by fusion of 85. Bachelot T, Bourgier C, Cropet C, et al. Randomized phase II trial of evero- the PDGFRA and FIP1L1 genes as a therapeutic target of imatinib in idio- limus in combination with tamoxifen in patients with hormone receptor-pos- pathic hypereosinophilic syndrome. N Engl J Med 2003;348(13):1201–1214. itive, human epidermal growth factor receptor 2-negative metastatic breast 25. Apperley JF, Gardembas M, Melo JV, et al. Response to imatinib mesylate cancer with prior exposure to aromatase inhibitors: a GINECO study. J Clin in patients with chronic myeloproliferative diseases with rearrangements of Oncol 2012;30(22):2718–2724. the platelet-derived growth factor receptor beta. N Engl J Med 2002;347(7): 86. Baselga J, Campone M, Piccart M, et al. Everolimus in postmenopaus- 481–487. al hormone-receptor-positive advanced breast cancer. N Engl J Med 27. Davies H, Bignell GR, Cox C, et al. Mutations of the BRAF gene in human 2012;366(6):520–529. cancer. Nature 2002;417(6892):949–954. 87. Carver BS, Chapinski C, Wongvipat J, et al. Reciprocal feedback regulation 28. Baxter EJ, Scott LM, Campbell PJ, et al., Acquired mutation of the ty- of PI3K and androgen receptor signaling in PTEN-deficient prostate cancer. rosine kinase JAK2 in human myeloproliferative disorders. Lancet Cancer Cell 2011;19(5):575–586. 2005;365(9464):1054–1061. 88. Krueger DA, Care MM, Holland K, et al. Everolimus for subependymal 29. James C, Ugo V, Le Couédic JP, et al. A unique clonal JAK2 muta- giant-cell astrocytomas in tuberous sclerosis. N Engl J Med 2010;363(19): tion leading to constitutive signalling causes polycythaemia vera. Nature 1801–1811. 2005;434(7037):1144–1148. 89. Iyer G, Hanrahan AL, Milowsky MI, et al. Genome sequencing identifies a 30. Levine RL, Wadleigh M, Cools J, et al. Activating mutation in the tyrosine basis for everolimus sensitivity. Science 2012;338(6104):221. kinase JAK2 in polycythemia vera, essential thrombocythemia, and myeloid 90. O’Reilly KE, Rojo F, She QB, et al. mTOR inhibition induces upstream metaplasia with myelofibrosis. Cancer Cell 2005;7(4):387–397. receptor tyrosine kinase signaling and activates Akt. Cancer Res 2006;66(3): 31. Samuels Y, Wang Z, Bardellli A, et al. High frequency of mutations of the 1500–1508. PIK3CA gene in human cancers. Science 2004;304(5670):554. 91. Gonzalez-Angulo AM, Juric D, Argilis G, et al. Safety, pharmacokinetics, 32. International Cancer Genome Consortium, Hudson TJ, Anderson and preliminary activity of the alpha-specific P13K inhibitor BYL719: results W, et al. International network of cancer genome projects. Nature from the first-in-human study. J Clin Oncol 2013;31(15 Suppl):2531. 2010;464(7291):993–998. 92. Byrd JC, Furman RR, Coutre SE, et al. Targeting BTK with ibrutinib in 33. Vogelstein B, Papadopoulos N, Velculescu VE, et al. Cancer genome land- relapsed chronic lymphocytic leukemia. N Engl J Med 2013;369(1):32–42. scapes. Science 2013;339(6127):1546–1558. 93. Wang ML, Rule S, Martin P, et al. Targeting BTK with ibrutinib in relapsed 34. Lawrence MS, Stojanov P, Mermel CH, et al. Discovery and saturation anal- or refractory mantle-cell lymphoma. N Engl J Med 2013;369(6):507–516. ysis of cancer genes across 21 tumour types. Nature 2014;505(7484):495–501. 95. Engelman JA, Chen L, Tan X, et al. Effective use of PI3K and MEK inhibi- 35. Eisen T, Ahmad T, Flaherty KT, et al. Sorafenib in advanced mela- tors to treat mutant Kras G12D and PIK3CA H1047R murine lung cancers. noma: a Phase II randomised discontinuation trial analysis. Br J Cancer Nat Med 2008;14(12):1351–1356. 2006;95(5):581–586. 96. Gorre ME, Mohammed M, Ellwood K, et al. Clinical resistance to STI-571 36. Solit DB, Garraway LA, Pratilas CA, et al. BRAF mutation predicts sensitivity cancer therapy caused by BCR-ABL gene mutation or amplification.Science to MEK inhibition. Nature 2006;439(7074):358–362. 2001;293(5531):876–880. 38. Pratilas CA, Taylor BS, Ye Q, et al. (V600E)BRAF is associated with disabled 97. Shah NP, Nicoll JM, Nagar B, et al. Multiple BCR-ABL kinase domain mu- feedback inhibition of RAF-MEK signaling and elevated transcriptional out- tations confer polyclonal resistance to the tyrosine kinase inhibitor imatinib put of the pathway. Proc Natl Acad Sci U S A 2009;106(11):4519–4524. (STI571) in chronic phase and blast crisis chronic myeloid leukemia. Cancer 42. Flaherty KT, Robert C, Hersey P, et al. Improved survival with MEK inhibi- Cell 2002;2(2):117–125. tion in BRAF-mutated melanoma. N Engl J Med 2012;367(2):107–114. 98. Shah NP, Tran C, Lee FY, et al. Overriding imatinib resistance with a novel 45. Flaherty KT, Puzanov I, Kim KB, et al. Inhibition of mutated, activated ABL kinase inhibitor. Science 2004;305(5682):399–401. BRAF in metastatic melanoma. N Engl J Med 2010;363(9):809–819. 99. Talpaz M, Shah NP, Kantarjian H, et al. Dasatinib in imatinib-resistant CANCER THERAPEUTICS 46. Bollag G, Hirth P, Tsai J, et al. Clinical efficacy of a RAF inhibitor needs Philadelphia chromosome-positive leukemias. N Engl J Med 2006;354(24): broad target blockade in BRAF-mutant melanoma. Nature 2010;467(7315): 2531–2541. 596–599. 101. Sawyers CL. Even better kinase inhibitors for chronic myeloid leukemia. N 47. Hauschild A, Grob JJ, Demidov LV, et al. Dabrafenib in BRAF-mutated met- Engl J Med 2010;362(24):2314–2315. astatic melanoma: a multicentre, open-label, phase 3 randomised controlled 104. Pao W, Miller VA, Politi KA, et al. Acquired resistance of lung adenocarci- trial. Lancet 2012;380(9839):358–365. nomas to gefitinib or erlotinib is associated with a second mutation in the 50. Soda M, Choi YL, Enomoto M, et al. Identification of the transform- EGFR kinase domain. PLoS Med 2005;2(3):e73. ing EML4-ALK fusion gene in non-small-cell lung cancer. Nature 105. Antonescu CR, Besmer P, Guo T, et al. Acquired resistance to imatinib in 2007;448(7153):561–566. gastrointestinal stromal tumor occurs through secondary gene mutation. Clin 52. Shaw AT, Kim DW, Nakagawa K, et al. Crizotinib versus chemotherapy in Cancer Res 2005;11(11):4182–4190. advanced ALK-positive lung cancer. N Engl J Med 2013;368(25):2385–2394. 106. Shah NP, Kantarjian HM, Kim DW, et al. Intermittent target inhibition with 53. Chen Y, Takita J, Choi YL, et al. Oncogenic mutations of ALK kinase in dasatinib 100 mg once daily preserves efficacy and improves tolerability in neuroblastoma. Nature 2008;455(7215):971–974. imatinib-resistant and -intolerant chronic-phase chronic myeloid leukemia. J 63. Sawyers CL. Finding the next Gleevec: FLT3 targeted kinase inhibitor ther- Clin Oncol 2008;26(19):3204–3212. apy for acute myeloid leukemia. Cancer Cell 2002;1(5):413–415. 107. Shah NP, Kasap C, Weier C, et al. Transient potent BCR-ABL inhibition is 69. Zarrinkar PP, Gunawardane RN, Cramer MD, et al. AC220 is a uniquely sufficient to commit chronic myeloid leukemia cells irreversibly to apoptosis. potent and selective inhibitor of FLT3 for the treatment of acute myeloid Cancer Cell 2008;14(6):485–493. leukemia (AML). Blood 2009;114(14):2984–2992. 112. Ward PS, Patel J, Wise DR, et al. The common feature of leukemia-associat- 71. Smith CC, Wang Q, Chin CS, et al. Validation of ITD mutations in ed IDH1 and IDH2 mutations is a neomorphic enzyme activity converting FLT3 as a therapeutic target in human acute myeloid leukaemia. Nature alpha-ketoglutarate to 2-hydroxyglutarate. Cancer Cell 2010;17(3):225–234. 2012;485(7397):260–263. 113. McCabe MT, Ott HM, Ganji G, et al. EZH2 inhibition as a thera- 72. Kaelin WG Jr. The von Hippel-Lindau tumour suppressor protein: O2 sens- peutic strategy for lymphoma with EZH2-activating mutations. Nature ing and cancer. Nat Rev Cancer 2008;8(11):865–873. 2012;492(7427):108–112. 73. Yang JC, Haworth L, Sherry RM, et al. A randomized trial of bevacizumab, 114. Morin RD, Johnson NA, Severson TM, et al. Somatic mutations altering an anti-vascular endothelial growth factor antibody, for metastatic renal can- EZH2 (Tyr641) in follicular and diffuse large B-cell lymphomas of germinal- cer. N Engl J Med 2003;349(5):427–434. center origin. Nat Genet 2010;42(2):181–185. 83. Guertin DA, Sabatini DM. Defining the role of mTOR in cancer.Cancer 115. Bernt KM, Zhu N, Sinha AU, et al. MLL-rearranged leukemia is dependent Cell 2007;12(1):9–22. on aberrant H3K79 methylation by DOT1L. Cancer Cell 2011;20(1):66–78. 84. Thomas GV, Tran C, Mellinghoff IK, et al. Hypoxia-inducible factor de- 116. Daigle SR, Olhava EJ, Therkelsen CA, et al. Selective killing of mixed lin- termines sensitivity to inhibitors of mTOR in kidney cancer. Nat Med eage leukemia cells by a potent small-molecule DOT1L inhibitor. Cancer 2006;12(1):122–127. Cell 2011;20(1):53–65. Inhibitorsors 23 and Demethylating Agents

Steven D. Gore, Stephen B. Baylin, and James G. Herman

INTRODUCTION or translocations), epigenetic changes do not alter the coding se- quence of targeted genes. Thus, reversal of epigenetic changes The past decade has seen an explosive growth, especially at a can potentially restore the normal function of affected genes and genome-wide level, in our understanding of the role of chromatin their encoded proteins. Second, the heritable nature of epigenetic in the normal regulation of gene expression and in the concept changes—that is, the ability of a cell to pass on regulation of gene of the epigenome.1–3 Concomitant with these advances has been expression through DNA replication—suggests that such changes, the increasing appreciation of the role of epigenetic abnormali- while relatively stable, can be reversed. Thus, therapeutic repro- ties in the progression of cancer4–7 and the concept of the cancer gramming of patterns of gene expression could theoretically result epigenome. The translational consequences of this research in a long-term change in the cancer cell phenotype, even after the include the possibilities for developing therapies in cancer that inducing drugs are removed, although to date, this has not been target epigenetic abnormalities. These are being explored in clini- accomplished. cal trials and several have entered clinical practice.4,5,7,8 Of these The fundamental unit that determines epigenetic states is the epigenetic abnormalities, the most thoroughly examined is the nucleosome that contains an octamer of histone proteins around occurrence of abnormal cytosine guanine (CpG) promoter region which approximately 160 base pairs of DNA are wrapped.13 It is the DNA methylation and associated altered chromatin involving positioning of these structures, and the three-dimensional aspects histone modifications, in the transcriptional silencing of genes, of their spacing, and the regulation of this process by posttransla- including a group of well-defined tumor suppressor genes.4,5,7,8 tional modifications of the constituent histones that underpins the However, targeting epigenetic processes to downregulate the ac- functions of the epigenome.13,14 tion of overexpressed genes is also an emerging area of research.9,10 This chapter describes the basis of epigenetic changes in cancer Abnormal Gene Silencing and discusses some of the latest approaches that target epigenetic 11 abnormalities in cancer, including those designed to induce the One key alteration in cancer, which can be associated with altered reexpression of silenced genes, for cancer therapy. The two ap- epigenetic control, is abnormal gene silencing. Normally, such si- proaches most mature in development are the inhibition of DNA lencing is fundamental and required at the level of chromatin and methyltransferases, which mediate the abnormal promoter DNA DNA methylation regulation for the life of multicellular eukary- methylation, and the inhibition of histone deacetylases, which otic organisms. The silencing is critical for regulating important remove histone modifications associated with active chromatin biologic processes, including all aspects of development, differen- that alone, or in association with DNA methylation, are associ- 4,5,7,8 tiation, imprinting, and silencing of large chromosomal domains, ated with transcriptional repression. However, several excit- including the X chromosome of female mammals.13 For example, ing newer approaches are now in clinical trials and these will be the diversity of structure and function of cells derived from epi- mentioned. thelial or mesenchymal origin, ultimately differentiating into cells Aberrant gene function and altered patterns of gene expression 4 lining the intestine or lung or forming mature granulocytes and are key features of cancer. Although genetic alterations remain the myocytes, result from heritable changes in gene expression that best characterized in the development and progression of cancer, are not the result of a change in DNA sequence. Although in many increasingly it is appreciated that epigenetic abnormalities coop- species, silencing can be initiated and maintained solely by pro- erate with genetic alterations in multiple ways to cause dysfunc- cesses involving the covalent modifications of histones and other tion of key regulatory pathways. Through genomic approaches to chromatin components, vertebrates utilize an additional layer of mutation discovery, there is growing recognition of the frequency gene regulation. This process involves the only natural covalent of mutations in genes encoding for proteins that regulate the 12 modification of DNA in humans and is characterized by DNA epigenome. This chapter will outline the understanding of how cytosine methylation that occurs nearly exclusively at the fifth po- each of these epigenetic alterations contribute to cancer and how sition of the cytosine ring in cytosines preceding guanine, the so- derivation of therapeutic approaches may depend on understand- called CpG dinucleotide (Fig. 23.1).13,15 ing the biology of these changes. Like most biologic processes, the normal patterns of silenc- ing can be altered, resulting in the development of disease states. Thus, activation of genes normally not expressed, or silencing of a EPIGENETIC ABNORMALITIES AND GENE gene that should be expressed, can contribute to the dysregulation EXPRESSION CHANGES IN CANCER of gene function that characterizes cancer and, when stably pres- ent, represent epigenetic alterations.4–7 Most studies have focused Epigenetic changes are defined as heritable alterations of gene on the silencing of normally expressed genes. For the purposes of expression patterns and cell phenotypes, which are not accompa- understanding the rationale behind epigenetic therapy, it is impor- nied by changes in DNA sequence.13 This definition clearly delin- tant to understand the mechanisms through which such silencing eates the two key features of epigenetic regulation important for occurs. Alterations in gene expression associated with epigenetic an understanding of therapies described in this chapter. Specifi- changes that give rise to a growth advantage would be expected to cally, in contrast to genetic alterations (point mutations, deletions, be selected for in the host tissue, leading to progressive dysregulated 248

tahir99 - UnitedVRG Chapter 23 Histone Deacetylase Inhibitors and Demethylating Agents 249

Silenced gene Condensed chromatin Methylated cytosines Histones deacetylated

Active gene Open chromatin Unmethylated cytosines CpG island Histones acetylated

Transcription DNA Nucleosome Histone Methyl group Acetyl group

Figure 23.1 Epigenetic regulation of gene expression. In the promoter region, gene expression is controlled by a combination of DNA methylation and chromatin configuration. In normal cells, gene expression is silenced by condensing chromatin, methylating DNA, and deacetylating histones. By contrast, active genes are those with open nucleosome spacing around the transcription start site, are unmethylated, and are associated with acetylated histones. In cancer cells, CpG islands that are rich in cytosine and guanine—and are typically unmethylated to promote gene expression—can be epigenetically silenced by hypermethylation. (Redrawn with permission from Azad N, Zahnow CA, Rudin CM, et al. The future of epigenetic therapy in solid tumours—lessons from the past. Nat Rev Clin Oncol

2013;10:256–266.) CANCER THERAPEUTICS

growth of the tumor. Such dysregulation is commonly associated cell state.18 Abnormal promoter DNA methylation of such genes with increases in promoter region DNA methylation and is associ- renders them more repressed and could be a factor in the fact that ated with repressive chromatin changes. cancers inevitably exhibit cell populations with enhanced self- renewal or refractoriness to full differentiation.18 Changes in DNA Methylation Recent studies have also suggested that DNA regions other than promoter CpG islands may undergo changes of DNA The importance of abnormal cytosine methylation and gene si- methylation in cancer. For example, non–CpG-rich sequences lencing has been clearly established in the past 2 decades and been surrounding promoter CpG islands, termed CpG island shores, shown convincingly to be involved in cancer development.4–7 The are abnormally methylated in cancers19 and may be altered in CpG dinucleotide, usually underrepresented in the genome, is stem cell populations. 20 Thus, the relative cancer specificity of clustered in the promoter regions of approximately 50% of human changes of DNA methylation in multiple CpG regions makes genes in regions termed CpG islands. These regions are largely reversal of these changes by targeting DNA methyltransferases, protected from DNA methylation in normal cells, with the ex- the enzymes that catalyze DNA methylation, logical for cancer ception of genes on the inactive X chromosome and imprinted therapeutics. genes.16 This protection is critical, because the methylation of pro- As a key example of the previous points, perhaps the most stud- moter region CpG islands is associated with a loss of gene expres- ied tumor suppressor gene for promoter hypermethylation is the sion.4–7 Abnormal de novo DNA methylation of gene promoter p16 gene, currently designated CDKN2A, a cyclin-dependent ki- CpG islands is a very frequent abnormality in virtually all cancer nase inhibitor that functions in the regulation of the phosphoryla- types and is associated with a process that can serve as an alter- tion of the Rb protein. Hypermethylation associated with loss of native mechanism for loss of tumor suppressor gene function.4–7 expression of the CDKN2A gene has been found to be one of the Although a limited number of classic tumor suppressor genes can most frequent alterations in neoplasia being common in the lung, be affected by this process, a patient’s individual cancer may har- head and neck, gliomas, colorectal, and breast carcinomas21,22 and bor hundreds of such genes.4–7 Which of these latter genes are other cancer types. A member of the same gene family, p15 or CD- drivers of cancer, individually or in groups, versus those which are KN2B, also regulates Rb and is silenced in association with pro- passengers reflecting only the widespread effects of a global epi- moter methylation in many forms of leukemia and in the chronic genetic abnormality is a leading question in the field and the target myeloid neoplasm myelodysplastic syndrome (MDS).23 These two of much research.5,6 A clue to the importance of at least groups previous changes are of much relevance for the clinical uses of of the previous DNA hypermethylated genes may come from the epigenetic therapies discussed later. fact that an inordinate number of them are involved in holding As mentioned, many hundreds of genes may be inactivated in a normal embryonic and adult stem cells in the self-renewal state single cancer by promoter methylation,5,6,18,24 providing potential and/or rendering such cells refractory to differentiation cues.17,18 targets for gene reactivation using epigenetic therapies.25–27 The Normally, these genes are then in a poised expression state and can latter represents one of the potential ways in which epigenetic be induced to be activated or repressed as needed for changes in therapy may be effective: Multiple genes and gene pathways, all 250 Cancer Therapeutics

Stemlike behavior Metastasis or EMT p16 (CDKN2A, p16Ink4A), IGFBP3, MMP9, GATA4, NONOG SOX family genes, GATA5, FBN2, NRCAM, CDH13 POU5F1 (OCT-4), miR-34n, DMBX1 (PaxB), HOX family genes

M G2 G1 S

Chemoresistance Cell proliferation WRN, LGR5, ASCL2, and survival CDKN1C (p57kip2), RASSF1A, PTEN, SNK/PLK2 DKK4, BMP3 Alter tumor biology

Immune responsiveness Apoptosis or chemotherapy sensitivity CD58, MAGE antigens, DAPK1, APAF1, SPARC, TMS1/ASC (PYCARD), B2M, STAT1, MHC1 antigens ESR1 (estrogen receptor α), BNIP3

Figure 23.2 Concurrent widespread changes in gene expression with epigenetic therapy. Anticancer efficacy of treatment with epigenetic-modulating agents is associated with extensive changes in gene expression that influence several biologic processes. Gene expression is increased through the direct reversal of epigenetic modifications of genomic DNA, whereas for cancer-promoting genes, gene expression is reduced by the regression of their regulatory genes. EMT, epithelial-membrane transition. (Redrawn with permission from Azad N, Zahnow CA, Rudin CM, et al. The future of epigenetic therapy in solid tumours—lessons from the past. Nat Rev Clin Oncol 2013;10:256–266.)

repressed by changes in DNA methylation and chromatin modi- generally works together with histone modifications to mediate a fication, can be reactivated by DNA methyltransferase inhibitors repressive state for that enhancer.31–33 The status of enhancers is and histone deacetylase (HDAC) inhibitors (HDACi), thereby also emerging as important for cancer risk states.34 restoring normal cell cycle control, differentiation, and apop- totic signaling (Fig. 23.2).8,26,28 In general, methylated CpG is- lands are not capable of the initiation of transcription unless the Chromatin in Gene Regulation methylation signal can be overridden by alterations in factors that modulate chromatin, such as the removal of methylated cytosine- Heritable gene silencing involves the interplay between DNA binding proteins. However, reversal of DNA methylation with methylation and histone covalent modifications. Complexes of secondary changes in histone modification or directed reversal of proteins that can regulate how nucleosomes are positioned per- repressive histone modifications represent a target for epigenetic form nucleosomal remodeling.35–37 What was initially termed therapies.8,26,28 the histone code, with reference to how histones are modi- Most studies of DNA methylation, particularly in the study of fied, has emerged to be much more complex than originally cancer, have focused on CpG island promoter methylation. How- envisioned. An explosion of research findings during the last ever, about 40% of human genes do not contain bona fide CpG several years now allows for an appreciation of how the epig- islands in their promoters.29 The primary focus on CpG islands has enome is controlled by a complex interplay between a myriad resulted from the clear demonstration that CpG-island promoter of posttranslational histone modifications that occur on key methylation permanently silences genes both physiologically and amino acid residues of these proteins.37 Acetylation, deacety- pathologically in mammalian cells. However, recent work has shown lation, methylation, phosphorylation, and other modifications correlations between tissue-specific expression and methylation of all modify chromatin structure and thereby alter gene expres- non-CpG islands, including, for example, the maspin gene,30 and sion.38 Some of the enzymes that catalyze these modifications as mentioned previously, regions near CpG islands,19,20 suggesting include HDACs, histone methyltransferases (HMT), and most that many additional genes could be regulated, either normally or recently, histone demethylases.13,14,39,40 These modifications abnormally, by changes in DNA methylation. help establish heritable states at the start site of genes, but also An exciting new area of DNA methylation research involves the at enhancers and other transcribed DNA regions not encoding role of this change in regulating gene enhancers: small DNA re- for canonical genes. The latter areas contain noncoding RNAs gions that regulate the expression of multiple target genes.31–33 The (ncRNAs) and micro-RNAs (miRNAs), which play key modula- presence of DNA methylation in these areas, which can reside tory roles for overall gene expression and protein patterns that considerable distances from the genes that are being regulated, can be altered in cancer.41–43 Again, much research is being

tahir99 - UnitedVRG Chapter 23 Histone Deacetylase Inhibitors and Demethylating Agents 251

focused on epigenetic changes in these DNA regions, which Of the previously listed HDACs, class I and 2 HDACs have may be important to cancer development and, potentially, to been most closely tied to gene silencing associated with ab- cancer management. normal promoter DNA hypermethylation.48 These are bound A link between covalent histone modifications and DNA meth- to the nucleosome remodeling complex, NuRD.48,49 Experi- ylation has been clearly established.44–46 In this interaction, cyto- mental decreases in NURD, after use of a DNA demethylating sine methylation attracts methylated DNA-binding proteins and agent, can augment reactivation of many abnormally silenced HDACs to methylated CpG sites during chromatin compaction and DNA hypermethylated genes in colon cancer cells.48 Ma- and gene silencing.46,47 In addition, the DNA methylation binding nipulation of these HDACs is under study in clinical trials, protein (MBD2) interacts with the nucleosomal remodeling com- with and without the use of DNA methyltransferase inhibitors, plex (NuRD) and directs the complex to methylated DNA.48 This and is discussed later. Another HDAC, SIRT1 in the class III complex also binds HDACs and has recently been identified as a of these proteins, is also involved with gene silencing.54,55 This central player for the abnormal silencing of genes associated with deacetylase has been linked to silencing of DNA hypermeth- promoter DNA hypermethylation in cancer.47 Thus, the three ylated genes, and blocking its activity can be associated with processes of DNA cytosine methylation, histone modification, and reactivation of such genes.55 nucleosomal remodeling are intimately linked, and alterations in these processes can result in abnormalities of gene expression in Reversal of Layers of Gene Silencing cancer-relevant genes. The interaction between DNA methylation and HDAC activity Enzymes Regulating DNA Methylation and and repressive chromatin marks in maintaining aberrant silenc- Histone Acetylation ing of hypermethylated genes in cancer has therapeutic implica- tions for epigenetic therapies. Experimental evidence suggests that DNA methylation functions as a dominant event that stably DNA methylation involves the covalent addition of a methyl group establishes transcriptional repression. Inhibition of HDAC activity to the 5′ position of cytosine. In mammals, three enzymes have alone, by potent and specific HDACis, does not generally result in been shown to catalyze this transfer of a methyl group from the the reactivation of aberrantly silenced and densely hypermethyl- methyl donor S-adenosylmethionine. Most of the methyltransfer- ated genes in tumor cells.56 In contrast, treatment with HDACis ase activity present in differentiated cells is derived from the expres- can reactivate densely silenced genes if the cells are first treated sion of DNMT1.49 This enzyme is thought to be most important with demethylating drugs, such as 5-.56 The clinical im- in maintaining DNA methylation patterns following DNA replica- plications of this observation are discussed in more detail in the tion and thus is referred to as a maintenance methyltransferase. following section (Table 23.1). However, the enzyme does possess the ability to methylate pre- 50 viously unmethylated DNA sequences (de novo activity). In CANCER THERAPEUTICS contrast, the other enzymes, DNMT3a and DNMT3b, are effi- DNA Methyltransferase Inhibitors cient at methylating previously unmethylated DNA and thus are considered de novo methyltransferases. Each of these enzymes Originally synthesized as cytotoxic drugs in the possesses a similar catalytic site,51 a fact important for the inhibi- 1960s,57 azacytosine nucleosides were recognized as inhibitors of tion of DNMT enzymes by nucleoside analogs, discussed later in DNA methylation in the early 1980s. The inhibitors 5-azacitidine this chapter. (5AC) and 2′-deoxy-5-azacytidine induced muscle, fat, and chon- DNA methylation is closely associated with changes in the drocyte differentiation in mouse embryo cells, in association with histone modifications. As previously discussed, histone proteins a reversal of DNA methylation.58,59 The incorporation of azacyto- are the central components of the nucleosome, and modifica- sine nucleosides into DNA in lieu of cytosine residues was shown tions of the histone tails of core histones are associated with ac- to be associated with inhibition of DNMT activity.59,60 DNMT tive or repressed chromatin.52 Although it is beyond the scope of inhibition requires the incorporation of triphosphate this chapter to fully discuss the complex series of modifications to into DNA. The incorporated azacytosine nucleoside forms an the histone tails of histone H3 and H4, a few well-characterized irreversible inactive adduct with DNMT. The sequential reversal modifications should be mentioned that are relevant to therapies of DNA methylation then results when DNA replication proceeds designed to target epigenetic abnormalities in cancer. In refer- in the absence of active DNMT.61 The inhibitor 5AC must be ence to currently investigated epigenetic therapies, changes in phosphorylated and converted to decitabine diphosphate by ribo- histone acetylation are of importance. Acetylation of histones H3 nucleotide reductase before it can be activated through triphos- and H4 at key amino acids is associated with the active chromatin phorylation, whereas decitabine does not require ribonucleotide present at the promoters of transcribed genes, whereas the ab- reductase. The inhibitor 5AC can also be incorporated into RNA. sence of histone acetylation is associated with repressed, silenced DNMT2, a misnamed protein that is actually an RNA-specific genes.13,14,53 Histone acetyltransferases (HAT) HDACs have op- methyltransferase,62 becomes inhibited, leading to the depletion posing functions to maintain the proper level of histone acety- of methylated tRNA.60 This may contribute to the inhibition of lation for gene expression.13,14,53 HDACs specifically deacetylate protein synthesis and is a potential difference between azacitidine the lysine residues of the histone tails, and this deacetylation is and decitabine.63 The previous DNA methyltransferase inhibitors associated with condensation of nucleosome positions in what not only block the catalytic activities of DNMTs, but also trigger is termed a closed chromatin formation. This scenario is key to degradation of these proteins, especially DNMTs 1 and 3B.64–68 transcriptional repression. There are four classes of HDACs.53 This latter activity is potentially important for their activities for Class I HDACs are characterized by their similarity to the gene reexpression because each of these two proteins, experimen- yeast Rpd3 HDAC. In humans, this class of enzymes includes tally, possess transcriptional repression properties independent of HDAC1, -2, -3, and -8. These HDACs are thought to be ubiqui- their DNA methylation catalytic sites.69,70 tously expressed in tissue throughout the body. In contrast, class The azacytosine nucleosides exhibit complex dose–response II HDACs are similar to yeast Hda1 and include HDAC4, -5, -6, characteristics. At low concentrations (0.2 to 1 μM), the epigenetic -7, -9, and -10, and they have a greater degree of tissue specificity. activities of these drugs predominate, with dose-dependent reversal Class III HDACs are similar to yeast Sir2 and are set apart from of DNA methylation71,72 and induction of terminal differentia- the other classes by their dependence on nicotinamide adenine tion in some systems.28,71 As concentrations are increased, DNA dinucleotide (NAD+) as a . Finally, class IV includes damage and apoptosis become more prominent.28,72 Cell lines HDAC11.53 with 30-fold resistance to the cytotoxic effects of , 252 Cancer Therapeutics

TABLE 23.1 Small Molecules Targeting Epigenetic Abnormalities in Clinical Development

Route of Drug Class Target Dose Range Schedule Administration 5-Azacitidine Nucleoside DNA methyl- 30–75 mg/m2/d Daily × 7–14 d/28 d Subcutaneous or intravenous 2′-Deoxy-5- Nucleoside DNA methyl-transferase 10–45 mg/m2/d Daily × 3–5 d/4–6 wk Intravenous azacytidine SG110 Nucleoside DNA methyl-transferase Being determined Being determined Subcutaneous Valproic acid Small chain fatty acid Histone deacetylase 25–50 mg/kg/d Daily Oral or intravenous (class I and II) Hydroxamic acid Histone deacetylase 400–600 mg/d Divided doses Oral (class I and II) Benzamide –8 mg/m2 Weekly Oral (class I) Hydroxamic acid Histone deacetylase 600–1,000 mg/m2 Daily × 5/28 d Intravenous (class I and II) Cyclic tetrapeptide Histone deacetylase 13–18 mg/m2 Weekly Intravenous (class I and II) LBH-589 Hydroxamic acid –11 mg/m2 Daily × 3 Intravenous (class I and II) MGCD-0103 Benzamide Histone deacetylase 40–125 mg/m2 Twice weekly Oral (class I) CI-994 Benzamide Histone deacetylase 5–8 mg/m2 Daily Oral (class I)

adriamycin , cyclophosphamide (DAC) continue to reverse (approximately 22 minutes).75 DAC given over 1 hour IV at 15 methylation in response to this nucleoside, suggesting that the to 20 mg/m2 produced plasma concentrations of 1.1 to 1.6 μM methylation reversing and cytotoxic activities of this compound during the infusion,77 whereas in a phase 1 study in patients with can be separated.73 The ability of these drugs to inhibit the cell thoracic malignancies, patients were treated with escalating doses cycle, at least in part through induction of p21WAF1/CIP1 expression, of decitabine for 72-hour IV infusions for two 35-day cycles. The complicates the goal of reversing DNA methylation, because the maximum tolerated total dose was 60 to 75 mg/m2 with neutrope- latter requires DNA replication with the azacytosine nucleoside nia as the dose-limiting toxicity. Steady-state plasma concentra- incorporated into the DNA. tions ranged from 25 to 40 nM, which is less than those usually The importance of low doses of the two azacytosine nucleo- used to induce expression of methylated genes in tissue culture sides to achieve a targeted therapeutic effect has been recently models.78 An oral formulation of 5AC has also been studied. The explored in a series of laboratory observations. Transient exposure oral bioavailability of oral azacitidine ranged from 6% to 20%. of both leukemia and solid tumor cells to submicromolar doses in- Nonetheless, MDS and acute myelogenous leukemia (AML) duce such cells to undergo cellular reprogramming, accompanied patients receiving oral azacitidine developed clinical responses by decreases in ability to clone in long-term self-renewal assays and similar to patients receiving parenteral azacitidine. Oral azaciti- to grow as explants in immune-incompetent mice.28 These effects dine has also been safely administered on 14-daily and 21-daily occur with partial genome-wide DNA demethylation and changes schedules repeated monthly. The extended administration of in gene expression in multiple pathways potentially key for driving lower daily doses may provide favorable pharmacodynamics of tumorigenesis. DNA methylation reversal given the need for ongoing cell cycling The pharmacokinetic properties of the two azacytosine nucle- to effect methylation reversal.79 osides are also very important to consider for their clinical use. SGI-110 is a dinucleoside that acts as a prodrug for decitabine. In this regard, a major potential challenge for their usage is the This drug is being studied in myelodysplasia and AML.80 fact that these drugs are highly unstable in an aqueous solution, resulting in their rapid hydrolysis and resultant inactivation.74 In clinical practice, the drugs must be administered shortly after HISTONE DEACETYLASE INHIBITORS reconstitution. The drugs are also metabolized by cytidine de- aminase,74 leading to a short half-life in plasma. When injected The increasing recognition of the critical importance of histone subcutaneously, 5AC reaches a maximal plasma concentration modifications in regulating the transcriptional permissively of at 30 minutes, with a terminal half-life of 1.5 to 2.3 hours.75,76 chromatin has led to intense interest in compounds that can in- At the U.S. Food and Drug Administration (FDA) approved hibit the activity of HDAC proteins, facilitating the acetylation of dose of 5AC (75 mg/m2 administered subcutaneously daily for 7 lysines associated with transcriptional activation of genes. As with days), peak plasma concentrations were 3 to 5 μM, which is well the DNMT inhibitors discussed previously, there are multiple, within the range of DNMT inhibitory concentrations.75,76 Intra- sometimes dose-dependent, effects of HDACis in preclinical stud- venous (IV) administration of the same dose has led to higher ies. Some of these may truly be epigenetic, others strictly cytotoxic, peak plasma concentrations (11 μM) with a shorter half-life and others a combination of both.9,81–84 Some actions of HDACis

tahir99 - UnitedVRG Chapter 23 Histone Deacetylase Inhibitors and Demethylating Agents 253 may relate to altering how chromatin is central to the repair of study.97 Eight patients achieved a partial response, with a median DNA. Thus, at especially high doses, these compounds can blunt time to response of 12 weeks and a median duration of response of efficient repair and even induce DNA breaks.84,85 These effects 15 weeks. Overall, 45% of patients had relief of pruritus. Fatigue, may underlie cell cycle arrest and induction of cell death as is diarrhea, nausea, and thrombocytopenia were common toxicities. often observed in preclinical studies of HDACis.81–84 In a multicenter phase 2 trial, 74 patients with relapsed or refrac- Perhaps novel uses of these drugs may be inferred by results tory CTCL were treated with 400 mg daily.98 Similar to the prior from recent studies suggesting they could be extremely powerful study, 29% of patients responded, consisting almost entirely of par- epigenetic therapy agents when used in proper doses, for targeted tial responses. Median time to response was 56 days, and median purposes, and at key time intervals. Recent studies by Settleman duration of response was greater than 6 months. In phase 1 trials, and colleagues86 suggest that histone acetylation changes, and responses to vorinostat have developed in other non- Hodgkin’s thus epigenetic mechanisms, could be a key factor for cancer and Hodgkin’s lymphoma cases.99 More recently, in a trial com- therapy resistance to both targeted therapy agents and conven- bining vorinostat with carboplatin and paclitaxel in patients with tional chemotherapy. The mechanisms involved may involve the untreated, advanced, non–small-cell lung cancer (NSCLC), emergence of drug-tolerant stem-like cells.86 In such cells, gene response rates increased significantly from 12.5% to 34%, and a expression studies suggest that a protein upregulated in resistance trend to improved progression-free survival and overall survival was is a histone demethylase, which diminishes a key histone modifica- observed.100 tion for active transcription, H3K4methyl.86 A very similar enzyme (LBH589), a cinnamic hydroxamic acid HDACi, has been shown in other studies to be central to self-renewal of reduced peripheral blood blast percentage but did not induce re- stem-like melanoma cells.87 Key to the therapies under discus- missions in a phase 1 trial of daily times 7 oral dosing in patients sion is that, in the previous drug-resistance studies, low doses of with a variety of relapsed hematologic malignancies.101 Asymptom- HDACis, could reversibly reduce drug-resistant cells induced by atic changes in electrocardiographic T waves developed in 80% of the various anticancer drugs.86 It is essential going forward to sort treated patients. Gastrointestinal symptoms and thrombocytopenia out which of these effects are dose-related off-target effects and were common. Panobinostat has recently been approved by the which are desired on-target effects that can be optimized for effica- FDA for the treatment of multiple myeloma.102 cious therapy strategies. Cyclic Tetrapeptides Types of Histone Deacetylase Inhibitors Romidepsin is FDA approved for the treatment of CTCL103 and Small Chain Fatty Acids peripheral T-cell lymphoma.104,105 Antitumor activity, including tumor lysis syndrome, was demonstrated in a phase 1 study that The earliest report of the use of an HDACi to treat leukemia de-

enrolled patients with chronic lymphocytic leukemia and AML, CANCER THERAPEUTICS scribed the treatment of a child with refractory AML with intrave- but no complete or partial remissions were seen.75 The administra- nous sodium butyrate, with a concomitant clearance of peripheral 88 tion of romidepsin induces electrocardiographic changes, includ- blood blast cells and a decrement in bone marrow blasts. No ing T-wave flattening and ST-T wave depression in greater than responses developed in a subsequent study of nine AML patients 89 half of the posttreatment tracings; however, no changes in serum who were treated with intravenous butyrate. Phase 1 studies of cardiac troponin levels or left ventricular ejection fraction have sodium phenylbutyrate (NaPB) in MDS and AML explored 7-day been reported.106 continuous infusions administered monthly or biweekly, and 21-day continuous infusions administered monthly.90,91 At the maximum tolerated dose (375 mg per kilogram per day), the mean Benzamides steady-state plasma concentration was 0.3 mM, within the range of HDAC inhibition.90–92 Isolated patients developed hematologic Entinostat, formerly known as MS-275, was administered weekly improvement in response to NaPB. times four to patients with relapsed and refractory AML in a phase Similar to NaPB, valproic acid (VPA) requires near millimolar 1 study. Infections, unsteady gate, and somnolence were dose- concentrations to effectively inhibit HDACs. Of 18 patients with limiting toxicities. No clinical responses developed, although im- MDS or AML with trilineage dysplasia treated with VPA to tar- provements in neutrophil counts were observed.107 Entinostat did get plasma concentrations of 0.3 to 0.7 mM, 6 patients developed not increase the response rate in patients with higher risk MDS hematologic improvement.93 Of 20 elderly patients with AML and AML with MDS-related changes when combined with azacit- treated with VPA, only 11 could remain in control long enough idine compared to azacitidine alone.108 Most recently, however, to be considered evaluable for response. Five had improvement in studies NSCLC suggest that entinostat could be a valuable thera- platelet counts.94 VPA induced hematologic improvement in com- peutic agent in solid tumors when used with established therapies. bination with all-transretinoic acid in two of eight patients treated When combined with the epidermal growth factor inhibitor er- with AML; a fluorescence in situ hybridization analysis showed lotinib, in a randomized phase 2 trial for patients with recurrent definitive evidence of terminal differentiation of the malignant advanced NSCLC, entinostat was not efficacious alone but ap- cells.95 A larger study of this combination induced hematologic peared to combine with erlotinib to benefit a group of patients response in only 2 of 26 elderly patients with AML.96 It appears whose tumors contained baseline high E-cadherin levels. Overall unlikely that the small chain fatty acids will develop an impor- survival in these latter patients yielded an increased survival ben- tant role in the treatment of malignancy given the availability of efit of 9.4 versus 5.4 months.109 Finally, entinostat significantly in- HDACis with vastly greater potency. creased survival when combined with an aromatase inhibitor in a phase 2 trial for patients with breast cancer.110 Hydroxamic Acids Pharmacodynamic Properties The FDA approved vorinostat as the first commercially available HDACi. The approval was based on activity of this agent in cu- The administration of oral vorinostat was associated with a tran- taneous T-cell lymphoma (CTCL). Thirty-three patients with a sient increase in acetylation of histone H3 in peripheral blood median number of five prior systemic therapy regimens received lymphocytes, which peaked at 2 hours post dosing and reverted to one of three dose schedules of vorinostat in a single institution baseline by 8 hours; similar changes were observed in the lymph 254 Cancer Therapeutics node of a treated patient with lymphoma.99 Treatment with vorin- alive compared with 26.2% (95% CI, 18.7 to 34.3) in the conven- ostat was associated with translocation of phosphorylated signal tional care group (p < 0.0001). Median time to AML transforma- transducer and activator of transcription 3 (STAT-3) from nucleus tion was 17.8 months (IQR 8.6 to 36.8; 95% CI, 13.6 to 23.6) in to cytoplasm in responding patients and with reduced microvessel the 5AC group compared with 11.5 months (4.9 not reached; 8.3 density.97 to 14.5) in the conventional care group (HR 0.50; 95% CI, 0.35 Similar changes in the acetylation of histones 2B and 3 were to 0.70; p < 0.0001). Subsequent unplanned analyses of AZA001 observed in peripheral blood cells from patients treated with included an examination of elderly patients with what would LBH589.101 Romidepsin induced acetylation of H3 and H4 now be classified as AML (blast count 20% to 30%). In these 113 in peripheral blood tumor cells within 4 hours of dosing111; of patients, there remains a statistically significant improvement in interest, p21WAF1/CIP1 protein levels also increased, associated with survival of 24.5 months versus 16.0 months (HR 0.47; 95% CI; an increase in acetylation of H4 at the p21 promoter (using chro- p = 0.0001).117 matin immunoprecipitation). Treatment with entinostat led to The early development of decitabine in MDS took place pri- increased acetylation of H3 and H4 in both peripheral blood and marily in Europe under the leadership of Wijermans et al.118,119 bone marrow. This increase was detectable within 8 hours and These investigators pursued intravenous scheduling of decitabine remained above baseline throughout the treatment cycle. Thus, administered three times daily for 3 days (45 mg/m2 per day total this compound may provide the most prolonged inhibition of dose). This cycle was repeated every 6 weeks. Phase 2 studies sug- protein deacetylation of HDACis and is under current investiga- gested a response rate of approximately 50% in MDS patients. In tion.107 Increases in p21WAF1/CIP1 and activation of caspase 3 were a randomized trial of DAC versus observation, patients with Inter- also demonstrated in these samples. national Prognostic Score risk categories intermediate 1 to high received the previously listed schedule of decitabine or observa- tion. No crossover was allowed in this trial. Response rates reported were: complete response: 9%, partial response: 8%, and hemato- EPIGENETIC THERAPY FOR HEMATOLOGIC 120 MALIGNANCIES logic improvement: 13%. A 10% induction death rate occurred, suggesting that this schedule of DAC may be more toxic than the CALGB schedule of 5AC (1% induction mortality). DAC has also DNA Methyltransferase Inhibitors been investigated in low-dose daily intravenous dosing121 and in daily-times-five schedules. The latter appears convenient and well Epigenetic therapy has seen the most widespread use to date and tolerated. A daily-times-five schedule (20 mg/m2 per day) has been achieved the greatest efficacy in hematologic malignancies. The FDA approved121; 99 patients with MDS (de novo or secondary) of therapeutic efficacy of 5AC and DAC for patients with the chronic any French-American-British (FAB) subtype and an International myeloid neoplasm myelodysplasia (MDS) and AML has been well Prognostic Scoring System (IPSS) score equal to or greater than reviewed.26,27 Their FDA approval for MDS/AML emerged only 0.5 were treated, with an overall response rate of 32% (17 complete after doses were reduced, with resultant diminishing toxicities for responses [CR] plus 15 marrow CRs [mCR]).122 Among patients patients. The successful development of 5AC for the treatment who improved, 82% demonstrated responses by the end of cycle of MDS can be credited largely to Silverman et al.25,112,114 in the two. This well-tolerated regimen allows outpatient administration Cancer and Leukemia Group B (CALGB). The inhibitor 5AC had and, as noted previously, provides plasma levels of decitabine that successfully induced the expression of hemoglobin F in patients inhibit DNMTs. with sickle cell anemia.25,112 Viewing this compound as a poten- The 3-day intravenous schedule of DAC has been studied in tial inducer of terminal differentiation, Silverman et al. conducted two randomized trials compared to supportive care in patients with a series of phase 2 trials of 5AC administered as a continuous higher risk MDS. The first trial confirmed the hematologic activ- intravenous infusion or as subcutaneous injections for the treat- ity of decitabine in this patient population but failed to show an ment of MDS.113,114 Based on significant hematologic responses, improvement in survival in the DAC-treated patients.123 Survival the group performed a phase 2 trial (CALGB 9221) in which pa- was also not increased in the subsequent trial, performed by the tients with low- and high-risk MDS with significant hematopoietic European Organization for Research and Treatment of Cancer compromise were randomly assigned to receive subcutaneous 5AC (EORTC).124 The failure of the randomized decitabine trials to (75 mg/m2 per day daily for 7 days, repeated on a 28-day cycle) or show a survival benefit may be partially due to study design. Both observation. Patients on the observation arm with progressive dis- randomized trials of 5AC continued treatment until disease pro- ease could cross over to receive 5AC. This study firmly established gression for patients who did not achieve complete remission; in the ability of 5AC to induce hematologic improvement, and, less fact, this meant that most patients received maintenance therapy. frequently, complete and partial responses.113,115 The median In contrast, both randomized trials of decitabine allowed a maxi- time to development of AML (defined by 30% bone marrow blast mum of eight cycles of treatment. The need for maintenance cells) or death was greater in the 5AC arm by 9 months (21 versus therapy in patients treated with DNMT inhibitors has not been 12 months); of note, the observation arm included patients who tested in prospective randomized trials. An additional difference in subsequently crossed over to 5AC treatment. the conduct of the two sets of DNMT inhibitor trials involves the In a subsequent phase 3 trial (AZA001),116 patients with higher duration of therapy administered. The median number of cycles of risk myelodysplastic syndromes were randomly assigned one-to- treatment administered in the two randomized trials of decitabine one to receive 5AC (75 mg/m2 per day for 7 days every 28 days) was three, compared to nine in the azacytidine trials. This may re- or conventional care (best supportive care, low-dose cytarabine, flect greater toxicity of the originally 3-day schedule of decitabine or intensive chemotherapy as selected by investigators before compared to that of the approved schedule of 5AC. Although randomization). Three hundred fifty-eight patients were ran- the differences in survival may reflect differences in trial design domly assigned to receive 5AC (n = 179) or conventional care and trial conduct, emerging data suggests that despite similarities regimens (n = 179). After a median follow-up of 21.1 months in methylation reversal, the two drugs differ in other potentially (interquartile range [IQR] 15.1 to 26.9), median overall survival important biologic parameters, which may contribute to clinical was 24.5 months (9.9 not reached) for the azacitidine group versus outcomes.62,63,125 15.0 months (5.6 to 24.1) for the conventional care group (haz- Two randomized phase 3 trials have been published treating el- ard ratio [HR] 0.58; 95% confidence interval [CI], 0.43 to 0.77; derly AML patients (greater than 20% blasts) with decitabine, both p = 0.0001). At 2 years, on the basis of Kaplan-Meier estimates, demonstrating improvement in survival that was not statistically 50.8% (95% CI, 42.1 to 58.8) of patients in the 5AC group were significant. In the European study, 233 patients received either

tahir99 - UnitedVRG Chapter 23 Histone Deacetylase Inhibitors and Demethylating Agents 255

DAC at 15 mg/m2 × 9 doses over 3 days on 42-day cycles or best key biologic roles for this gene and its low basal expression are supportive care. The patients received a median of 4 cycles (0 to probable. Moreover, in one study, the clinical response was closely 9), and the overall survival was improved in the decitabine-treated associated with the reversal of methylation of p15 or CDH-1 dur- patients, but did not reach statistical significance (median overall ing the first cycle of treatment with 5AC followed by the HDACi survival [OS], 10.1 versus 8.5 months, respectively; HR, 0.88; 95% NaPB.25 In that study, it was noteworthy that the administration CI, 0.66 to 1.17; two-sided, log-rank p = 0.38).126 In the M.D. of 5AC prior to the addition of an HDACi was associated with Anderson Cancer Center–led multicenter trial,127 485 patients 65 the induction of histone acetylation. Although the mechanism years or older were randomly assigned to receive decitabine 20 underlying this activity is unknown, histone acetylation has been mg/m2 per day as a 1-hour intravenous infusion for 5 consecutive observed following DNA damage due to gamma irradiation.136 days every 4 weeks or best supportive care or low-dose cytarabine Subsequent studies have found demethylation following treatment (20 mg/m2 per day for 10 days every 4 weeks). There was a simi- with either DAC or 5AC137–140 but not consistently associated with lar improvement in OS with decitabine (7.7 months; 95% CI, 6.2 response.137,138,140 More work will be required to answer the im- to 9.2) versus the control group (5.0 months; 95% CI, 4.3 to 6.3; portant mechanistic question underpinning the clinical activity of p = 0.108; HR, 0.85; 95% CI, 0.69 to 1.04).127 azacytosine analogs. The azacytosine nucleosides require prolonged administration to demonstrate hematologic improvement in MDS. Median time to development of first clinical response in the CALGB studies Combining Inhibitors in the Treatment of of 5AC was three cycles; 90% of responses developed by cycle Hematologic Malignancies six.114 In the phase 3 trial of decitabine, the median time to re- sponse was two cycles,123 as also seen in the alternative regimen of It is almost certain that the biggest promise of epigenetic therapy decitabine.122 It is, therefore, extremely important when treating lies in strategies to combine existing and newer drugs with each patients with azacytosine nucleosides to commit to administering other and with current and targeted therapies. To between four and six cycles of therapy before determining whether date, the example for existing agents is the combination of DNMT a patient is responding to treatment. Furthermore, survival benefit inhibitors and HDAC inhibitors based on the hypothesis from the is seen even in patients not showing bone marrow improvement laboratory that this paradigm leads to optimal reexpression of tran- for 5AC, perhaps related to decreased transfusion requirements or scriptionally silenced genes with promoter methylation.56,141 This delayed progression to AML.116 in vitro treatment paradigm has led to a variety of clinical stud- Because AML in the context of MDS is arbitrarily defined ies that have attempted to apply this concept to the treatment of based on marrow blast count, activity of the azanucleoside ana- hematologic malignancies. Much remains to be determined with logs in AML should not be surprising. In CALGB 9221, 20 pa- regard to its efficacy and precisely what determines this. The first tients were reclassified upon central pathology review as meeting study of sequential DNMT/HDAC inhibitors administered a vari- criteria for AML (greater than 30% blasts). Their outcomes were ety of doses of 5AC for 5 to 14 days followed by 7 days of NaPB by CANCER THERAPEUTICS comparable to the overall population in the study.115 In all three continuous infusion at its maximum tolerated dose to patients with CALGB studies among patients meeting current World Health MDS and AML.25 The combination was well tolerated, and clini- Organization (WHO) criteria for AML (greater than 20% blasts), cal responses were frequent in patients receiving 5AC at 50 mg/m2 a complete response was achieved in 9% and hematologic im- per day daily for 10 days and 25 mg/m2 per day daily for 14 days, provement in 26%.114 A retrospective review of 20 patients with with 5 of 14 patients at those dose schedules achieving complete AML, including 8 patients with bone marrow blasts greater than or partial response. 29% treated with 5AC, reported a complete remission in 4 pa- In a pilot study, 10 patients with MDS or AML were treated tients, a partial response in 5, and a hematologic improvement with 5AC at 75 mg/m2 per day daily times seven followed by 5 days in 3. The median duration of response was 8 months (range: 3 to of NaPB given at 200 mg per kilogram per day as a 1- to 2-hour 33 months).128 DAC induced a complete hematologic response infusion. Three patients developed a partial response.142 in 2 of 20 patients treated who had the blastic phase of chronic In a similar study, investigators at the M.D. Anderson Cancer myeloid leukemia.129 These studies suggest activity of the azacy- Center treated leukemic patients with decitabine (15 mg/m2 per tosine nucleosides in the treatment of a subset of AML patients. day IV daily times 10) and concomitant VPA at a variety of doses. Current studies do not allow for the determination of whether Of 54 patients, 12 achieved complete remission or complete re- this subset is limited to MDS-associated AML (AML with MDS- mission with incomplete platelet recovery.143 The inhibitors 5AC, related changes), which tends to have low white blood cell VPA, and all-transretinoic acid have been administered to patients counts and have a low proliferative rate, or whether these com- with AML and MDS. Of 33 previously untreated patients, 14 over pounds are also active for those with AML without a history of the age of 60 years developed a complete remission or a complete antecedent hematologic disorder. Several reports describe the remission with inadequate platelet recovery.144 A subsequent study sensitivity MDS and AML, characterized by abnormalities of of 5AC and VPA suggests increased efficacy of this combination in chromosome 7 and associated with poor outcomes in response high-risk MDS.145 to cytarabine-based therapy to azanucleosides. In one nonran- Entinostat has been successfully combined with azacytidine domized retrospective study, survival of such patients following in patients with myeloid malignancies.140 The US Leukemia the administration of DNMT inhibitors surpassed survival in re- Intergroup recently completed a randomized phase 2 trial of sponse to conventional cytotoxic chemotherapy, similar to the this combination compared with 5AC alone. In this study, of outcomes of AZA001.130–132 149 patients, the primary endpoint of hematologic normaliza- Although the mechanisms underlying the clinical activity of tion was statistically similar, with 32% (95% CI, 22% to 44%) of azacytosine analogs may involve reversal of gene methylation, the 5AC group reaching hematologic normalization (HN) ver- other actions need to be considered. The administration of DAC sus 27% (95% CI, 17% to 39%) in the AZA + entinostat group. has been shown to induce transient decrements of methylation in Median overall survivals were 18 months for the AZA group and noncoding regions, including long interspersed nuclear element 13 months for the AZA + entinostat group, but were also not (LINE) and ALU elements.133 Early studies that examined meth- statistically significant.108 In the latter study, the administration ylation reversal of the target gene p15INK4B in response to DAC of the combination was associated with less DNA methylation showed no correlation between methylation reversal and clinical reversal compared to azacitidine monotherapy, likely due to response.134,135 Clinical responders to DAC developed significantly cell cycle inhibitory effects of the HDACi. This highlights the higher expression of this gene following treatment, and certainly complexity of effectively targeting epigenetic gene regulation. 256 Cancer Therapeutics

It remains to be established whether combination therapies are uridine monophosphate (UMP) kinase, in colorectal cancer more effective than single-agent demethylating therapies. cell lines.152 These increases correlated with a reversal of 5FU resistance. Similar to studies discussed previously, DAC plus the Epigenetically Targeted Therapy in HDACi, trichostatin A, decreased marker identified self-renewal populations in ovarian cancer while simultaneously inducing Nonhematologic Malignancies increased sensitivity to cisplatin.153 In advanced ovarian cancer, 5AC or DAC plus carboplatin have yielded durable responses The efficacies that have emerged in the application of epigeneti- and induced stable disease in ovarian cancer patients.154,155 cally targeted drugs to hematologic malignancies has spurred in- These early results are being extrapolated for verification in terest in using epigenetic therapy for other types of cancer. As larger, ongoing clinical trials. outlined as follows, laboratory studies and clinical trials support this approach. Studies in the lab have been directed by lessons learned from therapy in hematologic malignancies, suggesting that low doses of drugs like DAC and 5AC, in the nanomolar NEW APROACHES TO EPIGENETIC range, may avoid excess toxicities due to off-target effects of the THERAPY drugs and may maximize epigenetic effects of the agents.28 The desired effects may require minimizing initial cellular cytotox- As we have outlined previously, the emerging promise for epige- icity, giving tumor cells time to accrue maximal cellular repro- netic therapy and the future of the approaches may lie in combina- gramming responses to the inhibition of DNMTs.28 DAC and torial drug strategies. Although this is already being explored with 5AC are effective only when they have been incorporated into older agents, new drugs for new targets are now entering the pic- DNA, after which they irreversibly inhibit DNMT catalytic activ- ture.9,11,156–158 In these efforts, several themes we have introduced ity and target these proteins for degradation.64–68 In cell culture in this chapter will likely dominate. and mouse explants, low nanomolar doses appear to induce both Most epigenetic therapies will not induce, when used at truly human leukemic and solid tumor cells to exhibit blunting of self- targeting doses, immediate cytotoxic effects. Therapeutic efficacy renewal and tumorigenic activity of tumor stem-like cells.28 These based on cellular reprogramming may require significant time preclinical results suggest a key possibility that use of epigenetic to manifest. Clinical trial designs may need adaptation so that therapies might inhibit these latter cell populations, which often effective therapies are not discarded due to premature response are difficult to eradicate and are a factor in resistance to many evaluations. Finally, the ultimate promise for epigenetic ther- standard cancer therapies.146 Exhaustion of such cells over time apy may lie with newer drugs now entering clinical trials. Out- during therapy with DAC or 5AC might explain the observation comes with DNMT inhibitors may be improved with alternative that most patients with MDS/AML take several months to reach scheduling of oral azacitidine or through prolonged pharmacoki- best response.147 Leukemic stem cells were not eliminated in one netics of the decitabine prodrug SGI110.79 Also, drugs targeting study in MDS and AML patients treated with 5AC in combina- other proteins including BET family bromodomain proteins are tion with VPA, although their frequency decreased in clinical generating much excitement.9,82,156–159 BET inhibitors may inter- responders.148 fere with localization of the oncogene C-MYC to acetylated lysines Clinical trials for common solid tumors, informed through in regulatory regions of target genes.9,82,156–159 These inhibitors are the previous laboratory studies, have been initiated includ- now entering clinical trials. Other promising approaches include ing phase 2 designs using low-dose strategies with 5AC often the use of inhibitors of EZH2, the enzyme in the PcG system, combined with use of histone deacetylase inhibitors. Sixty-five which catalyzes the repressive histone mark H3K27me3.9,82,156–159 patients with advanced, multiply treated NSCLCs were treated Another clinical trial underway employs targeting of the transloca- with 5AC plus entinostat.149 Only 3% of patients developed Re- tion in which the protein mixed lineage leukemia (MLL) is fused sponse Evaluation Criteria (RECIST)-measureable responses; with several targets, such as in infant leukemias. These transloca- however, these two patients had durable responses, with survival tions result in abnormal recruitment of the histone methyltransfer- of 3 to 4 years.149 Upregulation of immunogenic pathways in ase, DOT1L, to target genes like HOXA9.158 This fusion induces NSCLC and other solid tumor cells, observed in laboratory stud- hypermethylation of H3K79 and abnormal activation of MLL ies, suggest a potential for sequencing DNMT inhibitors with target genes.158,160 Very selective inhibitors of DOT1L are now in immune checkpoint inhibitors.150 This drug is also reported clinical trials. to induce antitumor responses and immune recognition in Epigenetically targeted therapies continue to hold great prom- a model of pancreatic cancer.151 Other laboratory results and ise that reprogramming of malignant cells could alter approaches emerging clinical trials also suggest the promise of combining to cancer management. Strategies to merge older drugs, which epigenetic therapy approaches to sensitize cancers other than we have focused on in this chapter, with the newer agents briefly NSCLC to subsequent therapies. Low-dose DAC appears able discussed in this section, will underpin future trials to test this to upregulate a key mediator of 5-fluorouracil (5FU) action, approach.

SELECTED REFERENCES

The full reference list can be accessed at lwwhealthlibrary.com/oncology. 18. Easwaran H, Johnstone SE, Van Neste L, et al. A DNA hypermethylation module for the stem/progenitor cell signature of cancer. Genome Res 4. Herman JG, Baylin SB. Gene silencing in cancer in association with pro- 2012;22:837–849. moter hypermethylation. N Engl J Med 2003;349:2042–2054. 25. Gore SD, Baylin S, Sugar E, et al. Combined DNA methyltransferase and 5. Jones PA, Baylin SB. The epigenomics of cancer. Cell 2007;128:683–692. histone deacetylase inhibition in the treatment of myeloid neoplasms. Can- 7. Esteller M. Cancer epigenomics: DNA methylomes and histone-modification cer Res 2006;66:6361–6369. maps. Nat Rev Genet 2007;8:286–298. 26. Azad N, Zahnow CA, Rudin CM, et al. The future of epigenetic therapy in 8. Yoo CB, Jones PA. Epigenetic therapy of cancer: past, present and future. solid tumours—lessons from the past. Nat Rev Clin Oncol 2013;10:256–266. Nat Rev Drug Discov 2006;5:37–50. 28. Tsai HC, Li H, Van Neste L, et al. Transient low doses of DNA-demethy- 14. Kouzarides T. Chromatin modifications and their function. Cell 2007;128: lating agents exert durable antitumor effects on hematological and epithelial 693–705. tumor cells. Cancer Cell 2012;21:430–446.

tahir99 - UnitedVRG Chapter 23 Histone Deacetylase Inhibitors and Demethylating Agents 257

38. Jones PA. Functions of DNA methylation: islands, start sites, gene bodies and 59. Jones PA, Taylor SM. Hemimethylated duplex DNAs prepared from beyond. Nat Rev Genet 2012;13:484–492. 5-azacytidine-treated cells. Nucleic Acids Res 1981;9:2933–2947. 39. Bannister AJ, Kouzarides T. Reversing histone methylation. Nature 2005; 64. Kelly TK, De Carvalho DD, Jones PA. Epigenetic modifications as therapeu- 436:1103–1106. tic targets. Nat Biotechnol 2010;28:1069–1078. 40. Bannister AJ, Kouzarides T. Regulation of chromatin by histone modifica- 65. Ferguson AT, Vertino PM, Spitzner JR, et al. Role of estrogen receptor gene tions. Cell Res 2011;21:381–395. demethylation and DNA methyltransferase. DNA adduct formation in 5-aza- 48. Cai Y, Geutjes EJ, de Lint K, et al. The NuRD complex cooperates with 2′deoxycytidine-induced cytotoxicity in human breast cancer cells. J Biol DNMTs to maintain silencing of key colorectal tumor suppressor genes. Chem 1997;272:32260–32266. Oncogene 2014;33:2157–2168. 66. Gabbara S, Bhagwat AS. The mechanism of inhibition of DNA (cytosine-5-)- 49. Bestor TH. Cloning of a mammalian DNA methyltransferase. Gene 1998; methyltransferases by 5-azacytosine is likely to involve methyl transfer to the 74:9–12. inhibitor. Biochem J 1995;307:87–92. 50. Jair KW, Bachman KE, Suzuki H, et al. De novo CpG island methylation in 69. Rountree MR, Bachman KE, Baylin SB. DNMT1 binds HDAC2 and a human cancer cells. Cancer Res 2006;66:682–692. new co-repressor, DMAP1, to form a complex at replication foci. Nat Genet 52. Jenuwein T, Allis CD. Translating the histone code. Science 2001;293: 2000;25:269–277. 1074–1080. 70. Bachman KE, Rountree MR, Baylin SB. Dnmt3a and Dnmt3b are 55. Pruitt K, Zinn RL, Ohm JE, et al. Inhibition of SIRT1 reactivates silenced transcriptional repressors that exhibit unique localization properties to cancer genes without loss of promoter DNA hypermethylation. PLoS Genet heterochromatin. J Biol Chem 2001;276:32282–32287. 2006;2:344–352. 75. Gore SD, Weng LJ, Figg WD, et al. Impact of prolonged infusions of 56. Cameron EE, Bachman KE, Myohanen S, et al. Synergy of demethylation the putative differentiating agent sodium phenylbutyrate on myelodys- and histone deacetylase inhibition in the re-expression of genes silenced in plastic syndromes and acute myeloid leukemia. Clin Cancer Res 2002;8: cancer. Nat Genet 1999;21:103–107. 963–970. CANCER THERAPEUTICS 24 Proteasome Inhibitors

Christopher J. Kirk, Brian B. Tuch, Shirin Arastu-Kapur, and Lawrence H. Boise

BIOCHEMISTRY OF THE PROTEASOME INHIBITORS UBIQUITIN-PROTEASOME PATHWAY Chemical Classes of Proteasome Inhibitors in The ubiquitin proteasome system is involved in the degrada- tion of more than 80% of cellular proteins, including those Clinical Development that control cell-cycle progression, apoptosis, DNA repair, and the stress response.1 A key step in this process is the tagging of As of the writing of this overview, six different proteasome in- proteins targeted for degradation with multiple copies of ubiq- hibitors comprising three distinct chemical classes have been tested in clinical trials (Table 24.1) and include: (1) dipeptide uitin, a 76–amino acid protein whose primary sequence and β 18,19 structure is highly conserved in organisms ranging from yeasts boronic acids, (2) peptide epoxy ketones, and (3) -lactones. to mammals.2,3 Once polyubiquitinated, proteins targeted for Bortezomib (PS-341, Velcade), a dipeptide boronic acid, was developed by Millennium Pharmaceuticals (Cambridge, MA) degradation bind to the 26S proteasome, a holoenzyme com- 20 posed of two 19S regulatory complexes capping a central 20S and was the first PI approved for clinical use. Two additional proteolytic core. The 20S core is a hollow “barrel” consisting dipeptide boronic acids have entered clinical development, ixa- of four stacked heptameric rings. The subunits of the rings are zomib/MLN 9708 ( Millennium), currently in phase III studies, classified as either β subunits (outer two rings) or β subunits and delanzomib/CEP-18770 (Teva Pharmaceuticals; Frazer, PA), (inner two rings). The 19S regulatory complex consists of a lid the clinical development of which has been halted. that recognizes ubiquitinated protein substrates with high fi- (Onyx Pharmaceuticals; San Francisco, CA), a tetrapeptide epoxy ketone, received U.S. Food and Drug Administration (FDA) ap- delity, and a base that contains six adenosine triphosphatases, 21 unfolds protein substrates, removes the polyubiquitin tag, and proval in 2012. A second peptide epoxy ketone proteasome in- hibitor, (Onyx), entered clinical study in 2010. The threads them into the catalytic chamber of the 20S particle in β an adenosine triphosphate–dependent manner.4,5 Unlike typi- third class of proteasome inhibitors, -lactones, is represented by cal proteases, the 20S proteasome in eukaryotic cells contains NPI-0052 (salinosporamide A [Marizomib]) and is currently being multiple proteolytic activities resulting in the cleavage of pro- developed by Nereus Pharmaceuticals, Inc. (San Diego, CA). tein targets after many different amino acids. In most cells, the The initial approvals for both bortezomib and carfilzomib were in 20S core particle contains the catalytic subunits β5 (PSMB5), multiple myeloma (MM), a plasma cell neoplasm and the second β1 (PSMB1), and β2 (PSMB2), accounting for chymotrypsin- most common hematologic cancer. However, the activity of PIs in like (CT-L), caspaselike (C-L), and trypsinlike (T-L) activities, other B-cell neoplasms has resulted in an expansion of the clinical respectively, each differing in their substrate preference.6 How- utilization of this drug class. ever, in cells of hematopoietic origin, such as lymphocytes and monocytes, the proteasome catalytic subunits are encoded by Preclinical Activity of Proteasome Inhibitors homologous gene products: LMP7 (PSMB8), LMP2 (PSMB9), and MECL-1 (PSMB10).7 These immunoproteasome subunits Each of the three classes of inhibitors has a distinct chemical mech- are also induced in nonhematopoietic cells following exposure anism of proteasome inhibition.22 Peptide boronates form stable to inflammatory cytokines such as interferon-γ (IFN-γ) and but reversible tetrahedral intermediates with the γ-hydroxyl ( γ-OH) tumor necrosis factor alpha (TNF-α).8 In the immunoprotea- group of the catalytic N-terminal threonine of the proteasome ac- some, the 19S regulatory complex can be replaced with protea- tive sites.23,24 β-lactones also interact with this γ-OH, but form a some activators such as PA28, whose expression is also induced completely irreversible interaction.25 Similarly, peptide epoxy ke- in cells following exposure to IFN-γ. Hybrid proteasomes, both tones form irreversible covalent adducts with the threo- for the catalytic subunits and regulatory particles, have been nine but do so via a dual covalent adduction of γ-OH group and described.9 the free amine.26 This interaction is highly specific for N-terminal Given its key role in maintaining cellular homeostasis, the threonine-containing and renders peptide epoxy ke- ubiquitin proteasome system appeared to be an unlikely target for tones the most selective proteasome inhibitors yet described.27,28 pharmaceutical intervention. However, a variety of groundbreak- The primary targets of these PIs within the constitutive and ing studies in the 1990s suggested that inhibitors of proteasome immunoproteasomes are the CT-L subunits, β5 and LMP7, re- function might prove to be viable therapeutic agents.10 Initial spectively. Despite accounting for less than 50% of total protein studies used substrate-related peptide aldehydes to investigate the turnover by the proteasome, these subunits are essential for cell sur- proteolytic functions and specificity of the proteasome.11 In vitro vival.29 In MM cell lines, inhibiting both subunits (β5 and LMP7) and in vivo studies with these inhibitors demonstrated their abil- is necessary and sufficient for tumor cell death.30 Cytotoxicity of ity to induce apoptosis as well as inhibit tumor growth.12–15 It was other tumor cell types requires the inhibition of multiple active subsequently discovered that several natural products with antitu- sites beyond the CT-L activity. The combination of inhibitors mor activity exert their action via proteasome inhibition, providing specific for either the T-L or C-L activities, which have no cyto- additional rationale for the development of selective proteasome toxic activity on their own, augments the cytotoxic potential of the inhibitors (PIs).16,17 CT-L–specific inhibitors.31,32

258

tahir99 - UnitedVRG Chapter 24 Proteasome Inhibitors 259

TABLE 24.1 Proteasome Inhibitors in Clinical Development

Drug Stage of Route of Schedule of Agent Other Names Class Development Tumor Types Administration Dose Levels Administration Bortezomib Velcade Peptide FDA/EMEA Multiple Intravenous, 1.3 mg/m2 Days 1, 4, 8, & PS-341 boronate approved myeloma, subcutaneous 11 (21-day cycle) mantle cell lymphoma MLN 9708 Peptide Phase III Multiple Oral 4 mg Once weekly MLN 2238 boronate myeloma, AL, (21-day cycle) amyloidosis Delanzomib CEP-18770 Peptide Phase I Multiple Intravenous 0.1–1.8 mg/m2 Days 1, 4, 8, & boronate (discontinued) myeloma 11 (21-day cycle) Carfilzomib Kyprolis Peptide FDA approved, Multiple Intravenous 20/27 mg/m2 Days 1, 2, 8, 9, PR-171 epoxy phase III myeloma 15, & 16 ketone (28-day cycle) Oprozomib ONX 0912 Peptide Phase I/II Multiple Oral 150–240 mg Days 1, 2, 8, & PR-047 epoxy myeloma (dose escalation 9 (14-day cycle) ketone ongoing) Days 1–5 (14- day cycle) Marizomib NPI-0052 β-lactone Phase II Multiple Intravenous 0.075–0.6 mg/m2 Days 1, 4, 8, & Salinosporamide A myeloma 11 (21-day cycle)

EMEA, European Medicines Agency; AL, amyloid light chain.

Given its status as the first approved for mammalian cyclins A, B, D, and E; and transcription factors E2F marketed use, the antitumor potential and preclinical activity of and Rb.49,50 The transcription factor nuclear factor kappa B (NF- CANCER THERAPEUTICS other proteasome inhibitors have generally been compared to κB), an important regulator of cell survival and cytokine/growth bortezomib.19 Carfilzomib showed equivalent antitumor activity factor production,51 is also affected by proteasome inhibition in to bortezomib in vitro against a panel of tumor cell lines under multiple ways. The net effect on NF-κB signaling is not consis- standard culture conditions but was >10-fold more potent at in- tent across various assays and cell lines, and its relative importance ducing tumor cell death when cells were exposed to drug for a in the antitumor effects of PIs remains unclear. Although it is 1-hour pulse, which mimics the pharmacokinetics of both com- interesting to note that patients whose myeloma harbor NF-κB– pounds.33 MLN2238 (the active agent of ixazomib) was active in activating mutations (∼20%) respond better to bortezomib than the same mouse models of human tumors as bortezomib, but dem- those without NF-κB–activating mutations.52–54 In MM cell lines, onstrated greater levels of proteasome inhibition in the tumors.34 there is growing evidence that the major determinant of sensitivity In biochemical assays of proteasome activity, delanzomib had an to proteasome inhibition is the relative load of protein flux to the identical potency and subunit activity profile to bortezomib, but proteasome.55–57 These data suggest that induction of the terminal in tumor cytotoxicity assays, potency relative to bortezomib was unfolded protein response may drive cell death. Whether proteo- 2- to 10-fold less.35 In addition, delanzomib appeared to be less toxic stress induced cell death reflects sensitivity to proteasome in- cytotoxic than bortezomib to normal cells and had a differential hibitors in other tumor types remains to be determined. effect on cytokine release in bone marrow stromal cells, suggest- ing a different pharmacologic activity. Oprozomib is 10-fold less potent than carfilzomib in proteasome activity assays, but showed Pharmacokinetics and Pharmacodynamics of similar antitumor activity in mouse tumor models.36,37 Marizomib Proteasome Inhibitors in Animals displayed greater potency against the non–CT-L active sites of the proteasome than bortezomib.38 Interestingly, this agent synergized Following intravenous (IV) administration to animals and hu- with bortezomib in killing tumor cells in vitro.39 All of the second- mans, proteasome activity is inhibited in a dose-dependent fashion generation inhibitors have shown activity in tumor cells made within minutes; however, PIs such as bortezomib and carfilzomib resistant to bortezomib and/or MM cells isolated from patients are also rapidly cleared from circulation.55,56,58–61 Recovery of pro- relapsed from bortezomib-based therapies35,36,40–42 teasome activity in animals occurs in tissues with a half-life of ap- The inhibition of tumor cells with proteasome inhibitors in- proximately 24 hours, mirroring the recovery time of cells exposed duces cell death via the induction of apoptosis through death effec- to sublethal concentrations of PIs in vitro and likely reflecting new tor caspase activation.10 Although the mechanism underlying the protein synthesis.33,62 induction of cell death remains to be fully elucidated, extensive re- search suggests a complex interplay of multiple pathways. PIs have been shown to affect the half-life of the BH3-only members of the PROTEASOME INHIBITORS IN CANCER Bcl-2 family, specifically BH3–interacting-domain death agonist (Bid) and Bcl-2 interacting killer (Bik).43 Moreover the BH3-only protein NOXA is upregulated at the transcription level by PIs.44–48 Clinical Activity of Bortezomib Proteasome inhibition also upregulates the expression of several key cell-cycle checkpoint proteins that include p53 (an inducer Bortezomib is typically administered on days 1, 4, 8, and 11 of a of G0/G1 cell-cycle arrest through accumulation of the cyclin- 3-week cycle either as an IV bolus or subcutaneous administra- dependent kinase [CDK] inhibitor p27); the CDK inhibitor p21; tion. Increasing doses of bortezomib inhibit proteasome activity in 260 Cancer Therapeutics blood in a dose-dependent fashion, reaching a maximum of 74% to weekly bortezomib administration, which yielded similar TTP inhibition at a dose of 1.38 mg/m2. Daily dosing schedules in ani- but reduced the incidence (21% versus 43%) and severity of PN mal studies have been associated with severe toxicity and have not (2% grade 3/4 versus 14%).64 The bortezomib, lenalidomide, and been attempted in humans. In clinical trials, thrombocytopenia and dexamethasone combination in newly diagnosed MM resulted in peripheral neuropathy (PN) were common adverse events.20,63,64 a ORR of 100% in 66 patients, 29% of whom achieved a CR.76 Bortezomib has shown remarkable single-agent antitumor activity Bortezomib has also shown activity in other hematologic cancers, in a wide range of B-cell neoplasms, including MM, non Hodgkin most notably mantle cell lymphoma (MCL).77,78 As a single agent lymphoma (NHL), and Waldenström macroglobulinemia (WM). In in 155 relapsed and refractory MCL patients, bortezomib yielded 2003, bortezomib was approved by the FDA for use as a single agent an ORR of 33% (8% CR), a median duration of response of 9.2 for the treatment of patients with MM following two prior therapies months, and a TTP of 6.2 months.78 Toxicities observed were similar and who demonstrated disease progression with their most recent to those seen in patients with MM and included thrombocytopenia, therapy. The primary efficacy data for this approval was derived from PN, and fatigue. When bortezomib was used to treat both newly the SUMMIT trial in which 202 patients with heavily pretreated dis- diagnosed and refractory MCL, a response rate of 46% was observed ease were treated with bortezomib at 1.3 mg/m2.65 In this trial, the in both populations,77 leading to FDA approval late in 2006. overall response rate (ORR), defined as patients achieving at least a Bortezomib has been tested in a variety of solid tumors in phase 50% reduction in serum or urine levels of the myeloma M protein, I and II studies.79 Partial responses (PR) were reported in 8% of was 35%. This clinical trial was supported by the CREST trial, in patients with refractory non–small-cell lung cancer (NSCLC), which the activity of 1.3 mg/m2 dose was determined to be superior although the TTP was 1.5 months.80 Exacerbation of PN was to a dose of 1.0 mg/m2.66 Bortezomib is also active as a single agent common. Bortezomib was subsequently tested in combination in earlier stage MM patient populations. A single-agent ORR of 38%, with paclitaxel, irinotecan, and gemcitabine/carboplatin; how- with a 6% complete response (CR) rate, was seen in the phase III ever, results have not been encouraging. Bortezomib continues to APEX study in early relapsed MM, with a time to progression (TTP) be tested in combination with other agents in a variety of tumor of 6.2 months and a median duration of response of 8 months.67 In types.81,82 this study, the major grade 3 and 4 toxicities were PN, 12%; dysesthe- Recent clinical activity and preclinical data suggest that protea- sia and related symptoms, 8% to 10%; anemia, 8%; diarrhea, 8%; neu- some inhibition may extend to nononcology applications. Single- tropenia, 14%; and fatigue, 12%. In the frontline setting, bortezomib agent bortezomib therapy in kidney transplant patients undergoing demonstrated a single-agent response rate of 41% (5% CR rate).68 antibody-mediated rejection resulted in a reduction of donor-spe- Bortezomib is also approved for newly diagnosed MM in com- cific antibodies and improved renal function.83 In mouse models of bination with velcade, and prednisone (VMP). The lupus nephritis, bortezomib resulted in a reduction of pathogenic phase III VISTA trial evaluated VMP in patients with untreated plasma cells and the prevention of disease progression.84 These data MM who were ineligible for high-dose therapy.69 The addition of suggest that PIs may be useful in a wide range of B-cell–mediated bortezomib to the melphalan prednisone (MP) backbone signifi- diseases. However, toxicities with bortezomib, particularly PN, may cantly improved response rates in this setting with an ORR of 71% prevent wider application of this particular agent. for VMP (including 30% CR) versus 35% (with only 4% CR) for MP.52 VMP was associated with a TTP of ∼24 months, compared with ∼16.6 months with MP. After a 5-year follow-up, there was a Carfilzomib 31% reduced risk of death for the VMP group versus MP-treated patients.70 Parallel phase I studies of carfilzomib have been conducted in Bortezomib has also shown promise when combined with other patients with multiple tumor types, and two phase I dose-finding agents in relapsed and refractory MM patients. The combination studies targeting B-cell malignancies have been completed. The of bortezomib with pegylated doxorubicin (Doxil, Centocor Ortho first study used daily IV bolus dosing with doses up to 20 mg/m2 Biotech Products, L.P.; Horsham, PA) resulted in an ORR of 79% in for 5 consecutive days followed by 9 days of rest and resulted in relapsed patients, and toxicities were similar to those observed with substantial inhibition of proteasome activity.85 In the second each agent administered separately.71 A phase III study in 646 pa- study, carfilzomib was administered daily for 2 days for 3 consecu- tients with relapsed and refractory MM compared this treatment tive weeks (days 1, 2, 8, 9, 15, and 16), followed by 12 days of with bortezomib alone; the combination produced a 44% ORR and recovery.86 Hematologic toxicities were the most frequent adverse extended the TTP from 6 to 9.3 months.72,73 The combination of events, observed along with transient, noncumulative elevations bortezomib with revlimid, lenalidomide and dexamethasone (Rd), a in serum creatinine, usually with increases in serum urea nitro- standard of care in the treatment of MM, resulted in an ORR of 64% gen and consistent with a prerenal etiology. New onset PN was and a median duration of response of 8.7 months.74 This activity is infrequent. Among 20 evaluable patients (including bortezomib- striking given that 53% of patients had received prior bortezomib refractory patients), 4 PRs and 1 minor response were seen. Re- and 75% of patients had received prior thalidomide, a closely related sponses were also durable, lasting more than 1 year in some cases. analog of lenalidomide. Other agents tested in combination with Although the maximum tolerated dose of carfilzomib was not es- bortezomib include vorinostat, the anti-CS1 mAb, elotuzumab, the tablished in this study, a dose of 20 mg/m2 was initially selected for Hsp90 inhibitor tanespimycin, and the Akt inhibitor perifosine.75 the phase II studies. Frontline combinations with bortezomib in MM patients have Based on the phase I studies, an open-label, single-arm, phase shown high ORRs with a notable improvement in CR rates. In II study of single-agent carfilzomib in relapsed and refractory MM longer term studies, CR rates with bortezomib-based combina- was initiated in 2007.87,88 Carfilzomib was administered as an IV tions have been shown to be associated with improved clinical bolus on the twice-weekly dose schedule. Patients enrolled in the outcomes.63,64 A community-based phase IIIb study evaluating initial phase of the study (003-A0) had received a median of five bortezomib + dexamethasone (VD) versus bortezomib + thalido- prior therapies, and 78% of patients had grade 1/2 PN at entry.87 mide + dexamethasone (VTD) versus VMP found similar ORR Among 39 evaluable patients in 003-A0, 10 (26%) achieved a minor (60%, 70%, and 52%, respectively) and CR rates (13%, 18%, and response or better, including 5 PRs, and 16 additional patients 15%, respectively).63 Bortezomib + melphalan + prednisone + with stable disease. Based on new safety information from phase thalidomide (VMPT) followed by bortezomib + thalidomide I studies, the protocol was amended and the carfilzomib dose was (VT) maintenance resulted in a superior CR rate compared with escalated to 27 mg/m2 after the first cycle (003-A1).89 In this trial, VMP with no maintenance (34% versus 21%) and improved 266 patients were enrolled and all patients had previously been 2-year progression-free survival (70% versus 58.2%).64 A protocol treated with an immunomodulatory agent (IMiD) and bortezomib modification in this trial involved changing from twice weekly and were refractory to their last therapy. An ORR of 24% with a

tahir99 - UnitedVRG Chapter 24 Proteasome Inhibitors 261 median duration of response of 8 months was reported. Adverse weekly and twice weekly dosing schedules.96,97 Oral administration events were predominantly hematopoietic (thrombocytopenia, resulted in potent proteasome inhibition of ∼65%. Clinical ac- lymphopenia, and anemia) and there was a <1% rate of grade 3 tivity in patients with relapsed MM was 16%.98 In patients with PN, despite 77% having a history of PN. Based on these findings, newly diagnosed MM, ixazomib plus lenalidomide and low-dose carfilzomib was granted conditional approval by the FDA in 2012 dexamethasone resulted in an ORR of 93% with 24% achieving a for the treatment of patients with relapsed and refractory myeloma CR.99 This combination is also being investigated in a phase III who had received prior bortezomib and IMiD therapy. trial comparing this to Rd in patients with relapsed MM. The parallel PX-171-004 trial enrolled patients with relapsed MM following one to three prior treatments and who may have been refractory to one or more of these therapies.90,91 Of the Oprozomib 155 patients enrolled in this trial, 120 had not received prior bortezomib-based therapy. In patients with relapsed disease, non- Initial clinical testing of oprozomib in patients with solid tumors hematologic and hematologic toxicity profiles were similar. De- investigated a dosing schedule consisting of a 14-day cycle with 100 spite high rates of baseline PN, reports of worsening neuropathic once daily administration for 5 consecutive days. In patients symptoms were infrequent (2% incidence of grade 3 and no with relapsed and/or refractory B-cell neoplasms, two dosing grade 4 events). Carfilzomib demonstrated considerable activity schedules are being utilized: the schedule described previously 101 in bortezomib-naïve patients, inducing PR or better in 46% of 54 and one involving 2 consecutive days of dosing repeated weekly. evaluable patients at 20 mg/m2 and 53% of patients at 27 mg/m2.91 Proteasome inhibition following the administration of oprozomib > The response rate in patients previously exposed to bortezomib reached 80% and clinical activity was noted in patients with was lower (18%).90 Responses across groups are durable, typically MM and WM. In patients receiving the 5 consecutive day sched- 8 to 9 months.90,91 ule, 5 of 19 MM patients (26%) and 8 of 10 WM patients (80%) Based on findings in animal studies in which a 30-minute infu- achieved a partial response or better. Exploration of the dose and sion of carfilzomib resulted in reduced toxicities,61 the effect of schedule continues as a single agent and in combination with infusional administration was tested in patients with relapsed and other anti-MM therapies. refractory myeloma. In a dose escalation study, PX-171-007, the 2 MTD dose of carfilzomib was determined to be 56 mg/m , more Biomarkers for Proteasome Inhibitors than twice the dose used in the studies described previously. In a cohort of 24 patients receiving this dose and who had received a As described previously, PI-based therapies have proven highly median of five prior lines of therapy (including two prior bortezo- effective in the treatment of MM and other B-cell neoplasms. mib-containing regimens), the ORR was 60%.92 This enhanced Given that response rates in single-agent trials are generally efficacy also correlated with a greater level of inhibition of all three <50%, there would be a distinct clinical benefit to identify those subunits of the immunoproteasome measured in isolated periph- CANCER THERAPEUTICS patients most likely to respond to proteasome inhibition prior to eral blood mononuclear cells (Lee S, et al., unpublished).93 This treatment initiation. Gene expression analysis from bone marrow– same dose and infusion time is currently being explored in a phase derived MM tumor cells from 169 bortezomib-treated patients III trial of nearly 900 patients comparing carfilzomib plus low-dose and 70 dexamethasone-treated patients revealed a 100-gene sig- dexamethasone (Cd) to bortezomib plus low-dose dexamethasone nature that provided a stratification for patients likely to respond (Vd) in MM patients with relapsed disease. that performed better than standard staging systems.102 However, Trials of carfilzomib in combination with other agents in MM this signature provided only a modest increase in predictive power have been initiated, including a phase Ib/II safety and efficacy study for treatment with bortezomib versus dexamethasone. More re- of carfilzomib in combination with lenalidomide and low-dose cently, Keats et al.54 reanalyzed this dataset based on a pathway dexamethasone (CRd) in relapsed and/or refractory MM. At the analysis of NF-κB and the realization that TRAF3, a key regula- maximum planned dose, the ORR was 77% with a median duration tory of the noncanonical NF-κB pathway, is a tumor suppressor in of response of 22 months.94 The CRd combination is now being MM cell lines. They found a dramatic enrichment for response tested in an international, multicenter, randomized, open-label to bortezomib in patients with low levels of TRAF3 expression. phase III study in comparison with lenalidomide and low-dose However, these data remain to be validated in a separate sample dexamethasone (Rd) in approximately 780 patients with relapsed set. A transcriptomic analysis of samples derived from single-agent MM following one to three prior therapies. The CRd regimen has carfilzomib trials suggest that patients with the highest level of im- also been explored in newly diagnosed MM patients.95 When carfil- munoglobulin heavy chain expression were the most sensitive to zomib is combined with Rd at a dose of 36 mg/m2, 62% of the carfilzomib therapy.103 Similar findings were noted in the expres- 53 patients treated achieved a CR. In addition, 20 of 21 patients an- sion data from bortezomib-treated patients described previously.103 alyzed for signs of minimal residual disease (MRD), utilizing mul- These data are supported by phenotypic data from patients pro- tiparameter flow cytometry were determined to be free of MRD. gressing on bortezomib-based therapy, in which resistance to bortezomib was associated with a dedifferentiated (and lower im- Ixazomib munoglobulin expressing) B-cell phenotype.104 Taken together, these findings suggest that biomarkers, potentially those involving Initial clinical studies of ixazomib involved dose escalation studies an analysis of protein load of immunoglobulin expression, may be in patients with hematologic malignancies and explored both developed to predict those patients most likely to respond to PIs.

SELECTED REFERENCES

The full reference list can be accessed at lwwhealthlibrary.com/oncology. 5. Groll M, Ditzel L, Lowe J, et al. Structure of 20S proteasome from yeast at 2.4 A resolution. Nature 1997;386:463–471. 1. Ciechanover A. Intracellular protein degradation: from a vague idea thru the 8. Griffin TA, Nandi D, Cruz M, et al. Immunoproteasome assembly: coop- lysosome and the ubiquitin-proteasome system and onto human diseases and erative incorporation of interferon gamma (IFN-gamma)-inducible subunits. drug targeting. Biochim Biophys Acta 2012;1824:3–13. J Exp Med 1998;187:97–104. 3. Wilkinson KD. Ubiquitination and deubiquitination: targeting of proteins 10. Adams J. The proteasome: a suitable antineoplastic target. Nat Rev Cancer for degradation by the proteasome. Semin Cell Dev Biol 2000;11:141–148. 2004;4:349–360. 262 Cancer Therapeutics

11. Vinitsky A, Michaud C, Powers JC, et al. Inhibition of the chymotrypsin- 54. Keats JJ, Fonseca R, Chesi M, et al. Promiscuous mutations activate the like activity of the pituitary multicatalytic proteinase complex. Biochemistry noncanonical NF-kappaB pathway in multiple myeloma. Cancer Cell 2007; 1992;31:9421–9428. 12:131–144. 12. Orlowski RZ, Eswara JR, Lafond-Walker A, et al. Tumor growth inhibition 56. Obeng EA, Carlson LM, Gutman DM, et al. Proteasome inhibitors induce induced in a murine model of human Burkitt’s lymphoma by a proteasome a terminal unfolded protein response in multiple myeloma cells. Blood inhibitor. Cancer Res 1998;58:4342–4348. 2006;107:4907–4916. 16. Meng L, Mohan R, Kwok BH, et al. Epoxomicin, a potent and selective pro- 57. Shabaneh TB, Downey SL, Goddard AL, et al. Molecular basis of differential teasome inhibitor, exhibits in vivo antiinflammatory activity. Proc Natl Acad sensitivity of myeloma cells to clinically relevant bolus treatment with bort- Sci U S A 1999;96:10403–10408. ezomib. PLoS One 2013;8:e56132. 20. Bross PF, Kane R, Farrell AT, et al. Approval summary for bortezomib 59. Papandreou CN, Daliani DD, Nix D, et al. Phase I trial of the proteasome in- for injection in the treatment of multiple myeloma. Clin Cancer Res hibitor bortezomib in patients with advanced solid tumors with observations 2004;10:3954–3964. in androgen-independent prostate cancer. J Clin Oncol 2004;22:2108–2121. 21. Herndon TM, Deisseroth A, Kaminskas E, et al. U.S. Food and Drug Admin- 60. Wang Z, Yang J, Kirk C, et al. Clinical pharmacokinetics, metabolism, and istration approval: carfilzomib for the treatment of multiple myeloma. Clin drug-drug interaction of carfilzomib. Drug Metab Dispos 2013;41:230–237. Cancer Res 2013;19:4559–4563. 62. Meiners S, Heyken D, Weller A, et al. Inhibition of proteasome activity in- 23. Adams J, Behnke M, Chen S, et al. Potent and selective inhibitors of the duces concerted expression of proteasome genes and de novo formation of proteasome: dipeptidyl boronic acids. Bioorg Med Chem Lett 1998;8: mammalian proteasomes. J Biol Chem 2003;278:21517–21525. 333–338. 63. Lonial S, Waller EK, Richardson PG, et al. Risk factors and kinetics of throm- 27. Kisselev AF, van der Linden WA, Overkleeft HS. Proteasome inhibitors: an bocytopenia associated with bortezomib for relapsed, refractory multiple my- expanding army attacking a unique target. Chem Biol 2012;19:99–115. eloma. Blood 2005;106:3777–3784. 28. Arastu-Kapur S, Anderl JL, Kraus M, et al. Nonproteasomal targets of the 64. Richardson PG, Briemberg H, Jagannath S, et al. Frequency, characteristics, proteasome inhibitors bortezomib and carfilzomib: a link to clinical adverse and reversibility of peripheral neuropathy during treatment of advanced mul- events. Clin Cancer Res 2011;17:2734–2743. tiple myeloma with bortezomib. J Clin Oncol 2006;24:3113–3120. 30. Parlati F, Lee SJ, Aujay M, et al. Carfilzomib can induce tumor cell death 65. Richardson PG, Barlogie B, Berenson J, et al. A phase 2 study of bortezomib through selective inhibition of the chymotrypsin-like activity of the protea- in relapsed, refractory myeloma. N Engl J Med 2003;348:2609–2617. some. Blood 2009;114:3439–3447. 67. Richardson PG, Sonneveld P, Schuster MW, et al. Bortezomib or high-dose dexa- 31. Britton M, Lucas MM, Downey SL, et al. Selective inhibitor of proteasome’s methasone for relapsed multiple myeloma. N Engl J Med 2005;352:2487–2498. caspase-like sites sensitizes cells to specific inhibition of chymotrypsin-like 69. San Miguel JF, Schlag R, Khuageva NK, et al. Bortezomib plus melphalan sites. Chem Biol 2009;16:1278–1289. and prednisone for initial treatment of multiple myeloma. N Engl J Med 33. Demo SD, Kirk CJ, Aujay MA, et al. Antitumor activity of PR-171, a novel 2008;359:906–917. irreversible inhibitor of the proteasome. Cancer Res 2007;67:6383–6391. 76. Richardson PG, Weller E, Lonial S, et al. Lenalidomide, bortezomib, and 34. Kupperman E, Lee EC, Cao Y, et al. Evaluation of the proteasome inhibi- dexamethasone combination therapy in patients with newly diagnosed mul- tor MLN9708 in preclinical models of human cancer. Cancer Res 2010;70: tiple myeloma. Blood 2010;116:679–686. 1970–1980. 78. Fisher RI, Bernstein SH, Kahl BS, et al. Multicenter phase II study of bort- 36. Chauhan D, Singh AV, Aujay M, et al. A novel orally active proteasome ezomib in patients with relapsed or refractory mantle cell lymphoma. J Clin inhibitor ONX 0912 triggers in vitro and in vivo cytotoxicity in multiple my- Oncol 2006;24:4867–4874. eloma. Blood 2010;116:4906–4915. 79. Milano A, Iaffaioli RV, Caponigro F. The proteasome: a worthwhile target for 38. Chauhan D, Catley L, Li G, et al. A novel orally active proteasome inhibitor the treatment of solid tumours? Eur J Cancer 2007;43:1125–1133. induces apoptosis in multiple myeloma cells with mechanisms distinct from 84. Neubert K, Meister S, Moser K, et al. The proteasome inhibitor bortezomib Bortezomib. Cancer Cell 2005;8:407–419. depletes plasma cells and protects mice with lupus-like disease from nephri- 40. Chauhan D, Tian Z, Zhou B, et al. In vitro and in vivo selective antitumor tis. Nat Med 2008;14:748–755. activity of a novel orally bioavailable proteasome inhibitor MLN9708 against 88. Siegel DS, Martin T, Wang M, et al. A phase 2 study of single-agent carfil- multiple myeloma cells. Clin Cancer Res 2011;17:5311–5321. zomib (PX-171-003-A1) in patients with relapsed and refractory multiple my- 41. Kuhn DJ, Chen Q, Voorhees PM, et al. Potent activity of carfilzomib, a eloma. Blood 2012;120:2817–2825. novel, irreversible inhibitor of the ubiquitin-proteasome pathway, against 91. Vij R, Wang M, Kaufman JL, et al. An open-label, single-arm, phase 2 (PX- preclinical models of multiple myeloma. Blood 2007;110:3281–3290. 171-004) study of single-agent carfilzomib in bortezomib-naive patients with 44. Fernandez Y, Verhaegen M, Miller TP, et al. Differential regulation of noxa relapsed and/or refractory multiple myeloma. Blood 2012;119:5661–5670. in normal melanocytes and melanoma cells by proteasome inhibition: thera- 95. Jakubowiak AJ, Dytfeld D, Griffith KA, et al. A phase 1/2 study of carfilzomib peutic implications. Cancer Res 2005;65:6294–6304. in combination with lenalidomide and low-dose dexamethasone as a front- 48. Mannava S, Zhuang D, Nair JR, et al. KLF9 is a novel transcriptional regu- line treatment for multiple myeloma. Blood 2012;120:1801–1809. lator of bortezomib- and LBH589-induced apoptosis in multiple myeloma 102. Mulligan G, Mitsiades C, Bryant B, et al. Gene expression profiling and cells. Blood 2012;119:1450–1458. correlation with outcome in clinical trials of the proteasome inhibitor bort- 49. Koepp DM, Harper JW, Elledge SJ. How the cyclin became a cyclin: regu- ezomib. Blood 2007;109:3177–3188. lated proteolysis in the cell cycle. Cell 1999;97:431–434. 103. Loehr A, Degenhardt JD, Kwei KA, et al. Immunoglobulin expression is 53. Chapman MA, Lawrence MS, Keats JJ, et al. Initial genome sequencing and a major determinant of patient sensitivity to proteasome inhibitors. Blood analysis of multiple myeloma. Nature 2011;471:467–472. 2013;122:1903.

tahir99 - UnitedVRG Poly (ADP-ribose) 25 Polymerase Inhibitors

Alan Ashworth

INTRODUCTION Both GC and SSA rely on sequence homology for repair whereas NHEJ uses no, or little, homology.2,3 NHEJ is the most impor- Cancer cells may harbor defects in DNA repair pathways lead- tant pathway for the repair of DSBs during G0, G1, and early S ing to genomic instability. This can foster tumorigenesis but also phases of the cell cycle, although it is likely active throughout the provides a weakness that can be exploited therapeutically. Tumors cell cycle.7,8 This form of DSB repair usually results in changes in with compromised ability to repair double-strand DNA breaks by DNA sequence at the break site and, occasionally, in the joining of homologous recombination, including those with defects in the previously unlinked DNA molecules, potentially resulting in gross BRCA1 and BRCA2 genes, are highly sensitive to blockade of chromosomal rearrangements such as translocations.9 GC uses a the repair of DNA single-strand breaks, via the inhibition of the homologous sequence, preferably the sister chromatid, as a tem- enzyme poly(ADP-ribose) (PARP). This provides the basis for a plate to resynthesize the DNA surrounding the DSB, and therefore synthetic lethal approach to cancer therapy, which is showing con- generally results in accurate repair of the break. Repair by GC is siderable promise in the clinic. critically dependent on the recombinase function of RAD51 and is facilitated by a number of other proteins. SSA also involves the use of homologous sequences for the repair of DSBs, but unlike GC, CELLULAR DNA REPAIR PATHWAYS SSA is RAD51-independent and involves the annealing of DNA strands formed after resection at the DSB. The detailed mecha- nism of SSA is still obscure but it frequently results in the loss of DNA is continually damaged by environmental exposures and one of the homologous sequences and deletion of the intervening endogenous activities, such as DNA replication and cellular free- CANCER THERAPEUTICS sequence.9 SSA is a potentially important pathway of mutagenesis radical generation, which cause diverse lesions including base mod- because a significant fraction of mammalian genomes consist of ifications, double-strand breaks (DSB), single-strand breaks (SSB), repetitive elements. GC and SSA are cell-cycle regulated and are and intrastrand and interstrand cross-links.1 These aberrations are most active in S-G phases of the cell cycle.10 repaired by distinct repair pathways, which are coordinated to 2 maintain the stability and integrity of the genome. This faithful repair of DNA damage is an essential prerequisite for the mainte- nance of genomic integrity and cellular and organismal viability. THE DEVELOPMENT OF PARP INHIBITORS Where one DNA strand is affected and the intact complementary strand is available as a template, the base-excision repair (BER), PARP inhibitors were originally developed as chemopotentiators, nucleotide-excision repair, or mismatch repair pathways are used which are agents that enhance the effects of DNA damage— and these pathways are highly efficient at repairing damage. DSBs, a common mechanism of action of drugs used to treat cancer. more problematic than SSBs because the complementary strand is The rationale was that inhibition of the repair of chemotherapy- not available as a template, are repaired by the homologous recom- induced DNA damage might give greater efficacy. Early studies 1 using relatively nonspecific PARP inhibitors such as 3-aminoben- bination (HR) or nonhomologous end-joining (NHEJ) pathways. 11 Endogenous base damage, including SSBs, is the most com- zamide, demonstrated potential synergy with alkylating agents. mon DNA aberration and it has been estimated that the average Subsequent studies with more potent PARP inhibitors demon- strated synergy with temozolomide, an observation that was taken cell may repair 10,000 such lesions every day. BER is an important 12 pathway for the repair of SSBs and involves the sensing of the lesion into a clinical trial with AG014699, a PARP inhibitor developed followed by the recruitment of a number of other proteins. PARP-1 by Pfizer. This agent is now being developed by Clovis. Although (poly[ADP]ribose polymerase) is a critical component of the major the major focus of this chapter is the use of PARP inhibitors in “short-patch” BER pathway. PARP is an enzyme, discovered over synthetic lethal therapeutic strategies, their use in chemopoten- 40 years ago,2 that produces large branched chains of poly(ADP) tiation in combination with chemotherapy remains under active ribose (PAR) from NAD+. In humans, there are 17 members of investigation, as described later. the PARP gene family but most of these are poorly characterized.3,4 The abundant nuclear protein PARP-1 senses and binds to DNA nicks and breaks, resulting in activation of catalytic activity caus- BRCA1 AND BRCA2 MUTATIONS AND ing poly(ADP)ribosylation of PARP-1 itself as well as other accep- DNA REPAIR tor proteins including histones. This modification may signal the recruitment of other components of DNA repair pathways as well Heterozygous germline mutations in the BRCA1 and BRCA2 as modify their activity. The highly negatively charged PAR that genes confer a high risk of breast (up to 85% lifetime risk) and ovar- is produced around the site of damage may also serve as an an- ian (10% to 40%) cancer in addition to a significantly increased tirecombinogenic factor. In addition to the BER pathway PARP risk of pancreatic, prostate, and male breast cancer.13 The genes enzymes have been implicated in numerous cellular pathways.3,4 have been classified as tumor suppressors, because the wild-type Two main DSB repair pathways are available within eukaryotic BRCA allele is frequently lost in tumors, a phenomenon that oc- cells: NHEJ and HR.5,6 HR can be further subdivided into the gene curs by a variety of mechanisms. The BRCA1 and BRCA2 genes conversion (GC) and single-strand annealing (SSA) subpathways.1 encode large proteins that likely function in multiple cellular 263 264 Cancer Therapeutics pathways, including transcription, cell-cycle regulation, and the DNA lesions that might be repaired by GC. As a loss of function of maintenance of genome integrity. However, the roles of BRCA1 either BRCA1 or BRCA2 impairs GC,14,15 a loss of PARP-1 func- and BRCA2 in DNA repair have been best documented.14 tion in a BRCA1- or BRCA2-defective background could result BRCA1- and BRCA2-deficient cells are highly sensitive to ion- in the generation of replication-associated DNA lesions normally izing radiation and display chromosomal instability, which is likely repaired by sister chromatid exchange. If so, this might lead to cell- to be a direct consequence of unrepaired DNA damage.14 The sim- cycle arrest and/or cell death. Therefore, PARP inhibitors could be ilar genomic instability in BRCA1- and BRCA2-deficient cells and selectively lethal to cells lacking functional BRCA1 or BRCA2 but the interaction of both BRCA1 and BRCA2 with RAD51 suggested might be minimally toxic to normal cells. This would indicate a a functional link between the three proteins in the RAD51-medi- synthetic lethal interaction between PARP and BRCA1 or BRCA2. ated DNA damage repair process. However, although BRCA2 is Exemplifying this principle, potent inhibitors of PARP were ap- directly involved in RAD51-mediated repair, affecting the choice plied to cells deficient in either BRCA1 or BRCA2. Cell survival between GC and SSA, BRCA1 acts upstream of these pathways15; assays showed that cell lines lacking wild-type BRCA1 or BRCA2 both GC and SSA are reduced in BRCA1-deficient cells, placing were extremely sensitive to these agents compared with heterozy- BRCA1 before the branch point of GC and SSA.15 gous mutant or wild-type cells.24,25 BRCA1 has a role in signaling DNA damage and cell-cycle To explain these observations, a model was proposed whereby checkpoint regulation,14,15 whereas BRCA2 has a more direct role persistent single-strand gaps in DNA caused by PARP inhibition in DNA repair itself. BRCA2 is thought to promote genomic stabil- when encountered by a replication fork might trigger fork arrest, ity through a role in the error-free repair of DSBs by GC via asso- collapse, and/or a DSB.26 Alternatively, PARP-1 trapped on DNA ciation with RAD51. Aberrations in BRCA2-deficient cells arise at by the inhibition of enzyme activity might also cause a fork col- least in part by the use of the SSA pathway. NHEJ, however, is ap- lapse. Normally, these DSBs would be repaired by RAD51-de- parently unaffected in BRCA2-deficient cells.14,15 Loss of BRCA2, pendent GC.14,15 However, in the absence of BRCA1 or BRCA2, therefore, results in the repair of DSBs by preferential utilization of the replication fork cannot be restarted and collapses, causing an error-prone mechanism, which potentially explains the appar- persistent chromatid breaks. When repaired by the alternative ent chromosome instability associated with BRCA2 deficiency.15 error-prone DSB repair mechanisms of SSA or NHEJ, large The physical interaction between BRCA2 and RAD51 is es- numbers of chromatid aberrations would be induced, leading to sential for error-free DSB repair. BRCA2 is required for the local- cell lethality.26 The idea that the defect in GC is being targeted ization of RAD51 to sites of DNA damage, where RAD51 forms in BRCA-deficient cells is supported by the demonstration that the nucleoprotein filament required for recombination. The foci deficiency in other genes implicated in HR also confers sensitiv- of the RAD51 protein are apparent in the nucleus after certain ity to PARP inhibitors.27 This further suggests that this approach forms of DNA damage and these likely represent sites of repair by may be more widely applicable in the treatment of sporadic can- HR; BRCA2-deficient cells do not form RAD51 foci in response cers with impairments of the HR pathway or BRCAness28 (see to DNA damage.15 Two different domains within BRCA2 interact the following). with RAD51, the eight BRC repeats in the central part of the pro- tein and a distinct domain, TR2, at the C-terminus.16 INITIAL CLINICAL RESULTS TESTING SYNTHETIC LETHALITY OF PARP PARP-1 INHIBITION AS A SYNTHETIC INHIBITORS AND BRCA MUTATION LETHAL THERAPEUTIC STRATEGY FOR THE TREATMENT OF BRCA-DEFICIENT Phase I studies29 established that (AstraZeneca, London, CANCERS UK; formerly KU-0059436, KuDOS Pharmaceuticals, Cam- bridge, UK) could be administered safely as a single agent at a Synthetic lethality is defined as the situation when a mutation dose of 400 mg twice per day. Side effects were classified as mild in either of two genes individually has no effect, but combining and were unlike those typically experienced with cytotoxic che- the mutations leads to death.17 This effect was first described and motherapy. Significant and durable responses were observed in studied in genetically tractable organisms such as Drosophila and patients with germ-line BRCA1 or BRCA2 mutations and breast yeast.17,18 This effect can arise because of a number of different ovary or prostate cancer. Of the 19 mutation carriers enrolled, gene–gene interactions. Examples include two genes in separate 9 had an objective response defined by Response Evaluation semiredundant or cooperating pathways, and two genes acting in Criteria in Sold Tumors (RECIST) criteria and 12 had stable the same pathway where loss of both critically affects flux through disease for more than 4 months in duration. A similar magni- the pathway. The implication is that targeting one of these genes tude of clinical responses was observed in an expanded cohort.30 in a cancer where the other is defective should be selectively lethal These observations are impressive because the cohort had been to the tumor cells but not toxic to the normal cells. In principle, heavily pretreated and most were resistant to a wide range of this should lead to a large therapeutic window.19 The original sug- chemotherapies. 29,30 gestion that the concept of synthetic lethality could be used in Phase II studies were subsequently performed in advanced the selection or development of cancer therapeutics came from breast and ovarian cancers arising in BRCA1 and BRCA2 muta- Hartwell et al.,18 and from experiments performed in yeast. Syn- tion carriers.31,32 The reported response rate was 41% in the breast thetic lethal screens have now been performed in a number of study and 52% in the ovarian group; both groups had been heavily model organisms20 and in human cells,21 and these have revealed pretreated. Again, the drug was well tolerated. Another study of multiple potential gene–gene interactions, some of which could BRCA1/2 carriers with ovarian cancer compared olaparib with pe- be exploited clinically. However, synthetic lethal therapies have gylated liposomal doxorubicin (PLD).33 There was no significant not been clinically used until recently, when evidence has been difference in the response rates, but there were some differences provided for PARP-1 inhibition as a potential synthetic lethal ap- in the patient characteristics and an unexpectedly high rate of re- proach for the treatment of BRCA-mutation–associated cancers. sponse to PLD. PARP-1 inhibition causes failure of the repair of SSB lesions There are also reports of responses to PARP inhibitors in but does not affect DSB repair.22 However, a persistent DNA SSB BRCA2 mutation carriers with prostate34 and pancreatic35 cancer. encountered by a DNA replication fork will cause stalling of the A number of other PARP inhibitors are in clinical development fork and may result in either fork collapse or the formation of a (Table 25.1), and some of these have shown efficacy in the treat- DSB.23 Therefore, the loss of PARP-1 increases the formation of ment of cancers arising in BRCA1 or BRCA2 mutation carriers.36,37

tahir99 - UnitedVRG Chapter 25 Poly (ADP-ribose) Polymerase Inhibitors 265

45 TABLE 25.1 III study showed no improvement in PFS. The reasons for this are uncertain, but significant questions have been raised about PARP Inhibitors in Late Stage Clinical Development whether iniparib is indeed a bona fide PARP inhibitor. Therefore, it is now generally conceded that studies of iniparib have no impli- cations for PARP inhibitors as a drug class.46 Agent Company Phase III Trials Which population of patients lacking a BRCA1 or BRCA2 mu- Olaparib AstraZeneca (formerly BRCA-mutant ovarian tation might benefit from PARP inhibitors remains unclear. This KuDOS) cancer is likely to require the development of a clinical test to identify Tesaro (formerly Merck) Platinum sensitive ovarian prospectively tumors with intrinsic sensitivity. Presently, most ef- 47 cancer forts are directed at developing assays of DNA repair deficiency. BRCA-mutant breast cancer Clovis (formerly Pfizer) Platinum sensitive ovarian MECHANISMS OF RESISTANCE TO cancer PARP INHIBITORS Veliparib AbbVie (formerly Abbot) Undisclosed Resistance to targeted therapy frequently occurs, but it was unclear BMN673 BioMarin (formerly Lead) BRCA-mutant breast cancer how resistance might arise to a synthetic lethal therapy.48 Poten- tial mechanisms of resistance to PARP inhibitors have, however, Adapted from Garber, K. PARP inhibitors bounce back. Nat Rev Drug Discov 2013;12:725–727. been elucidated both directly in vitro, in mouse models, and in the clinic.48 An in vitro model for resistance was developed by producing cells from the highly PARP inhibitor–sensitive BRCA2- deficient cell line CAPAN1, which carries a c.6174delTBRCA2 THE USE OF PARP INHIBITORS IN frameshift mutation. CAPAN1 cells cannot form damage-induced SPORADIC CANCERS RAD51 foci, are defective for HR, and are extremely sensitive to treatment with PARP inhibitors.49 PARP inhibitor–resistant clones Germline mutations in BRCA1 or BRCA2 are relatively com- were highly resistant (over 1,000-fold) to the drug and were also mon in hereditary breast and ovarian cancer. However, inactiva- cross-resistant to the DNA cross-linking agent cisplatin, but not tion of BRCA genes by mutation in sporadic cancers is rare, at to the microtubule-stabilizing drug docetaxel. PARP inhibitors least in breast cancer, which may seem to limit the application and cisplatin both exert their effects on BRCA-deficient cells by of PARP inhibitors to a wider range of patients. However, many increasing the frequency of misrepaired DSBs in the absence of

tumors display features in common with BRCA-deficient tumors, effective HR. Therefore, this observation indicates that the resis- CANCER THERAPEUTICS including similar defects in DNA repair due to either epigenetic tance of PARP inhibitor–resistant clones to PARP inhibitors might mutation of BRCA1, such as promoter methylation, or mutation be because of restored HR. This contention was supported by the of other components of BRCA-associated pathways.28 This BRCA- acquisition in PARP inhibitor–resistant clone cells of the ability ness may make these tumors also susceptible to PARP inhibition.28 to form RAD51 foci after PARP inhibitor treatment or exposure For example, phosphatase and tensin homolog (PTEN) mutations, to irradiation. which occur with a frequency estimated at 50% to 80% in spo- DNA sequencing of PARP inhibitor–resistant clones revealed radic tumors,38 may cause PARP inhibitor sensitivity in preclini- the unexpected presence of novel BRCA2 alleles that resulted cal models, possibly because PTEN-null cells display BRCAness in the elimination of the c.6174delT mutation and restoration phenotypes, such as the inability to efficiently repair certain forms of an open reading frame.49 Therefore, in this case, resistance of DNA damage.39 arises because of gain of function mutations in the synthetic le- Traditional histopathologic methods and, more recently, gene thal partner (BRCA2) rather than the direct drug target (PARP). expression profiling approaches have shown the phenotypic over- Alternative mechanisms of PARP inhibitor resistance have also lap between triple-negative breast cancers, basal-like breast can- been described.48 A mouse model of BRCA1-associated mam- cers, and BRCA1 familial breast cancers.40,41 In gene expression mary gland cancer demonstrated the efficacy of olaparib in vivo profiling studies, it has been observed thatBRCA1 familial cancers and was used to study mechanisms of resistance.50 Resistance strongly segregate with basal-like tumors and share features such seemed to be caused by the upregulation of ABCB1a/b, which en- as high-grade and pushing margins.28,40,41 Although the overlap is code P-glycoprotein pumps; this effect could be reversed with the not absolute, it leads to the hypothesis that there may be a subset of P-glycoprotein inhibitor tariquidar. In addition, other alterations sporadic breast cancers that exhibits features of BRCAness, includ- in DNA repair pathways have been proposed to compensate for ing deficiencies in HR and that may be susceptible to treatment BRCA1 deficiency resulting in PARP inhibitor deficiency.48 with drugs such as PARP inhibitors.26 Studies of the mechanisms of resistance to PARP inhibitors in There have been several studies of PARP inhibitors in sporadic patient material are still at an early stage. Initial studies addressed ovarian cancer. A study by Lederman42 showed in a maintenance the mechanism of resistance to platinum salts in BRCA mutation study following the response to platinum therapy a significant carriers. Cisplatin and carboplatin are part of the standard of care benefit in terms of progression-free survival (PFS) of olaparib for the treatment of ovarian cancer, including individuals with compared to placebo. This was even more pronounced when the BRCA1 or BRCA2 mutations. Platinum salts are thought to exert subgroup of BRCA mutation carriers were examined.43 In both their BRCA-selective effects by a similar mechanism to PARP in- cases, the overall survival (OS) advantage was less than the PFS, hibitors.15 Clinical observations suggest that BRCA mutation carri- but in the case of the BRCA mutation group, this reached statisti- ers with ovarian cancer usually respond better to these agents than cal significance. Gelmon44 also showed activity in sporadic ovarian patients without BRCA mutations51,52; however, resistance does cancer. In contrast, a study in sporadic triple-negative breast can- eventually occur. To investigate this effect, BRCA1 and BRCA2 cer failed to observe any benefit, although the study was small and have been sequenced in tumor material from mutation carri- the patients were heavily pretreated.44 ers.49,53 These studies revealed mutations in BRCA1 or BRCA2 Iniparib (initially reported as a PARP inhibitor) showed an that restored the open reading frame and likely contributed to overall survival benefit in a Phase II trial of triple-negative breast platinum resistance. These observations suggest that specific mu- cancer in combination with gemcitabine and carboplatin com- tations in BRCA1 or BRCA2 and sensitivity to therapeutics in cell pared with chemotherapy alone.45 However, a subsequent Phase lines and patients can be suppressed by intragenic deletion. Pre- 266 Cancer Therapeutics sumably, these mutations occur randomly and are then selected BRCA2 carrier, provides the rationale for this approach. Inhibiting for by differential drug sensitivity. Therefore, the best use of these the DNA repair protein PARP results in the generation of specific agents is likely to be earlier in the disease process when the disease DNA lesions that require BRCA1 and BRCA2 specialized repair burden is smaller, which will reduce the probability of resistance function(s) for their removal. Preclinical data indicate that tumors based on stochastic genetic reversion. Recently, similar observa- defective in wild-type BRCA1 or BRCA2 could be much more tions of revertant BRCA alleles were made in two patients who sensitive to PARP inhibition than unaffected heterozygous tissues, became resistant after an initial response to olaparib.54 Although providing a potentially large therapeutic window. The safety and preliminary, these results suggest that this mechanism is responsi- efficacy of this approach is currently being tested in clinical trials, ble for at least some of the clinical resistance observed. Doubtless, which, if successful, may lead to registration for routine clinical as with other targeted therapies, multiple resistance mechanisms use of one or more PARP inhibitors.37 will be implicated as further patients are studied.48 Synthetic lethality by combinatorial targeting of DNA repair pathways may have usefulness as a therapeutic approach beyond familial cancers. The majority of solid tumors also exhibit genomic PROSPECTS instability and aneuploidy. This suggests that pathways involved in the maintenance of genomic stability are dysfunctional in a sig- Currently, the treatments for cancers arising in carriers of BRCA1 nificant proportion of neoplastic disorders.47 Understanding which or BRCA2 mutations are the same as those that occur sporadically specialized DNA damage response and repair pathways are abro- matched for tumor pathology and age of onset. However, tumors gated in sporadic tumor subtypes may allow for the development of in BRCA1 or BRCA2 mutation carriers lack wild-type BRCA1 or therapies that target the residual repair pathways on which the can- BRCA2, but normal tissues retain a single wild-type copy of the cer, but not normal tissue, is now completely dependent. These relevant gene. This is a potentially targetable alteration that pro- potential therapies may significantly improve response rates while vides the basis for new mechanism-based approaches to the treat- causing fewer treatment-related toxicities. However, these ap- ment of cancer. The biochemical difference in capacity to carry proaches may be associated with mechanism-associated resistance, out HR between the tumor and normal tissues, in a BRCA1 or and careful consideration of their optimal use will be required.

SELECTED REFERENCES

The full reference list can be accessed at lwwhealthlibrary.com/oncology. 28. Turner N, Tutt A, Ashworth A. Hallmarks of ‘BRCAness’ in sporadic cancers. Nat Rev Cancer 2004;4:814–819. 1. Hoeijmakers JH. Genome maintenance mechanisms for preventing cancer. 29. Fong PC, Boss DS, Yap TA, et al. Inhibition of poly(ADP-ribose) polymerase Nature 2001;411:366–374. in tumors from BRCA mutation carriers. N Engl J Med 2009;361:123–134. 2. Chambon P, Weill JD, Mandel P. Nicotinamide mononucleotide activation 30. Fong PC, Yap TA, Boss DS, et al. Poly(ADP)-ribose polymerase (PARP) inhi- of new DNA-dependent polyadenylic acid synthesizing nuclear enzyme. Bio- bition: frequent durable responses in BRCA carrier ovarian cancer correlat- chem Biophys Res Commun 1963;11:39–43. ing with platinum-free interval. J Clin Oncol 2010;28:2512–2519. 11. Durkacz BW, Omidiji O, Gray DA, et al. (ADP-ribose)n participates in DNA 31. Audeh MW, Carmichael J, Penson RT, et al. Oral poly(ADP-ribose) poly- excision repair. Nature 1980;283(5747):593–596. merase inhibitor olaparib in patients with BRCA1 or BRCA2 mutations and 13. Wooster R, Weber BL. Breast and ovarian cancer. N Engl J Med recurrent ovarian cancer: a proof-of-concept trial. Lancet 2010;376:245–251. 2003;348:2339–2347. 32. Tutt A, Robson M, Garber JE, et al. Oral poly(ADP-ribose) polymerase in- 17. Dobzhansky T. Genetics of natural populations: Xiii. Recombination hibitor olaparib in patients with BRCA1 or BRCA2 mutations and advanced and variability in populations of Drosophila pseudoobscura. Genetics breast cancer: a proof-of-concept trial. Lancet 2010;376:235–244. 1946;31:269–290. 42. Ledermann J, Harter P, Gourley C, et al. Olaparib maintenance therapy in plat- 18. Hartwell LH, Szankasi P, Roberts CJ, et al. Integrating genetic approaches inum-sensitive relapsed ovarian cancer. N Engl J Med 2012;366;1382–1392. into the discovery of anticancer drugs. Science 1997;278:1064–1068. 43. Ledermann JA, Harter P, Gourley C. Olaparib maintenance therapy in pa- 19. Kaelin WG Jr. The concept of synthetic lethality in the context of anticancer tients with platinum-sensitive relapsed serous ovarian cancer (SOC) and a therapy. Nat Rev Cancer 2005;5:689–698. BRCA mutation (BRCAm). J Clin Oncol 2013;31 (suppl; abstr 5505). 21. Iorns E, Lord CJ, Turner N, et al. Utilizing RNA interference to enhance 46. Mateo J, Ong M, Tan DS, et al. Appraising iniparib, the PARP inhibitor that cancer drug discovery. Nat Rev Drug Discov 2007;6:556–568. never was—what must we learn? Nat Rev Clin Oncol 2013;10:688–696. 24. Farmer H, McCabe N, Lord CJ, et al. Targeting the DNA repair defect in 47. Lord CJ, Ashworth A. The DNA damage response and cancer therapy. Na- BRCA mutant cells as a therapeutic strategy. Nature 2005;434:917–921. ture 2012;481:287–294. 25. Bryant HE, Schultz N, Thomas HD, et al. Specific killing of BRCA2-defi- 48. Lord CJ, Ashworth A. Mechanisms of resistance to therapies targeting BR- cient tumours with inhibitors of poly(ADP-ribose) polymerase. Nature 2005; CA-mutant cancers. Nat Med 2013;19;1381–1388. 434:913–917. 49. Edwards S, Brough R, Lord CJ, et al. Resistance to therapy caused by intra- 26. Ashworth A. A synthetic lethal therapeutic approach: PARP inhibitors for the genic deletion in BRCA2. Nature 2008;451(7182):1111–1115. treatment of cancers deficient in double-strand break repair. J Clin Oncol 53. Sakai W, Swisher EM, Karlan BY, et al. Secondary mutations as a mecha- 2008;26:3785–3790. nism of cisplatin resistance in BRCA2-mutated cancers. Nature 2008;451: 27. McCabe N, Turner NC, Lord CJ, et al. Deficiency in the repair of DNA 1116–1120. damage by homologous recombination and sensitivity to poly(ADP-ribose) 54. Barber LJ, Sandhu S, Chen L, et al. Secondary mutations in BRCA2 associ- polymerase inhibition. Cancer Res 2006;66:8109–8115. ated with clinical resistance to a PARP inhibitor. J Pathol 2013;229:422–429.

tahir99 - UnitedVRG Miscellaneous 26 Chemotherapeutic Agents

M. Sitki Copur, Scott Nicholas Gettinger, Sarah B. Goldberg, and Hari A. Deshpande

HOMOHARRINGTONINE AND L-ASPARAGINASE OMACETAXINE l-Asparaginase (l- aminohydrolase, EC 3.5.1.1), which Homoharringtonine and its congener, harringtonine, are cepha- catalyzes the hydrolysis of the essential amino acid l-asparagine lotaxine esters isolated from the evergreen tree Cephalotaxus to l-aspartic acid and ammonia, is a naturally occurring enzyme 10,11 hainanensis, which are distributed throughout southern and in some microorganisms. Although cancer cells depend on an northeastern China. The two differ only by a single methylene exogenous source of l-asparagine for survival, normal cells can group, but both have a similar activity against murine leukemia.1 synthesize asparagine. In addition to the depletion of l-asparagine, The primary action of homoharringtonine appears to be the inhi- it may exert its antitumor activity through a effect, bition of protein synthesis and chain elongation through binding depleting essential glutamine stores and leading to the inhibition to 80S ribosome in eukaryotic cells.2 DNA effects may also be of DNA biosynthesis. It comes in three preparations, two of which important, involving a block in progression of cells from G1 phase are native forms purified from bacterial sources,Escherichia coli into S phase and from G2 phase into M phase.3 Homoharringto- and Erwinia carotovora. A third preparation, pegylated (PEG)- nine exhibits a triphasic plasma decay with a terminal half-life l-asparaginase, is a chemically modified form of the enzyme in of 65.3 hours and apparent volume of distribution of 2.4 L/kg.4 which native E. coli l-asparaginase has been covalently conju- 12 In early phase I studies, homoharringtonine was administered gated to polyethylene glycol. 5 as a 10 to 360 minute infusion daily for 10 days. Dose-limiting After an intramuscular (IM) injection, peak plasma levels, ap- CANCER THERAPEUTICS cardiovascular toxicity with hypotension began 4 or more hours proximately one-half of those achieved with IV administration, are after drug administration, which was alleviated by interrupting the reached within 14 to 24 hours. Plasma protein binding is 30%. The 13 infusion or by fluid administration and prolonging the duration pharmacokinetics vary depending on the source of the enzyme. of administration. Initial clinical studies with homoharringtonine Pharmacokinetic studies in newly diagnosed children with acute in China showed activity against acute myeloid leukemia (AML) lymphocytic leukemia (ALL) have shown peak serum concentra- and chronic phase chronic myeloid leukemia (CML).6 Variable tions in the range of 1 to 10 IU/mL in 24 to 48 hours of a single 2 activity was observed in the initial series of phase II trials in pe- dose of 2,500 to 25,000 IU/m of the enzyme derived from E. coli. 2 diatric and adult patients with acute leukemia. In early studies After a single dose of 25,000 IU/m , peak serum levels are reached of homoharringtonine, a continuous intravenous (IV) infusion at within 24 hours. PEG-l-asparaginase, when administered at a dose 2 2.5 mg/m2 per day for 10 to 14 days per month induced complete of 2,500 IU/m , achieves peak drug levels at 72 to 96 hours and has hematologic and cytogenetic responses in 72% and 31% of pa- a significantly longer half-life (5.7 days) than the E. coli l-asparagi- 13 tients, respectively, with chronic phase CML.7 nase preparation. Clinical trials have demonstrated the efficacy, The greater availability of homoharringtonine led to its further safety, and tolerability of PEG-l-asparaginase administered intra- testing and the development of a semisynthetic cephalotaxine muscularly, subcutaneously, or intravenously as part of multiagent ester, omacetaxine mepesuccinate.2 The mechanism of action of chemotherapy regimens in the management of newly diagnosed omacetaxine includes inhibition of protein synthesis and is in- and relapsed pediatric and adult ALL. l-Asparaginase can antago- dependent of direct Bcr-Abl binding. In vitro, it reduces protein nize antineoplastic effects of methotrexate if given concurrently levels of the Bcr-Abl oncoprotein and Mcl-1, an antiapoptotic or immediately before. These two drugs should be administered B-cell lymphoma 2 (Bcl-2) family member. The antileukemic sequentially at least 24 hours apart. l-Asparaginase has also been effect of omacetaxine is not affected by the presence of muta- shown to inhibit the metabolic clearance of vincristine and can tions in Bcr-Abl.8 Omacetaxine is absorbed following subcutane- result in increased neurotoxicity. Toxicity is less pronounced if ous administration of 1.25 mg/m2 twice daily for 11 days with a l-asparaginase is administered after vincristine. Hypersensitivity mean half-life of 6 hours, and a volume of distribution of 141 reactions occur in up to 25% of patients as a skin rash and urticaria +/−93.4 L. A phase 2 trial assessed the efficacy of omacetaxine or serious anaphylactic reactions. The risk increases with repeat in CML patients with T315I and tyrosine–kinase inhibitor fail- exposure, and as a single-agent use without steroids. PEG-l-aspar- ure. Patients received subcutaneous omacetaxine 1.25 mg/m2 aginase is less immunogenic than the native nonpegylated forms twice daily on days 1 through 14, every 28 days until hematologic of the enzyme. A number of other side effects are observed that response or a maximum of 6 cycles, and then days 1 through 7 are secondary to the inhibitory effects of l-asparaginase on cellular every 28 days as maintenance. Complete hematologic response protein synthesis. Decreased serum levels of insulin, key lipopro- was achieved in 77%, with a median response duration of 9.1 teins, and albumin have been reported. l-Asparaginase can cause months. Of patients, 23% achieved a major cytogenetic response, alterations in thyroid function tests as early as 2 days after an admin- including a complete cytogenetic response in 16%. Hematologic istered dose, possibly secondary to a reduction in the serum levels toxicity included thrombocytopenia (76%), neutropenia (44%), of thyroxine-binding globulin. Alterations in coagulation param- and anemia (39%) and was typically manageable by dose reduc- eters with prolonged thrombin time, prothrombin time, and partial tion. Nonhematologic adverse events were mostly grade 1/2 and thromboplastin time have been observed. Patients treated with l-as- included infection, diarrhea, and nausea.9 paraginase are at an increased risk for bleeding or thromboembolic 267 268 Cancer Therapeutics events.14 l-Asparaginase is contraindicated in patients with a prior can lead to death. The exact mechanism of these reactions is unclear, history of pancreatitis, because there is a 10% incidence of acute but is thought to be related to the release of endogenous pyrogens pancreatitis. Neurologic toxicity includes lethargy, confusion, agi- from the host cells. Supportive care, including hydration, steroids, an- tation, hallucinations, and/or coma. In contrast to the other anti- tipyretics, and antihistamines, may resolve the symptoms. cancer agents used to treat ALL, myelosuppression is rare. Clinicians should monitor their patients for any signs and symptoms of acute hyperpyrexic reactions during bleomycin ad- ministration. Because the onset of the reactions can occur with BLEOMYCIN any dose of bleomycin and at any time, routine test dosing does not seem to predict when drug reactions may occur.26 Mucocutane- Bleomycin is a glycopeptide antibiotic produced by the bacte- ous toxicity presents as mucositis, erythema, hyperpigmentation, rium Streptomyces verticillus. The most active chemotherapeuti- induration, hyperkeratosis, and skin peeling, which may progress 15 cal forms are bleomycin A2 and B2. The effect of bleomycin is to ulceration, and usually develops in the 2nd and 3rd week of cell cycle specific, because its main effects are mediated in the G2 treatment and after a cumulative dose of 150 to 200 U of the drug. and M phases of the cell cycle.16 The exact mechanism for DNA Levels of bleomycin are relatively low in lung and skin strand scission has been suggested to be due to bleomycin’s chelat- tissue, perhaps offering an explanation as to why these normal tis- ing of metal ions (primarily iron) and producing a pseudoenzyme sues are more adversely affected by bleomycin. Myelosuppression that reacts with oxygen to produce superoxide- and hydroxide-free and immunosuppression are relatively mild. In rare cases, vascular radicals, thus cleaving DNA. Alternatively, bleomycin may bind at events, including myocardial infarction, stroke, and Raynaud phe- specific sites in the DNA strand and induce scission by abstracting nomenon, have been reported. the hydrogen atom from the base, resulting in strand cleavage as the base undergoes a Criegee-type rearrangement, or bleomycin may form an alkali-labile lesion.17 Bleomycin is used in the treat- ment of Hodgkin lymphoma (as a component of the ABVD and BEACOPP regimen), squamous cell carcinomas, and testicular Originally prepared as a monoamine oxidase inhibitor, procarba- cancer; in the treatment of plantar warts,18 as a means of effecting zine is a prodrug, which after oxidation of the in the pleurodesis,19 as well as an intralesional agent with electrochemo- liver, undergoes a complex enzymatic and chemical breakdown to therapy in the management of cutaneous malignancies.20 its alkylating and methylating species.27,28 The precise mechanism The oral bioavailability is poor. It must be administered via IV of action is uncertain, but may involve damaging the DNA, RNA or IM routes. The initial distribution half-life is 10 to 20 minutes or transfer RNA, and the inhibition of protein synthesis. Procar- with a terminal half-life of 3 hours. Bleomycin can be administered bazine is a cell-cycle phase-nonspecific antineoplastic agent. This via the intracavitary route to control malignant pleural effusions or agent was initially approved by the U.S. Food and Drug Adminis- ascites, or both. Approximately 45% to 55% of an administered in- tration (FDA) in 1969 as part of the MOPP (mechlorethamine, tracavitary dose of bleomycin is absorbed into the systemic circula- vincristine, procarbazine, and prednisone) regimen for the treat- tion. Elimination is primarily via the kidneys, and approximately ment of Hodgkin lymphoma. Since then, it has also demonstrated 60% to 70% of an administered dose is excreted unchanged in the clinical activity in non-Hodgkin lymphoma, cutaneous T-cell lym- urine. Dose reductions are required if creatinine clearance is less phoma, and brain tumors. than 25 mL per minute. Procarbazine is rapidly and completely absorbed from the gas- Bleomycin-induced pneumonitis, the dose-limiting toxicity of trointestinal tract. Following oral administration, peak drug levels the drug, occurs in 10% of patients, and is dependent on the cumu- are reached within 10 to 15 minutes. Procarbazine crosses the lative dose.21 The risk increases in patients older than 70 years and blood–brain barrier and rapidly equilibrates between plasma and in those who receive a total cumulative dose greater than 400 U. In cerebrospinal fluid after oral administration. Peak cerebrospinal addition, patients with an underlying lung disease, prior irradiation fluid drug concentrations are reached within 30 to 90 minutes to the chest or mediastinum, and exposure to high concentrations after drug administration. The biologic half-life of procarbazine of inspired oxygen are at increased risk. Increased use of granu- hydrochloride in both plasma and cerebrospinal fluid is approxi- locyte colony-stimulating factor (G-CSF) has been paralleled by mately 1 hour. Procarbazine is metabolized to active and inactive an increased incidence of bleomycin-induced pulmonary toxicity. metabolites by chemical breakdown in an aqueous solution and The exacerbating effects of G-CSFs seem to be associated with a the liver microsomal P-450 system. Approximately 70% of procar- marked infiltration of activated neutrophils along with the lung bazine is excreted in urine within 24 hours, and less than 5% to injury caused by the direct effects of bleomycin.22,23 In a retrospec- 10% of the drug is eliminated in an unchanged form.29,30 tive review, 18% of a total of 141 patients with Hodgkin lymphoma A careful food and drug history is required before starting a pa- treated with a bleomycin-containing regimen developed pulmo- tient on procarbazine therapy, because there are several potential nary toxicity. G-CSF use was one of the key factors associated with drug–drug and drug–food interactions. Patients should avoid tyra- the development of this complication, and omission of bleomycin mine-containing foods, such as dark beer, wine, cheese, yogurt, ba- had no impact on clinical outcomes.24 Similarly the combination nanas, and smoked foods. Procarbazine produces a disulfiramlike of brentuximab vedotin and ABVD was associated with excessive reaction with concurrent use of alcohol. Acute hypertensive reac- pulmonary toxicity, indicating that brentuximab vedotin and bleo- tions may occur with coadministration of tricyclic antidepressants mycin should not be used together.25 and sympathomimetic drugs. Concurrent use of procarbazine with Patients with bleomycin-induced pulmonary toxicity may pre- antihistamines and other central nervous system (CNS) depres- sent with cough, dyspnea, dry inspiratory crackles, and infiltrates on sants can result in CNS and/or respiratory depression. chest radiograph. Pulmonary function testing is the most sensitive ap- Dose-limiting toxicity is myelosuppression, more commonly proach to monitor patients, and pulmonary function tests should be thrombocytopenia, and the nadir in platelet count is generally ob- obtained at baseline and before each cycle of therapy, with a specific served at 4 weeks. Patients with glucose-6-phosphate dehydrogenase focus on the carbon monoxide diffusion capacity and vital capacity. deficiency can develop hemolytic anemia while receiving procar- A decrease greater than 15% in either diffusion capacity of carbon bazine therapy. Stepwise dose increments over the first few days of monoxide or vital capacity should mandate immediate discontinua- drug administration may minimize gastrointestinal intolerance. On tion of bleomycin. Early clinical trials and isolated case reports sug- rare occasions, procarbazine may induce interstitial pneumonitis, gest that bleomycin-induced acute hypersensitivity reactions occur in which mandates the discontinuation of therapy. Azoospermia and 1% of patients with lymphoma and less than 0.5% of those with solid infertility after treatment with MOPP can be attributed, in part, to tumors. The reactions are mainly characterized by high-grade fever, procarbazine. Procarbazine is associated with an increased risk of chills, hypotension, and, in a few cases, cardiovascular collapse, which secondary malignancies, especially acute leukemia.

tahir99 - UnitedVRG Chapter 26 Miscellaneous Chemotherapeutic Agents 269

VISMODEGIB reactions, and cardiotoxicity. Nausea, fatigue, headaches, and ane- mia are also common. The left ventricular ejection fraction should Vismodegib (Erivedge, GDC-0449, Genentech) is a first-in-class, be monitored prior to and at least every 3 months during therapy small-molecule inhibitor of the Hedgehog pathway. It binds to and because of the potential for cardiac dysfunction. inhibits smoothened, a transmembrane protein that is involved in Hedgehog signaling.31 Pharmacodynamic downmodulation in the Hedgehog pathway was shown by a 90% decrease in transcription fac- SIROLIMUS AND TEMSIROLIMUS tor Gli1 mRNA in basal-cell carcinoma biopsy specimens of patients treated for a month. One-month vismodegib treatment also signifi- Sirolimus (rapamycin) was isolated from the soil bacteria Strepto- cantly reduced tumor proliferation, as assessed by Ki-67 expression, myces hygroscopicus, in the mid 1970s.37 This bacterial macrolide but did not change apoptosis, as assessed by cleaved caspase 3. The later became the preferred immunosuppressant for kidney trans- extent of Gli1 downmodulation does not seem to correlate with phar- plantation, because it was mildly immunosuppressive; however, in macokinetic levels of vismodegib in individual patients. Vismodegib contrast to cyclosporine A, it did not enhance tumor incidence.38 is absorbed from the gastrointestinal tract, with an oral bioavailability Sirolimus is the prototypic inhibitor of the mammalian target of 32%. Food does not affect drug exposure. Elimination is mainly he- of rapamycin (mTOR), a serine/threonine protein kinase that is patic, with excretion in feces. The median steady-state concentration a highly conserved regulatory protein involved in cell-cycle pro- is not changed by increasing the dose from 150 mg to 270 mg, and the gression, proliferation, and angiogenesis.39 Signaling pathways median time to steady state is 14 days. The half-life is estimated at 8 both upstream and downstream of mTOR have been shown to be days after a single dose. Intermittent doses (e.g., three times per week commonly dysregulated in cancer. mTOR functions through two or once per week) were associated with a decrease of 50% and 80% in main mechanisms, depending on the presence and activity of the effective plasma levels of unbound drug, respectively, thus reinforcing mTOR-associated protein complexes, mTORC1 and mTORC2. the recommended dose and schedule of 150 mg orally daily.32 Vismo- Sirolimus and its analog compounds, temsirolimus and everoli- degib is approved for the treatment of adults with metastatic basal-cell mus, form a complex with the FK-binding protein (FKBP) and carcinoma that has recurred following surgery or in those who are inhibit activation of a subset of mTOR proteins residing within not candidates for surgery and who are not candidates for radiation.33 mTORC1. In contrast, mTORC2 holds mTOR in a form that is No dose-limiting toxic effects or grade 5 events have been ob- not as readily inhibited by these rapamycin analogs, and upregula- served. However, 54% of patients receiving vismodegib discontin- tion of mTORC2 may represent a mechanism by which resistance ued the owing to side effects, and only one out of give can develop to this class of compounds. eligible patients was able to continue vismodegib for 18 months. Temsirolimus (CCI-779), a novel functional ester of sirolimus, Abdominal pain, fatigue, weight loss, dysgeusia, and anorexia were is a water-soluble dihydroxymethyl propionic acid compound that reasons for discontinuation of the drug. When vismodegib was rapidly undergoes hydrolysis to sirolimus after IV administration, withdrawn, dysgeusia and muscle cramps ceased within 1 month, reaching peak concentrations within 0.5 to 2.0 hours.40 This drug CANCER THERAPEUTICS and scalp and body hair started to regrow within 3 months. Other is widely distributed in tissues, and steady-state drug levels are side effects reported include hyponatremia, dyspnea, muscle reached in 7 to 8 days. Temsirolimus is metabolized primarily in spasm, atrial fibrillation, aspiration, back pain, corneal abrasion, the liver by CYP3A4 microsomal enzymes to yield sirolimus as the dehydration, keratitis, lymphopenia, pneumonia, urinary tract in- main metabolite. The terminal half-life of temsirolimus is 17 hours, fection, and a prolonged QT interval.34 whereas that of sirolimus is approximately 55 hours. When bound to temsirolimus, mTOR is unable to phosphorylate the key protein translation factors, such as 4E-BP1 and S6K1, leading to transla- ADO-TRASTUZUMAB EMTANSINE tional inhibition of several critical regulatory proteins involved in cell-cycle control. Several other cellular proteins involved in the Ado-trastuzumab emtansine (T-DM1), is a HER2-targeted anti- regulation of angiogenesis, such as hypoxia-inducible factor-1α body-drug conjugate (ADC). It is a novel compound composed of (HIF-1α) and vascular endothelial growth factor (VEGF), are sup- trastuzumab, a stable thioether linker, and DM1. DM1, a derivative pressed through mTOR inhibition by temsirolimus. of maytansine, is a microtubule polymerization inhibitor with activ- Phase I studies of temsirolimus have investigated various sched- ity similar to that of vinca alkaloids. T-DM1 is taken up into cells ules and doses, ranging from 7.5 mg to 220 mg given as weekly after binding to HER2, allowing for cytotoxic drug delivery specifi- 30-minute infusions.40 A phase II study in patients with cytokine- cally to cells overexpressing HER2. It has a drug-to-antibody ratio refractory renal cell cancer (RCC) investigated the efficacy and of approximately 3.5:1. T-DM1 is administered intravenously every safety of three different dose levels (25 mg, 75 mg, and 250 mg, re- 3 weeks and has been tested in a phase I trial at doses ranging from spectively) administered on a weekly schedule. This study showed 0.3 to 4.8 mg/kg. The maximally tolerated dose is 3.6 mg/kg, which promising antitumor activity for all three dose levels with no signifi- was the dose used in further phase II–III trials. T-DM1 is metabo- cant difference in efficacy or toxicity.41 As a result, the 25-mg dose lized by the liver, via CYP3A4/5, and has a half-life of 3.5 days.35 was eventually selected as the monotherapy dose for further study. A T-DM1 is approved for use in patients with metastatic HER2- phase III randomized trial compared interferon, temsirolimus, and positive breast cancer who have received prior trastuzumab and a the combination of the two agents in previously untreated patients taxane. This approval was based on the results of the EMILIA trial, with advanced RCC who had at least three of six poor prognostic which randomized 991 patients with HER2-positive unresectable, features.42 Once-weekly IV temsirolimus, 25 mg, prolonged the locally advanced or metastatic breast cancer to T-DM1 3.6 mg/ median overall survival of patients with poor prognostic features by kg IV every 21 days or lapatinib 1,250 mg daily plus capecitabine 49% from 7.3 months (95% CI, 6.1 to 8.8 months) in the interferon 1,000 mg/m2 on days 1 through 14 every 21 days. All patients arm to 10.9 months (95% CI, 8.6 to 12.7 months) in the temsiroli- were previously treated with trastuzumab and a taxane. T-DM1 mus arm (P = .008). The temsirolimus arm also had a prolonged resulted in a progression-free survival of 9.6 months compared median progression-free survival of 5.5 months compared to 3.1 to 6.4 months for lapatinib plus capecitabine (hazard ratio [HR] months in the interferon arm (P <.001). Moreover, temsirolimus 0.65; 95% confidence interval [CI] 0.55 to 0.77; p< 0.001). The was effective for both clear cell and non–clear cell histologies.43,44 response rate and overall survival was also higher with T-DM1 Mantle cell lymphoma was the first hematologic malignancy in compared to lapatinib plus capecitabine.36 which mTOR inhibition was explored as a treatment strategy. The Although maytansine itself is associated with significant toxic- rationale for this approach was that mantle cell lymphoma is char- ity, T-DM1 is very well tolerated overall, which is likely due to the acterized by overexpression of cyclin D1, which is a cyclin whose targeted nature of the compound. Side effects from T-DM1 in- expression appears to be tightly regulated by mTOR signaling. clude thrombocytopenia, hepatotoxicity, hypersensitivity/infusion The early-phase clinical trials of temsirolimus showed promising 270 Cancer Therapeutics

activity against non-Hodgkin lymphomas, multiple myeloma, and immunomodulating, antiangiogenic, and epigenetic effects. They myeloid leukemias, with some evidence of success thus far.45 are classified as class I (non–phosphophodiesterase-4 inhibitory) In terms of the safety profile, the most common adverse events immunomodulatory drugs (IMiDs). Although their primary mecha- associated with temsirolimus were asthenia and fatigue, dry skin with nism of activity against malignancy is uncertain, it is believed that acneiform skin rash, nausea/vomiting, mucositis, and anorexia. Hy- IMiDs exert their anticancer effects both directly on cancer cells perlipidemia with increased serum triglycerides and/or cholesterol and indirectly via effects on the tumor microenvironment and host as well as hyperglycemia occur in up to 90% of patients. Allergic, hy- antitumor immunity. Specific mechanisms include the inhibition persensitivity reactions have been observed in about 10% of patients, of nuclear factor kappa B (NF-κB) transcriptional activity in ma- and pulmonary toxicity, presenting as increased cough, dyspnea, lignant cells with a resultant decrease in the production of anti- fever, and pulmonary infiltrates, is a relatively rare event, occurring apoptotic molecules; the inhibition of surface adhesion molecule in less than 1% of patients. However, the risk of pulmonary toxicity expression on both multiple myeloma cells and bone marrow stro- increases in patients with an underlying pulmonary disease.46 mal cells; the inhibition of the production and release of various growth factors (including vascular endothelial growth factor, basic fibroblast growth factor, tumor necrosis factor alpha, and interleu- EVEROLIMUS kin [IL] 6) that regulate angiogenesis and tumor cell proliferation; and costimulation of IL-2 and interferon gamma (IFN-γ) release Everolimus (RAD001) is an orally active hydroxyethyl ether analog with T-helper 1 subset skewing and augmentation of cytotoxic T-cell of rapamycin that contains a 2-hydroxyethyl chain substitution. This and natural killer cell effector function.56,57 Unlike thalidomide, molecule is significantly more water soluble than sirolimus. As with both lenalidomide and pomalidomide result in cell cycle arrest and sirolimus and temsirolimus, everolimus targets mTOR by forming a apoptosis of myeloma cells in vitro, believed in part to be related to complex with mTOR and FKBP, resulting in inhibition of mTOR epigenetic effects.58 They are also more potent stimulators of IL-2 activity. Few data are available regarding the actual differences in the and INF-γ production and T-cell proliferation than thalidomide, ability of temsirolimus and everolimus to inhibit mTOR. One pre- and appear to additionally inhibit T-regulatory cells.57 Clinically, clinical in vitro study showed that the binding of everolimus to FKBP lenalidomide has activity in patients with thalidomide-resistant was approximately threefold weaker than that of sirolimus.47 In vivo multiple myeloma, and pomalidomide has additional activity in pa- studies, however, have documented similar efficacy of the two agents tients with lenalidomide-resistant disease.59,60 Recently, the protein in terms of immunosuppressive activity as well as antitumor activity. cereblon (cerebral protein with lon protease), a highly conserved In preclinical models, the administration of everolimus results in the E3 , was recognized as a primary target of IMiDs teratogenic inhibition of mTOR, similar to what has been observed with the effect, and appears to be an important target of IMiD anticancer ac- other rapamycin analogs.48 In terms of clinical pharmacology, peak tivity.61–63 Efforts are currently under way to evaluate the expression drug levels are achieved within 1 to 2 hours after oral administration, of Cereblon as a predictive biomarker of response to IMiDs.63 Due and food with a high fat content reduces oral bioavailability by up to the potential risk of significant teratogenicity, thalidomide, le- to 20%. This compound is metabolized in the liver, mainly by the nalidomide, and pomalidomide can only be prescribed by licensed CYP3A4 system, and six main metabolites have been identified. In prescribers who are registered in restricted distribution programs. general, these metabolites are less active than the parent compound. Elimination is mainly hepatic with excretion in feces, and caution should be used in patients with moderate liver impairment (Child- Thalidomide Pugh class B).49 In this setting, the daily dose of drug should be re- duced to 5 mg. In patients with severe liver dysfunction (Child-Pugh Thalidomide (2-[2,6-dioxopiperidin-3-yl]-2,3-dihydro-1H-isoindole- class C), the use of this drug is contraindicated. 1,3-dione; Thalomid) is a synthetic glutamic acid derivative that Encouraging clinical activity was initially observed in phase 1/2 trials was initially synthesized in 1953. It was used widely in Europe be- in patients with non–small-cell lung, gastric, and esophageal cancers, tween 1956 and 1962 as a sleeping aid and antiemetic for pregnant sarcomas, pancreatic neuroendocrine tumors, as well as hematologic women before it was discovered to cause severe congenital mal- malignancies.50–53 Presently, everolimus is indicated and approved for formations. Initial reports of its efficacy in multiple myeloma were the treatment of adults with advanced RCC after failure with sunitinib published in 1999, and the 200-mg daily dose combined with pulse or sorafenib; advanced hormone receptor-positive, HER2-negative dexamethasone (40-mg daily dose on days 1 through 4, 9 through breast cancer in combination with exemestane; and progressive un- 12, and 17 through 20 on a 28-day schedule) was approved by the resectable, locally advanced, or metastatic neuroendocrine tumors of FDA in 2006 for newly diagnosed multiple myeloma. The use of pancreatic origin (PNET).54,55 The recommended dose of everolimus thalidomide has dropped precipitously in the United States with for these indications is 10 mg taken orally once daily. the FDA approval of more efficacious and less toxic therapies for The safety profile of everolimus is similar to what has been myeloma. Thalidomide is poorly soluble, and it is absorbed slowly observed with temsirolimus. The most common adverse events from the gastrointestinal tract, reaching peak plasma concentra- include asthenia and fatigue, dry skin with acneiform skin rash, tion in 3 to 6 hours, with 55% to 66% bound to plasma proteins. nausea/vomiting, mucositis, and anorexia. Hyperlipidemia with The exact metabolic route and fate of thalidomide is not known. increased serum triglycerides and/or cholesterol as well as hyper- Thalidomide does not appear to be hepatically metabolized, but glycemia occur in up to 90% of patients. Allergic, hypersensitiv- rather undergoes spontaneous nonenzymatic hydrolysis in plasma ity reactions have been observed in about 10% of patients, and to multiple metabolites, with a half-life of elimination ranging from pulmonary toxicity, presenting as increased cough, dyspnea, fever, 5 to 7 hours. These metabolites are believed to be responsible for and pulmonary infiltrates, are a relatively rare event, occurring in the antitumor effects of thalidomide. Less than 1% is excreted into less than 1% of patients. However, the risk of pulmonary toxicity the urine as unchanged drug.64 increases in patients with an underlying pulmonary disease. Thalidomide frequently causes drowsiness, constipation, and fatigue. Peripheral neuropathy is a common and potentially severe and irreversible side effect occurring in up to 30% of patients. In- THALIDOMIDE, LENALIDOMIDE, AND creased incidences of venous thromboembolic events, such as POMALIDOMIDE deep venous thrombosis and pulmonary embolus, have also been observed with thalidomide, particularly when used in combina- Thalidomide and its amino-substituted analogs, lenalidomide tion with dexamethasone or anthracycline-based chemotherapy. and pomalidomide, are small-molecule glutamic acid deriva- Patients who are appropriate candidates may benefit from concur- tives that possess a wide range of biologic properties, including rent prophylactic anticoagulation or aspirin treatment.65 Other side

tahir99 - UnitedVRG Chapter 26 Miscellaneous Chemotherapeutic Agents 271

effects of thalidomide include rash, nausea, dizziness, orthostatic hy- myelosuppression in the form of neutropenia and thrombocyto- potension, bradycardia, and mood changes. In 2013, additional alerts penia can be dose limiting. As with thalidomide, the incidence were released linking thalidomide to an increased risk of developing of thromboembolic events is significant with the combination second primary malignancies (both acute myelogenous leukemia of dexamethasone and lenalidomide. A pooled analysis of 691 and myelodysplastic syndrome) and arterial thromboembolic events. patients enrolled in two randomized studies reported a 12% in- cidence of thrombotic or thromboembolic events with the combi- Lenalidomide nation, compared with 4% with dexamethasone alone.67

Lenalidomide (3-[4-amino-1-oxo-2,3-dihydro-1H-isoindol-2-yl]pipe- ridine-2,6-dione; Revlimid) is a thalidomide derivative that shares Pomalidomide the immunomodulatory and antineoplastic properties of its parent compound. However, lenalidomide appears to be more potent in Pomalidomide (4-amino-2-[2,6-dioxopiperidin-3-yl]-2,3-dihydro-1H- vitro with less nonhematologic toxicities in clinical studies. It ini- isoindole-1,3-dione; Pomalyst) is another thalidomide derivative tially received FDA approval (10-mg daily dose) in 2005 for the treat- designed to be more portent and less toxic than both thalidomide ment of patients with transfusion-dependent anemia secondary to and lenalidomide. It is currently FDA approved (4-mg once daily low or intermediate risk myelodysplastic syndromes associated with dose orally on days 1 through 21 of a 28-day cycle, with or with- a deletion 5q cytogenetic abnormality, with or without additional out dexamethasone) for use in patients with progressive multiple cytogenetic abnormalities. In 2006, lenalidomide (25-mg daily dose myeloma who have received at least two prior therapies, including on days 1 through 21 of a 28-day cycle) in combination with dexa- lenalidomide and bortezomid. Pomalidomide is administered orally methasone (40-mg daily dose on days 1 through 4, 9 through 12, and is rapidly absorbed. Maximum plasma concentration is reached and 17 through 20 on each 28-day cycle for the first four cycles, then 2 to 3 hours after ingestion, with approximately 12% to 44% protein 40 mg daily on days 1 through 4 every 28 days) was approved by the binding.68 The half-life of elimination is between 7.5 and 9.5 hours. FDA for the treatment of patients with multiple myeloma who had Pomalidomide is metabolized in the liver,via CYP1A2/CYP3A4 received at least one prior therapy for multiple myeloma. In 2013, (major) and CYP2C19/CYP2D6 (minor), and excretion occurs pri- lenalidomide 25 mg daily (days 1 through 21 on repeated 28-day marily through the kidneys (73%; 2% as unchanged drug). cycles) was additionally approved for use in refractory mantle cell Like lenalidomide, pomalidomide is better tolerated than tha- lymphoma (after relapse/ progression on two lines of therapy, one of lidomide at approved doses with less constipation, fatigue, and which contained bortezomid). Lenalidomide is administered orally neuropathy.69 The primary toxicity appreciated in myeloma trials and is rapidly absorbed from the gastrointestinal tract. Maximum has been myelosuppression, particularly neutropenia, which can plasma concentration is reached 0.625 to 1.5 hours after dosing, be dose limiting. The risk of thromboembolic events is similar to with approximately 30% bound to plasma proteins. The half-life of that seen with thalidomide and lenalidomide. Unlike thalidomide elimination is approximately 3 hours, with little information cur- or lenalidomide, dermatologic toxicity is rare with pomalidomide. CANCER THERAPEUTICS rently available concerning metabolism. Approximately 70% of an A summary of the characteristics of the miscellaneous drugs men- administered dose is excreted unchanged by the kidneys.66 tioned in this chapter is provided in Table 26.1. A summary of all Compared with thalidomide, lenalidomide is associated with hematology oncology drug approvals since the last edition of the less sedation, constipation, and peripheral neuropathy. However, textbook can be viewed in Table 26.2.

TABLE 26.1 Miscellaneous Chemotherapeutic Agents

Main Therapeutic Uses Clinical Pharmacology Major Toxicities Notes Omacetaxine CML Mean half-life of 6 h after Thrombocytopenia, Efficacy shown in subcutaneous injection anemia, nausea, diarrhea Bcr-Abl–mutated CML L-Asparaginase Pediatric and adult ALL Peak concentration 7–12 Hypersensitivity Myelosuppression is h after IV administration; reactions, alterations rare; hypersensitivity 30% plasma protein in thyroid function, reaction risk increases binding; PEG form has prolonged PT/PTT, with repeated exposure longer half-life of 5.7 decreased levels of and when used as single days; antagonize effects vitamin K–dependent agent; PEG form is less of methotrexate if given factors, acute pancreatitis immunogenic before or concurrently Bleomycin Hodgkin disease, Terminal half-life of 3h; Pulmonary toxicity dose- Not myelosuppressive; neoplastic pleural can be given intracavitary; limiting; more if age >70 immunosuppressive; effusion, non-Hodgkin 45%–55% of intracavitary y; cumulative dose >400 metabolizing enzyme; lymphoma, squamous dose absorbed U; acute hypersensitivity bleomycin hydrolase cell carcinoma of cervix, systemically; elimination reactions rare (1%); enzyme low in lung and squamous cell carcinoma via kidneys if CrCl mucositis, erythema, skin tissue; G-CSF use of nasopharynx, <25–35 mL/min dose hyperpigmentation seems to exacerbate squamous cell carcinoma reduction required pulmonary toxicity of penis, squamous cell carcinoma of the head and neck, squamous cell carcinoma of vulva, testicular cancer (continued) 272 Cancer Therapeutics

TABLE 26.1 Miscellaneous Chemotherapeutic Agents (continued)

Main Therapeutic Uses Clinical Pharmacology Major Toxicities Notes Procarbazine Hodgkin lymphoma Rapid complete oral Dose-limiting toxicity Avoid tyramine- absorption; peak is myelosuppression, containing foods; concentration, 10–15 more commonly disulfiramlike reaction min; crosses blood–brain thrombocytopenia with concurrent alcohol barrier; half-life 1 h; nadir at 4 wk; G-6PD– use; hypertensive several drug–drug and deficient patients can reaction with concurrent food–drug interactions; develop hemolytic tricyclic antidepressant metabolized by hepatic anemia, nausea, use; increased risk microsomal P-450 system; vomiting, diarrhea, for azoospermia/ 70% excreted in urine flulike symptoms, infertility and secondary peripheral neuropathy, malignancy hypersensitivity reactions Vismodegib Basal cell carcinoma of Oral bioavailability 32%; No dose limiting toxicity; Hedgehog-signaling the skin not affected by food abdominal pain, fatigue, pathway inhibitor weight loss, dysgeusia Ado- Advanced HER2-positive Peak concentration Thrombocytopenia, Monitor cardiac function trastuzumabemtansine breast cancer near the end of infusion hepatotoxicity, cardiac metabolized by CYP3A4/5; toxicity fatigue, nausea half-life, 3.5 days Temsirolimus Advanced renal cancer Peak concentration, 0.5–2 Asthenia, fatigue, Efficacy shown for both h; widely distributed in dry skin, acneiform clear cell and non– tissues; steady-state skin rash, mucositis, clear-cell histologies; levels reached in 7–8 d; anorexia, hyperlipidemia, efficacy in hematologic half-life, 17 h hyperglycemia malignancies (mantle cell lymphoma, non-Hodgkin lymphoma, multiple myeloma) Everolimus Advanced renal cell Peak concentration, 1–2 Asthenia, dry Contraindicated in Child- carcinoma, breast hr; reduced bioavailability skin, nausea, Pugh class C patients; cancer, pancreatic with high fat content food; vomiting, mucositis, encouraging activity neuroendocrine tumor metabolized by CYP3A4 hyperlipidemia, in gastric, non–small- system; mainly hepatic hyperglycemia, allergic cell, lung, esophageal excretion hypersensitivity reaction, cancers, sarcomas; pulmonary toxicity approved for organ rejection prophylaxis Thalidomide Multiple myeloma, Oral absorption slow; Drowsiness, Pregnancy category X; erythema nodosum peak concentration, 3–6 constipation, fatigue, may be present in semen; leprosum h; 55%–66% bound to skin rash, increased serious skin reactions plasma proteins; half-life, risk for thromboembolic including Stevens- 5–7 h; spontaneous complications Johnson syndrome nonenzymatic hydrolysis in plasma Lenalidomide Low-to-intermediate Rapid oral absorption; Less sedation, Pregnancy category X; risk myelodysplastic peak concentration, drowsiness, constipation caution in patients with syndrome associated 0.6–1.5 h; half-life, 3 h; than thalidomide; renal function impairment; with 5q deletion, multiple 70% excreted unchanged myelosuppression; neutropenia; myeloma by kidneys thromboembolic events; thrombocytopenia may be peripheral neuropathy dose limiting Pomalidomide Multiple myeloma who Rapid oral absorption; Myelosuppression; Better tolerated than have received at least two peak concentration, 2–3 h; thromboembolic events; thalidomide; effective prior therapies half-life, 7.5 h skin toxicity rare in prior bortezomib- and lenalidomide-receiving patients

PT, prothrombin time; PTT, partial thromboplastin time; CrCl, creatinine clearance.

tahir99 - UnitedVRG Chapter 26 Miscellaneous Chemotherapeutic Agents 273

TABLE 26.2 U.S. Food And Drug Administration Hematology Oncology Drug Approvals 2010–2013

Drug/Manufacturer Indication Approval Date Sorafenib (NEXAVAR tablets, Bayer Healthcare For the treatment of locally recurrent or metastatic, November 22, 2013 Pharmaceuticals Inc.) progressive, differentiated thyroid carcinoma (DTC) refractory to radioactive iodine treatment. Crizotinib (Xalkori, Pfizer, Inc.) capsules For the treatment of patients with metastatic non–small- November 20, 2013 cell lung cancer (NSCLC) whose tumors are anaplastic lymphoma kinase (ALK) positive as detected by an FDA- approved test. Ibrutinib (IMBRUVICA, Pharmacyclics, Inc.) For the treatment of patients with mantle cell lymphoma November 13, 2013 (MCL) who have received at least one prior therapy. Obinutuzumab (GAZYVA injection, for For use in combination with for the treatment November 1, 2013 intravenous use, Genentech, Inc.; previously of patients with previously untreated chronic lymphocytic known as GA101) leukemia (CLL). Pertuzumab injection (PERJETA, Genentech, Inc.) For use in combination with trastuzumab and docetaxel for September 30, 2013 the neoadjuvant treatment of patients with HER2-positive, locally advanced, inflammatory, or early stage breast cancer (either greater than 2 cm in diameter or node positive) as part of a complete treatment regimen for early breast cancer. Paclitaxel protein-bound particles (albumin- In combination with gemcitabine for the first-line treatment September 6, 2013 bound) (Abraxane for injectable suspension, of patients with metastatic adenocarcinoma of the Abraxis BioScience, LLC, a wholly owned pancreas. subsidiary of Celgene Corporation) Afatinib (Gilotrif tablets, Boehringer Ingelheim For the first-line treatment of patients with metastatic July 12, 2013 Pharmaceuticals, Inc.) NSCLC whose tumors have epidermal growth factor receptor (EGFR) exon 19 deletions or exon 21 (L858R) CANCER THERAPEUTICS substitution mutations as detected by an FDA-approved test. The safety and efficacy of afatinib have not been established in patients whose tumors have other EGFR mutations. Denosumab (Xgeva injection, for subcutaneous For the treatment of adults and skeletally mature June 13, 2013 use, Amgen Inc.) adolescents with a giant cell tumor of bone that is unresectable or where surgical resection is likely to result in severe morbidity. Lenalidomide capsules (REVLIMID, Celgene For the treatment of patients with MCL whose disease has June 5, 2013 Corporation) relapsed or progressed after two prior therapies, one of which included bortezomib. Trametinib (MEKINIST tablet, GlaxoSmithKline, For the treatment of patients with unresectable or May 29, 2013 LLC) metastatic melanoma with BRAF V600E or V600K mutation as detected by an FDA-approved test. Dabrafenib (TAFINLAR capsule, GlaxoSmithKline, For the treatment of patients with unresectable or May 29, 2013 LLC) metastatic melanoma with BRAF V600E mutation as detected by an FDA-approved test. Radium Ra 223 dichloride (Xofigo Injection,Bayer For the treatment of patients with castration-resistant May 15, 2013 HealthCare Pharmaceuticals Inc.) prostate cancer, symptomatic bone metastases, and no known visceral metastatic disease. Erlotinib (Tarceva, Astellas Pharma Inc.) For the first-line treatment of metastatic NSCLC patients May 14, 2013 whose tumors have EGFR exon 19 deletions or exon 21 (L858R) substitution mutations. Ado-trastuzumab emtansine (KADCYLA for For use as a single agent for the treatment of patients with February 22, 2013 injection, Genentech, Inc.) HER2-positive, metastatic breast cancer who previously received trastuzumab and a taxane, separately or in combination. Pomalidomide (POMALYST capsules, Celgene For the treatment of patients with multiple myeloma February 8, 2013 Corporation) who have received at least two prior therapies, including lenalidomide and bortezomib, and have demonstrated disease progression on or within 60 days of completion of the last therapy. (continued) 274 Cancer Therapeutics

TABLE 26.2 U.S. Food And Drug Administration Hematology Oncology Drug Approvals 2010–2013 (continued)

Drug/Manufacturer Indication Approval Date Doxorubicin hydrochloride liposome injection For the treatment of ovarian cancer in patients whose February 4, 2013 (Sun Pharma Global FZE), a generic version disease has progressed or recurred after platinum-based of DOXIL Injection (doxorubicin hydrochloride chemotherapy and for AIDS-related Kaposi sarcoma after liposome; Janssen Products, L.P.) failure of prior systemic chemotherapy or intolerance to such therapy. Bevacizumab (Avastin, Genentech U.S., Inc.) For use in combination with fluoropyrimidine–irinotecan- or January 23, 2013 fluoropyrimidine–oxaliplatin-based chemotherapy for the treatment of patients with metastatic colorectal cancer (mCRC) whose disease has progressed on a first-line bevacizumab-containing regimen. Ponatinib (Iclusig tablets, ARIAD For the treatment of adult patients with chronic phase, December 17, 2012 Pharmaceuticals, Inc.) accelerated phase, or blast phase chronic myeloid leukemia (CML) that is resistant or intolerant to prior tyrosine–kinase inhibitor (TKI) therapy or Philadelphia chromosome–positive acute lymphoblastic leukemia (Ph+ ALL) that is resistant or intolerant to prior TKI therapy. Abiraterone acetate (Zytiga Tablets, Janssen In combination with prednisone for the treatment of patients December 10, 2012 Biotech, Inc.) with metastatic castration-resistant prostate cancer. Cabozantinib (COMETRIQ capsules, Exelixis, For the treatment of patients with progressive metastatic November 29, 2012 Inc.) medullary thyroid cancer (MTC).Cabozantinib is a small molecule that inhibits the activity of multiple tyrosine kinases, including RET, MET, and VEGF receptor 2. Omacetaxine mepesuccinate (SYNRIBO For the treatment of adult patients with chronic or October 26, 2012 for injection, for subcutaneous use, Teva accelerated phase CML with resistance and/or intolerance Pharmaceutical Industries Ltd.) to two or more TKIs. Paclitaxel protein-bound particles for injectable For use in combination with carboplatin for the initial October 11, 2012 suspension, albumin-bound (ABRAXANE for treatment of patients with locally advanced or metastatic injectable suspension; Abraxis Bioscience a NSCLC who are not candidates for curative surgery or wholly owned subsidiary of Celgene Corporation) radiation therapy. Regorafenib (Stivarga tablets, Bayer HealthCare For the treatment of patients with mCRC who have been September 27, 2012 Pharmaceuticals, Inc.) previously treated with fluoropyrimidine-, oxaliplatin-, and irinotecan-based chemotherapy, an anti-VEGF therapy, and, if KRAS wild-type, an anti-EGFR therapy. Bosutinib tablets (Bosulif, Pfizer, Inc.) for the treatment of chronic, accelerated, or blast phase September 4, 2012 Ph+ CML in adult patients with resistance or intolerance to prior therapy. Enzalutamide (XTANDI Capsules, Medivation, For the treatment of patients with metastatic castration- August 31, 2012 Inc., and Astellas Pharma US, Inc.) resistant prostate cancer who have previously received docetaxel. Everolimus tablets for oral suspension (Afinitor For the treatment of pediatric and adult patients with August 30, 2012 Disperz, Novartis Pharmaceuticals Corp.) tuberous sclerosis complex (TSC) who have subependymal giant cell astrocytoma (SEGA) that requires therapeutic intervention, but that cannot be curatively resected. Vincristine sulfate LIPOSOME injection (Marqibo, For the treatment of adult patients with Ph- ALL in second August 9, 2012 Talon Therapeutics, Inc.) or greater relapse or whose disease has progressed following two or more antileukemia therapies. Ziv-aflibercept injection (ZALTRAP, Sanofi U.S., For use in combination with 5-fluorouracil, leucovorin, August 3, 2012 Inc.) irinotecan (FOLFIRI) for the treatment of patients with mCRC that is resistant to or has progressed following an oxaliplatin-containing regimen. Everolimus tablets (Afinitor, Novartis For the treatment of postmenopausal women with July 20, 2012 Pharmaceuticals Corporation) advanced hormone receptor–positive, HER2-negative breast cancer in combination with exemestane, after failure of treatment with letrozole or anastrozole. (continued)

tahir99 - UnitedVRG Chapter 26 Miscellaneous Chemotherapeutic Agents 275

TABLE 26.2 U.S. Food And Drug Administration Hematology Oncology Drug Approvals 2010–2013 (continued)

Drug/Manufacturer Indication Approval Date Carfilzomib injection (Kyprolis, Onyx For the treatment of patients with multiple myeloma July 20, 2012 Pharmaceuticals) who have received at least two prior therapies, including bortezomib and an immunomodulatory agent, and have demonstrated disease progression on or within 60 days of the completion of the last therapy. Cetuximab (Erbitux, ImClone LLC, a wholly For use in combination with FOLFIRI for first-line treatment July 9, 2012 owned subsidiary of Eli Lilly and Co.) of patients with K-ras mutation-negative (wild-type), EGFR- expressing mCRC as determined by FDA-approved tests for this use. Pertuzumab injection (PERJETA, Genentech, Inc.) For use in combination with trastuzumab and docetaxel for June 8, 2012 the treatment of patients with HER2-positive metastatic breast cancer who have not received prior anti-HER2 therapy or chemotherapy for metastatic disease. Pazopanib tablets (VOTRIENT, a registered For the treatment of patients with advanced soft tissue April 26, 2012 Trademark of GlaxoSmithKline) sarcoma (STS) who have received prior chemotherapy. Everolimus (Afinitor tablets, Novartis) For the treatment of adults with renal angiomyolipoma, April 26, 2012 associated with TSC who do not require immediate surgery. Imatinib mesylate tablets (Gleevec, Novartis For the adjuvant treatment of adult patients following January 31, 2012 Pharmaceuticals) complete gross resection of Kit (CD117)-positive gastrointestinal stromal tumors (GIST). Vismodegib (ERIVEDGE Capsule, Genentech, For the treatment of adults with metastatic basal cell January 30, 2012 Inc.) carcinoma or with locally advanced basal cell carcinoma that has recurred following surgery or who are not candidates for surgery and who are not candidates for radiation. CANCER THERAPEUTICS Axitinib tablets (Inlyta, Pfizer, Inc.) For the treatment of advanced renal cell carcinoma after January 27, 2012 failure of one prior systemic therapy. Glucarpidase injection (Voraxaze, BTG For the treatment of toxic plasma methotrexate January 17, 2012 International Inc.) concentrations (> 1 μmol/L) in patients with delayed methotrexate clearance due to impaired renal function. Asparaginase Erwinia chrysanthemi (Erwinaze, As a component of a multiagent chemotherapeutic regimen November 18, 2011 injection, EUSA Pharma [USA], Inc.) for the treatment of patients with ALL who have developed hypersensitivity to E. coli–derived asparaginase. Ruxolitinib (Jakafi oral tablets, Incyte Corporation) For the treatment of intermediate and high risk November 16, 2011 myelofibrosis, including primary myelofibrosis, postpolycythemia vera myelofibrosis, and postessential thrombocythemia myelofibrosis. Cetuximab (Erbitux, ImClone LLC, a wholly- In combination with platinum-based therapy plus November 7, 2011 owned subsidiary of Eli Lilly and Company) 5-fluorouracil (5-FU) for the first-line treatment of patients with recurrent locoregional disease and/or metastatic squamous cell carcinoma of the head and neck (SCCHN). Eculizumab (Soliris, Alexion, Inc.) For the treatment of pediatric and adult patients with September 23, 2011 atypical hemolytic uremic syndrome (aHUS). Denosumab (Prolia, Amgen Inc.) As a treatment to increase bone mass in patients at high September 16, 2011 risk for fracture receiving androgen-deprivation therapy (ADT) for nonmetastatic prostate cancer or adjuvant aromatase inhibitor (AI) therapy for breast cancer. Crizotinib (XALKORI Capsules, Pfizer Inc.) For the treatment of patients with locally advanced or August 26, 2011 metastatic NSCLC that is ALK-positive as detected by an FDA-approved test. (continued) 276 Cancer Therapeutics

TABLE 26.2 U.S. Food And Drug Administration Hematology Oncology Drug Approvals 2010–2013 (continued)

Drug/Manufacturer Indication Approval Date Brentuximab vedotin (Adcetris for injection, For treatment of patients with Hodgkin lymphoma after August 19, 2011 Seattle Genetics, Inc.) failure of autologous stem cell transplant (ASCT) or after failure of at least two prior multiagent chemotherapy regimens in patients who are not ASCT candidates and treatment of patients with systemic anaplastic large cell lymphoma (ALCL) after failure of at least one prior multiagent chemotherapy regimen. Vemurafenib tablets (ZELBORAF, Hoffmann-La For the treatment of patients with unresectable or August 17, 2011 Roche Inc.) metastatic melanoma with the BRAFV600E mutation as detected by an FDA-approved test. Sunitinib (Sutent capsules, Pfizer, Inc.) For the treatment of progressive, well-differentiated May 20, 2011 pancreatic neuroendocrine tumors (pNET) in patients with unresectable, locally advanced, or metastatic disease. Everolimus (Afinitor tablets, Novartis For the treatment of progressive PNET in patients with May 5, 2011 Pharmaceuticals Corporation) unresectable, locally advanced, or metastatic disease. Abiraterone acetate (Zytiga tablets, Centocor For use in combination with prednisone for the treatment of April 28, 2011 Ortho Biotech, Inc.) patients with metastatic castration-resistant prostate cancer (mCRPC) who have received prior chemotherapy containing docetaxel. Vandetanib tablets (Vandetanib tablets, For the treatment of symptomatic or progressive medullary April 6, 2011 AstraZeneca Pharmaceuticals LP) thyroid cancer in patients with unresectable, locally advanced, or metastatic disease. Peginterferon alfa-2b (Sylatron, Schering For the treatment of patients with melanoma with March 29, 2011 Corporation, Kenilworth, NJ 07033) microscopic or gross nodal involvement within 84 days of definitive surgical resection including complete lymphadenectomy. Ipilimumab injection (YERVOY, Bristol-Myers For the treatment of unresectable or metastatic melanoma. March 25, 2011 Squibb Company) Rituximab (Rituxan, Genentech, Inc.) For maintenance therapy in patients with previously January 28, 2011 untreated follicular, CD-20 positive, B-cell non-Hodgkin lymphoma who achieve a response to rituximab in combination with chemotherapy. mesylate (Halaven injection, Eisai Inc.) For the treatment of patients with metastatic breast cancer November 15, 2010 who have previously received an anthracycline and a taxane in either the adjuvant or metastatic setting, and at least two chemotherapeutic regimens for the treatment of metastatic disease. Everolimus (Afinitor, Novartis), an mTOR inhibitor For patients with SEGA associated with tuberous sclerosis October 29, 2010 (TS) who require therapy but who are not candidates for surgical resection. Dasatinib (Sprycel, Bristol-Myers Squibb) For the treatment of newly diagnosed adult patients with October 28, 2010 Ph+ CML in chronic phase (CP-CML). Trastuzumab (Herceptin, Genentech, Inc.) In combination with cisplatin and a fluoropyrimidine October 20, 2010 (capecitabine or 5-FU), for the treatment of patients with HER2-overexpressing metastatic gastric or gastroesophageal (GE) junction adenocarcinoma, who have not received prior treatment for metastatic disease. Nilotinib (Tasigna capsules, Novartis For the treatment of adult patients with newly diagnosed June 17, 2010 Pharmaceuticals Corporation) Ph+ CP-CML. Cabazitaxel (Jevtana injection, Sanofi-Aventis) For use in combination with prednisone for treatment of June 17, 2010 patients with metastatic hormone-refractory prostate cancer (mHRPC) previously treated with a docetaxel-containing regimen.

tahir99 - UnitedVRG Chapter 26 Miscellaneous Chemotherapeutic Agents 277

SELECTED REFERENCES

The full reference list can be accessed at lwwhealthlibrary.com/oncology. 44. Hudes G, Carducci M, Tomczak P, et al. Temsirolimus, interferon alfa, or both for advanced renal-cell carcinoma. N Engl J Med 2007;356:2271–2281. 3. Baaske DM, Heinstein P. Cytotoxicity and cell cycle specificity of homohar- 47. Schuler W, Sedrani R, Cottens S, et al. SDZ RAD, a new rapamycine de- rintonine. Antimicrob Agents Chemother 1977;12:298–300. rivative: pharmacological properties in vitro and in vivo. Transplantation 9. Jorge Cortes J, Lipton JF, Rea D, et al. Phase 2 study of subcutaneous omac- 1997;64:36–42. etaxine mepesuccinate after TKI failure in patients with chronic-phase CML 48. Dudkin L, Dilling MB, Cheshire PJ, et al. Biochemical correlates of mTOR with T315I mutation. Blood 2012;120:2573–2580. inhibition by the rapamycin ester CCI-779 and tumor growth inhibition. 13. Avramis VI, Panosyan EH. Pharmacokinetic/pharmacodynamic relation- Clin Cancer Res 2001;7:1758–1764. ships of asparaginase formulations: the past, the present and recommenda- 51. Yao JC, Lombard-Bohas C, Baudin E, et al. Daily oral everolimus activity in tions for the future. Clin Pharmacokinet 2005;44:367–393. patients with metastatic pancreatic neuroendocrine tumors after failure of 16. Chen J, Stubbe J. : towards better therapeutics. Nat Rev Cancer cytotoxic chemotherapy: a phase II trial. J Clin Oncol 2010;28:69–76. 2005;2:102–112. 54. Yao JC, Shah MH, Ito T, et al. Everolimus for advanced pancreatic neuroen- 22. Azulay E, Herigault S, Levame M, et al. Effect of granulocyte colony-stimu- docrine tumors. N Engl J Med 2011;364:514–523. lating factor on bleomycin-induced acute lung injury and pulmonary fibro- 55. Baselga J, Campone M, Piccart M, et al. Everolimus in postmenopausal sis. Crit Care Med 2003;31:1442–1448. hormone-receptor-positive advanced breast cancer. N Engl J Med 2012;366: 25. Younes A, Connors JM, Park SI et al. Brentuximab vedotin combined with 520–529. ABVD or AVD for patients with newly diagnosed Hodgkin’s lymphoma: a 56. Shortt J, Hsu AK, Johnstone RW. Thalidomide-analogue biology: immu- phase 1, open-label, dose-escalation study. Lancet Oncol 2013;14:1348–1356. nological, molecular and epigenetic targets in cancer therapy. Oncogene 30. Preiss R, Baumann F, Regenthal R, et al. Plasma kinetics of procarbazine and 2013;32:4191–4202. azo-procarbazine in humans. Anticancer Drugs 2006;17:75–80. 57. Zhu YX, Kortuem KM, Stewart AK. Molecular mechanism of action of 31. Von Hoff DD, LoRusso PM, Rudin CM, et al. Inhibition of the hedge- immune-modulatory drugs thalidomide, lenalidomide and pomalidomide in hog pathway in advanced basal-cell carcinoma. N Engl J Med 2009;361: multiple myeloma. Leuk Lymphoma 2013;54:683–687. 1164-1172. 61. Ito T, Ando H, Suzuki T, et al. Identification of a primary target of thalido- 35. Krop IE, Beeram M, Modi S, et al. Phase I study of trastuzumab-DM1, an mide teratogenicity. Science 2010;327:1345–1350. HER2 antibody-drug conjugate, given every 3 weeks to patients with HER2- 62. Zhu YX, Braggio E, Shi CX, et al. Cereblon expression is required for the positive metastatic breast cancer. J Clin Oncol 2010;28:2698–2704. antimyeloma activity of lenalidomide and pomalidomide. Blood 2011;118: 37. Sehgal SN, Baker H, Vézina C. Rapamycin (AY-22,989), a new antifungal 4771–4779. antibiotic. II. Fermentation, isolation and characterization. J Antibiot (Tokyo) 65. Bennett CL, Angelotta C, Yarnold PR, et al. Thalidomide- and lenalid- 1975;28:727–732. omide-associated thromboembolism among patients with cancer. JAMA 39. Wullschleger S, Loewith R, Hall MN. TOR signaling in growth and metabo- 2006;296:2558–2560. lism. Cell 2006;124:471–484. 69. Lacy MQ, McCurdy AR. Pomalidomide. Blood 2013;122:2305–2309. CANCER THERAPEUTICS 27 Hormonal Agents

Matthew P. Goetz, Charles Erlichman, Charles L. Loprinzi, and Manish Kohli

INTRODUCTION tamoxifen, the cumulative risk is 3.1% (mortality, 0.4%) versus 1.6% (mortality, 0.2%) for 5 years of tamoxifen.4 The incidence of a rarer form of uterine cancer, uterine sarcoma, is also increased Hormonal agents are commonly used as a treatment of hormon- 15 ally responsive cancers, such as breast, prostate, or endometrial after tamoxifen use. This form of endometrial cancer comprises carcinomas. Other uses for some hormonal therapies include the approximately 15% of all uterine malignancies that develop after tamoxifen use.15 Beneficial estrogenic effects from tamoxifen in- treatment of paraneoplastic syndromes, such as carcinoid syn- 16 drome, and symptoms caused by cancer, including anorexia. This clude a decrease in total cholesterol and the preservation of bone density in postmenopausal women.17 In premenopausal women, chapter discusses the major hormonal agents for such therapy, first 18 with an overview of their use in practice, then with more detailed however, tamoxifen has a negative effect on bone density. Al- pharmacologic information regarding them (Table 27.1). though most patients do not complain of vaginal symptoms, a few complain of vaginal dryness, whereas others have increased vagi- nal secretions and discharge, the latter of which is an indication of the estrogenic activity of tamoxifen on the vagina. In the Arimidex, SELECTIVE ESTROGEN RECEPTOR Tamoxifen, Alone or in Combination (ATAC) trial, a commonly MODULATORS observed tamoxifen side effect was vaginal bleeding, leading to a higher hysterectomy rate for patients randomized to tamoxifen 19 Tamoxifen (5%) compared to anastrozole (1%). An uncommon effect from tamoxifen is retinal toxicity. This drug can also increase the risk of cataracts. However, no difference in the rate of vision-threatening Tamoxifen continues to be an important hormonal therapy for the ocular toxicity has been seen among prospectively treated tamoxi- prevention and treatment of breast cancer worldwide. The contin- fen patients.20 Tamoxifen predisposes patients to thromboembolic ued importance of tamoxifen is reflected in the fact that it is the phenomena, especially if used with concomitant chemotherapy. only hormonal agent approved by the U.S. Food and Drug Admin- Depression has also been described, but the association with istration (FDA) for the prevention of premenopausal breast can- tamoxifen is not clear. Although liver cancers have been noted cer,1 the treatment of ductal carcinoma in situ (DCIS),2 and the in laboratory animals, there is no established association between treatment of surgically resected premenopausal estrogen receptor tamoxifen and liver cancers in humans. (ER)–positive breast cancer.3 The standard daily dose of tamoxifen is 20 mg, and the optimal duration depends on the underlying clinical setting. Although the Pharmacology recommended duration in the prevention and DCIS settings is 5 years, recently published prospective studies have demonstrated Tamoxifen acts by blocking estrogen stimulation of breast cancer that for the adjuvant treatment of invasive breast cancer, a duration cells, inhibiting both translocation and nuclear binding of the ER. of 10 years (compared to 5 years) further reduced the risk of breast This alters transcriptional and posttranscriptional events mediated cancer mortality and improved overall survival.4 by this receptor.21 Tamoxifen has agonistic, partial agonistic, or an- The most common toxicity from tamoxifen is hot flashes, af- tagonistic effects depending on the species, tissue, or endpoints fecting approximately 50% of treated women. These hot flashes are that have been assessed. Additionally, there are marked differ- of varying intensity and duration. Tamoxifen-induced hot flashes ences between the antiproliferative properties of tamoxifen and its appear to increase over the first 3 months of therapy and then pla- metabolites.22 teau. They appear to be more prominent in women with a history Resistance to tamoxifen can be intrinsic or acquired, and the of hot flashes or estrogen replacement use. Tamoxifen-induced potential mechanisms for this resistance are reviewed in the fol- hot flashes can be ameliorated by a number of different pharma- lowing paragraphs. At each step of the signal transduction path- cotherapies, including low doses of megestrol5; antidepressants way with which tamoxifen or its metabolites interferes, there is such as venlafaxine,6 desvenlafaxine,7 citalopram,8 escitalopram,9 the potential for an alteration in response. The most important and paroxetine10; and the anticonvulsant drugs gabapentin11 and factor appears to be the level of ER, which is highly predictive for pregabalin.12 There is evidence that drugs that inhibit CYP2D6 a response to tamoxifen. Tamoxifen is ineffective in ER-negative (e.g., paroxetine) alter the metabolic activation of tamoxifen to breast cancer. Although decreased or absent expression of the pro- endoxifen, a critical metabolite associated with in vivo tamoxifen gesterone receptor (PR) is associated with a worse prognosis, the efficacy.13 relative risk reduction in tamoxifen-treated patients is the same The estrogenic properties of tamoxifen are responsible for both regardless of the presence or absence of the PR. beneficial and deleterious side effects. Tamoxifen increases the Following binding to the ER, subsequent translocation of the incidence of endometrial cancer in postmenopausal (but not pre- tamoxifen/ER complex to the nucleus and binding to an estrogen- menopausal) women, with the increase in the annual incidence of response element may occur. This binding prevents transcrip- endometrial cancer being approximately 2.58 (ratio of incidence tional activation of estrogen-responsive genes. Laboratory and rates).14 The absolute risk depends on the duration of tamoxi- clinical data have demonstrated that ER-positive breast cancers fen administration. For women who receive 10 years of adjuvant that overexpress HER2 may be less responsive to tamoxifen and 278

tahir99 - UnitedVRG Chapter 27 Hormonal Agents 279

TABLE 27.1 Overview of Major Hormonal Agents Used in Cancer

Class of Drug Individual Drug Dose Route of Delivery Frequency of Delivery Selective estrogen Tamoxifen 20 mg Oral Once daily receptor modulator Toremifene 60 mg Oral Once daily Raloxifene 60 mg Oral Once daily Aromatase inhibitor Anastrozole 1 mg Oral Once daily Letrozole 2.5 mg Oral Once daily Exemestane 25 mg Oral Once daily Estrogen receptor Fulvestrant 500 mg IM Once monthly downregulator Luteinizing hormone Goserelin 7. 5 IM Once monthlya releasing hormone Leuprolide 3.6 IM Once monthlya agonist GnRH antagonist Degarelix 240 mg loading dose SC 80 mg SC monthly maintenance dose Antiandrogen Flutamide 250 mg Oral Three times daily Bicalutamide 50 mg Oral Once daily Nilutamide 300 mg for 30 d then Oral Once daily 150 mg Cytochrome P45017 Abiraterone Acetate 1,000 mg (four 250 mg Oral Once Daily alpha inhibitors capsules) AR “super antagonists” Enzalutamide 160–240 mg Oral Daily Androgen Fluoxymesterone 10 mg Oral Twice daily Estrogen Estradiol 10 mg Oral Up to three times daily b Somatostatin analog Octreotide Varies SC or IV Up to three times daily CANCER THERAPEUTICS Progestational agents Megestrol Varies Oral Once daily Medroxyprogesterone Varies Oral or IM Varies acetate a Longer acting depot preparations (every 3 months) are available. b Depot formulations are available. IM, intramuscular; SC, subcutaneous; GnRH, gonadotropin-releasing hormone; CYP, cytochrome P-450; AR, androgen receptor. to hormonal therapy in general.23–26 In these tumors, ligand-inde- such metabolic steps are being evaluated for their carcinogenic pendent activation of the ER by mitogen-activated protein kinase potential in vitro and in vivo. (MAPK) pathways may contribute to resistance.27–29 In addition, Multiple studies to evaluate tumor gene expression profiling the expression of AIB1, an estrogen-receptor coactivator, has been have identified gene expression patterns or specific genes associ- associated with tamoxifen resistance in patients whose breast can- ated with resistance to tamoxifen therapy. A commonly utilized cers overexpress HER2.30 In some cases, resistance may result gene expression assay, Oncotype DX 21 gene assay (Genomic from a decrease or loss of ER expression.31,32 Although mutations Health, Redwood City, California), measures the expression of in the ER ligand binding domain (LBD) are rare in newly diag- genes known to be involved in estrogen signaling (e.g., ER, PR), nosed breast cancer, ER mutations are present in up to 20% of HER2, proliferation (e.g., Ki-67), and others. In multiple differ- recurrent breast cancers.33–36 These mutations lead to a confor- ent data sets, the recurrence score has been associated with a mational change in the LBD, which mimics the conformation of higher risk of breast cancer recurrence in patients treated with hor- activated ligand-bound receptor and constitutive, ligand-indepen- monal therapy (e.g., tamoxifen or aromatase inhibitors) without dent transcriptional activity, resulting in resistance to hormonal concomitant chemotherapy.47–49 therapy. Preclinical studies suggest that some of these mutations, The pharmacokinetics of tamoxifen is complex. The chemi- although insensitive to aromatase inhibitors, retain sensitivity to cal structure and metabolic pathway of tamoxifen are shown in higher dose selective estrogen-receptor modulators (SERM), such Figure 27.1. Metabolic activation of tamoxifen is associated with as endoxifen, as well as fulvestrant.35 greater pharmacologic activity. The two most active tamoxifen The carcinogenic potential of tamoxifen has been recognized metabolites are 4-hydroxytamoxifen (4-OH tamoxifen) and 4-OH- in rat studies37–39 and in humans (endometrial cancer).40 It has N-desmethyltamoxifen (endoxifen). A series of studies carried out been proposed that the generation of reactive intermediates that to characterize endoxifen pharmacology have demonstrated that bind covalently to macromolecules underlies the process. Such it has equivalent potency in vitro to 4-hydroxytamoxifen in ER-α reactive intermediates have been demonstrated in vitro.40–43 In ad- and -beta (ER-β) binding,50 for the suppression of ER-dependent dition, the induction of covalent DNA adducts in rat livers treated human breast cancer cell line proliferation,22,50 and in global ER- with tamoxifen has been reported.44 Both constitutive and induc- responsive gene expression.51 A recent study suggests that endoxi- ible cytochrome P-450 (CYP) enzymes have been implicated in fen’s effect on the ER may differ from 4-hydroxytamoxifen based the formation of metabolites with tamoxifen,45,46 and the flavone- on the observation of ER-α degradation.52 containing monooxygenase has been implicated in the forma- In women who receive tamoxifen at a dose of 20 mg per day, tion of the N-oxide of tamoxifen. Reactive intermediates from plasma endoxifen steady-state concentrations are generally 6 to 280 Cancer Therapeutics

CH3 OCH2CH2N CH3

C C CH3 CH2

CYP3A4 CYP3A4 CYP3A5 CYP2B6 OH CYP2C9 CYP2D6 CYP2C19 CYP2C9 Tamoxifen CYP2D6 CYP2C19 CYP1A2 CH3 OCH2CH2N CH3 CH3 OCH2CH2N H SULT1A1 C C CH3 CH2 UGTs C C CH3 CH2

OH 4-hydroxtamoxifen N-desmethyltamoxifen CYP3A4 CYP2B6 CYP2C9 CYP2C19 CH3 CYP2D6 OCH2CH2N H

C C CH3 CH2

OH 4-hydroxy-N-desmethyltamoxifen (Endoxifen)

SULT1A1 UGTs

Figure 27.1 Metabolic pathway of tamoxifen biotransformation. (From Sideras K, Ingle JN, Ames MM, et al. Coprescription of tamoxifen and that inhibit CYP2D6. J Clin Oncol 2010;28:2768–2776.)

10 times higher than 4-hydroxytamoxifen.53 Although the me- efficacy.5 However, this label change has been delayed, in part tabolism of tamoxifen to 4-OH-tamoxifen is catalyzed by multiple because of conflicting data from secondary analyses of 5-year enzymes, endoxifen is formed predominantly by the CYP2D6- tamoxifen prospective trials (ATAC,59 BIG 1-98,60 and ABCSG861) mediated oxidation of N-desmethyltamoxifen, the most abundant as well as meta-analyses,62 which demonstrate that the CYP2D6 tamoxifen metabolite (see Fig. 27.1).54 Multiple clinical studies genotype is associated with tamoxifen efficacy when tamoxifen is have demonstrated that common CYP2D6 genetic variation (lead- administered as monotherapy for the adjuvant treatment of post- ing to low or absent CYP2D6 activity) or the drug-induced inhibi- menopausal, ER-positive breast cancer. Additional support for the tion of CYP2D6 significantly lowers endoxifen concentrations.53,55 importance of endoxifen concentrations came from a secondary The CYP2D6 gene is highly polymorphic, with more than 70 analysis of a prospective study, which demonstrated a higher risk of major alleles with four well-defined phenotypes: poor metaboliz- recurrence for women with low endoxifen concentrations.13 ers (PM), intermediate metabolizers (IM), extensive metabolizers Many drugs are known to inhibit CYP2D6 activity. In tamox- (EM), and ultrarapid metabolizers (UM). ifen-treated women, the coadministration of potent CYP2D6 The clinical studies to evaluate the association between inhibitors, such as paroxetine, converts a patient with normal CYP2D6 polymorphisms and tamoxifen outcomes have yielded CYP2D6 metabolism to a phenotypic PM.63 Many other clinically conflicting results. Initial56 and follow-up data57,58 demonstrated important drugs have been reported to inhibit the CYP2D6 en- that CYP2D6 PM had an approximately two- to threefold higher zyme system, but their effects on tamoxifen metabolism have not risk of breast cancer recurrence (compared to CYP2D6 EM) and been prospectively studied. As with the data regarding CYP2D6 these data led an FDA special emphasis panel to recommend genotype, the data regarding CYP2D6 inhibitors has additionally a tamoxifen label change to incorporate data that the CYP2D6 been controversial, including two studies that reported opposite genotype was an important biomarker associated with tamoxifen findings with regard to CYP2D6 inhibitor use and breast cancer

tahir99 - UnitedVRG Chapter 27 Hormonal Agents 281 recurrence or death.64,65 Although the CYP2D6 data remain con- plasma steady-state concentrations is 1 to 5 weeks. Plasma protein troversial, we conclude that until results from prospective adjuvant binding is more than 99%. As with tamoxifen, toremifene is pres- studies are available, women should be counseled regarding the ent at higher concentrations in tissues compared to plasma with potential impact of the CYP2D6 genotype on the effectiveness a high apparent volume of distribution (958 L). Seventy percent of adjuvant tamoxifen, and potent CYP2D6 inhibitors should be of the drug is excreted in feces as metabolites. Studies in patients avoided. Additional caution should be used with drugs that induce with impaired liver function or those on anticonvulsants known CYP3A, such as rifampicin, as a these drugs have been demon- to induce CYP3A have demonstrated that hepatic dysfunction de- strated to substantially reduce (up to 86%) the concentrations of creases the clearance of toremifene and N-desmethyltoremifene,95 tamoxifen and its metabolites.66 whereas those patients on anticonvulsants had an increased clear- Strategies to overcome low endoxifen concentrations include ance. Although toremifene appeared to be less carcinogenic than dose escalation of tamoxifen to 40 mg per day, which has been tamoxifen in preclinical models,43,96,97 of the rates of endometrial demonstrated to significantly increase endoxifen concentra- cancer in the adjuvant studies have been similar to tamoxifen.85 tions,67,68 as well as the direct administration of endoxifen itself. The latter strategy is ongoing in multiple different clinical trials, Raloxifene and early reports suggest clinical activity in aromatase inhibitors (AI)-resistant breast cancer.69 Raloxifene is an estrogen agonist and antagonist originally devel- Following the metabolic activation of tamoxifen, the hy- oped to treat osteoporosis. Large placebo-controlled randomized droxylated metabolites undergo both glucuronidation and sulfa- trials demonstrated reduced rates of osteoporosis and a reduction tion. Peak plasma levels of tamoxifen (maximum concentration in new breast cancers in treated women, leading to the develop- [Cmax]) are seen 3 to 7 hours after oral administration. Assum- ment of a second-generation breast cancer chemoprevention trial ing an oral bioavailability of 30%, the volume of distribution has (National Surgical Adjuvant Breast and Bowel Project, NSAPB been calculated to be 20 L/kg, and plasma clearance ranges from P2) in which raloxifene was compared with tamoxifen in high-risk 1.2 to 5.1 L per hour.70 The terminal half-life of tamoxifen has postmenopausal women. In this study, tamoxifen was superior to been reported to range between 4 and 11 days.71,72 The elimina- raloxifene in terms of both invasive and noninvasive cancer events, tion half-life of tamoxifen increases with successive doses, which but was associated with a higher risk of thromboembolic events is consistent with saturable kinetics.71,73 The drug’s distribution in and endometrial cancer.98 tissues is extensive. Levels of the parent drug and metabolites have been reported to be higher in tissue than in plasma in animal stud- ies.74,75 Reports of tamoxifen concentrations 10- to 60-fold higher Pharmacology than plasma concentrations in the liver, lungs, brain, pancreas, skin, and bones are reported.76,77 Elevated levels of tamoxifen with Raloxifene is partially estrogenic in bone99 and lowers cholesterol.100 biliary obstruction have been reported.78 It is antiestrogenic in mammary tissue101,102 and uterine tissue.103 CANCER THERAPEUTICS Tamoxifen has been reported to interact with warfarin,73,79–81 The pharmacokinetics of raloxifene have been studied princi- digitoxin, phenytoin,82 and medroxyprogesterone.73 Tamoxifen- pally in postmenopausal women.104–106 Pharmacokinetic param- induced activation of human transcription factor pregnane X eters of raloxifene show considerable interindividual variation. receptor (hPXR), resulting in the induction of CYP3A4, may in- Limited information is available on the pharmacokinetics of ral- crease the elimination of concomitantly administered CYP3A sub- oxifene in individuals with hepatic impairment, renal impairment, strates,83 such as anastrozole.84 or both. Raloxifene is rapidly absorbed from the gastrointestinal tract. Toremifene Because raloxifene undergoes extensive first-pass glucuronidation, oral bioavailability of unchanged drug is low. Although approxi- mately 60% of an oral dose is absorbed, the absolute bioavailability Toremifene is an agent similar to tamoxifen. It is available in the as unchanged raloxifene is only 2%. However, systemic availability United States for the treatment of patients with metastatic breast of raloxifene may be greater than that indicated in bioavailability cancer, and is approved in other countries for the adjuvant treat- studies, because circulating glucuronide conjugates are converted ment of ER-positive breast cancer. Clinical trials have demon- back to the parent drug in various tissues. strated no difference in either disease-free or overall survival when After the oral administration of a single 120- or 150-mg dose toremifene was compared with tamoxifen for the treatment of of raloxifene hydrochloride, peak plasma concentrations of raloxi- ER-positive breast cancer,85,86and evidence exists for major cross- fene and its glucuronide conjugates are achieved at 6 hours and resistance between tamoxifen and toremifene.87,88 1 hour, respectively. After the oral administration of radiolabeled raloxifene, less than 1% of total circulating radiolabeled material Pharmacology in plasma represents the parent drug. Results of a single-dose study in patients with liver dysfunc- Toremifene is an antiestrogen with a chemical structure that differs tion indicate that plasma raloxifene concentrations correlate with from that of tamoxifen by the substitution of a chlorine for a hy- serum bilirubin concentrations and are 2.5 times higher than indi- drogen atom that is retained when toremifene undergoes metabo- viduals with normal hepatic function. In postmenopausal women lism.89 Like tamoxifen, toremifene is metabolized by CYP3A,90 who received raloxifene in clinical trials, plasma concentrations with a secondary metabolism to form hydroxylated metabolites of raloxifene and the glucuronide conjugates in those with renal that appear to have similar binding affinities to 4-OH tamoxi- impairment (i.e., estimated creatinine clearance values as low as fen.89,91 The importance of these metabolites or the role of metab- 23 mL per minute) were similar to values in women with normal olism to the hydroxylated metabolites is unknown, but may play a renal function. role given the structural similarity of toremifene to tamoxifen. Al- Raloxifene and its monoglucuronide conjugates are more though the oral bioavailability has not been defined, toremifene’s than 95% bound to plasma proteins. Raloxifene binds to albu- oral absorption appears to be good. The time to peak plasma con- min and α1-acid glycoprotein. Raloxifene undergoes extensive centrations after oral administration ranges from 1.5 to 6.0 hours,92 first-pass metabolism to the glucuronide conjugates raloxifene with the terminal half-lives for toremifene and one metabolite, 4′-glucuronide, 6-glucuronide, and 6,4′-diglucuronide. UGT1A1 4-hydroxytoremifene, being 5 to 6 days.93,94 The apparent clear- and -1A8 have been found to catalyze the formation of both the ance is 5.1 L per hour. The terminal half-life for the major me- 6-β-and 4′-β-glucuronides, whereas UGT1A10 formed only the tabolite, N-desmethyltoremifene, is 21 days.95 The time to reach 4′-β-glucuronide.107 The metabolism of raloxifene does not ap- 282 Cancer Therapeutics pear to be mediated by CYP enzymes (such as CYP2D6), because fulvestrant, tamoxifen, or placebo. After single intramuscular injec- metabolites other than glucuronide conjugates have not been tions of fulvestrant, the time of maximal concentration (tmax) ranged identified. from 2 to 19 days, with the median being 7 days for each dose group. The plasma elimination half-life of raloxifene at steady state At the interval of 28 days, Cmin values were two- to fivefold lower averages 32.5 hours (range, 15.8 to 86.6 hours). Raloxifene is ex- than the Cmax values. For most patients in the 125- and 250-mg creted principally in feces as an unabsorbed drug and via biliary dose groups, significant levels of fulvestrant were still measurable 84 elimination as glucuronide conjugates, which, subsequently, are days after administration. Pharmacokinetic modeling of the pooled metabolized by bacteria in the gastrointestinal tract to the parent data from the 250-mg cohort was best described by a two-compart- drug. After oral administration, less than 0.2% of a raloxifene dose ment model in which a longer terminal phase began approximately is excreted as the parent compound and less than 6% as glucuro- 3 weeks after administration. Because of the long time needed to nide conjugates in urine. reach a steady state, the 500-mg loading dose regimen was pro- spectively studied and determined to be superior to the 250 mg per Fulvestrant month dose, both in terms of steady state concentrations achieved within 1 month119as well as progression-free and overall survival.118 Fulvestrant is an ER antagonist that has no known agonist activity 108–111 and results in ER downregulation. Like tamoxifen, fulvestrant AROMATASE INHIBITORS competitively binds to the ER but with a higher affinity—approxi- mately 100 times greater than that of tamoxifen,108,112–114—thus At menopause, the synthesis of ovarian hormones ceases. How- preventing endogenous estrogen from exerting its effect in target ever, estrogen continues to be converted from androgens (pro- cells. duced by the adrenal glands) by aromatase, an enzyme of the CYP Results from two phase III clinical trials using the 250 mg per superfamily. Aromatase is the enzyme complex responsible for month dose demonstrated fulvestrant to be as effective as anastro- the final step in estrogen synthesis via the conversion of andro- zole in the treatment of postmenopausal women with advanced gens, androstenedione and testosterone, to estrogens, estrone (E ) hormone receptor–positive breast cancer previously treated with 1 and E . This biologic pathway served as the basis for the devel- antiestrogen therapy (mainly tamoxifen).112–116 In the setting of 2 opment of the antiaromatase class of compounds. Alterations in first-line hormone-responsive metastatic breast cancer, a random- aromatase expression have been implicated in the pathogenesis ized phase III clinical trial to compare tamoxifen to fulvestrant of estrogen-dependent disease, including breast cancer, endome- (250 mg per month) demonstrated no differences in response or trial cancer, and endometriosis. The importance of this enzyme is time to progression.117 Because of pharmacology data (discussed in also highlighted by the fact that selective aromatase inhibitors are the following paragraphs), the 500 mg per day dose was developed. commonly used as first-line therapy for the treatment of postmeno- A randomized trial comparing the 250 mg per month with 500 mg pausal women with estrogen-responsive breast cancer. Aminoglu- per month dose demonstrated a 4-month improvement in median tethimide was the first clinically used aromatase inhibitor. When overall survival advantage for the higher dose.118 For this reason, it became available, it was used to cause a medical adrenalectomy. the higher dose is now the standard recommended dose. Because of the lack of selectivity for aromatase and the resultant Fulvestrant is well tolerated. The most common drug-related suppression of aldosterone and cortisol, aminoglutethimide is no events (greater than 10% incidence) from the randomized phase longer recommended for treating metastatic breast cancer. Ami- III studies were injection-site reactions and hot flashes. Common noglutethimide is also occasionally used to try to reverse excess events (1% to 10% incidence) included asthenia, headache, and hormone production by adrenocortical cancers.120 gastrointestinal disturbances such as nausea, vomiting, and diar- Aromatase (cytochrome P-450 19 [CYP19]) is encoded by the rhea, with minor gastrointestinal disturbances being the most com- CYP19 gene, which is highly polymorphic. Some of these vari- monly described adverse event. ants are functionally important121 and may have clinical signifi- cance.122,123 Pharmacology Aromatase inhibitors have been classified in a number of dif- ferent ways, including first, second, and third generation; steroidal Fulvestrant is a steroidal molecule derived from E2 with an al- and nonsteroidal; and reversible (ionic binding) and irreversible kylsulphonyl side chain in the 7-α position (Fig. 27.2). Because (suicide inhibitor, covalent binding).124 The nonsteroidal aro- fulvestrant is poorly soluble and has low and unpredictable oral bio- matase inhibitors include aminoglutethimide (first generation), availability, a parenteral formulation of fulvestrant was developed rogletimide and fadrozole (second generation), and anastrozole, in an attempt to maximize delivery of the drug.111 The intramus- letrozole, and vorozole (third generation). The steroidal aromatase cular formulation provides prolonged release of the drug over sev- inhibitors include formestane (second generation) and exemes- eral weeks. The pharmacokinetics of three different single doses of tane (third generation). fulvestrant (50, 125, and 250 mg) have been published.111 In this Steroidal and nonsteroidal aromatase inhibitors differ in their phase I/II multicenter study, postmenopausal women with primary modes of interaction with, and their inactivation of, the aroma- breast cancer who were awaiting curative surgery received either tase enzyme. Steroidal inhibitors compete with the endogenous

OH

O–

S+ CF CF HO

Figure 27.2 Structure of fulvestrant.

tahir99 - UnitedVRG Chapter 27 Hormonal Agents 283 substrates, androstenedione and testosterone, for the active site N of the enzyme and are processed into intermediates that bind ir- C reversibly to the active site, causing irreversible enzyme inhibi- 19 N tion. Nonsteroidal inhibitors also compete with the endogenous N substrates for access to the active site, where they then form a re- N versible bond to the heme iron atom so that enzyme activity can re- cover if the inhibitor is removed; however, inhibition is sustained whenever the inhibitor is present.19

Letrozole and Anastrozole

Both letrozole and anastrozole have been extensively studied in the metastatic and adjuvant settings. When compared to tamoxi- C fen, both letrozole and anastrozole have demonstrated superior response rates and progression-free survival in the metastatic set- N ting.124,125 In the adjuvant setting, two trials have been performed and demonstrated superiority in terms of relapse-free survivals of Figure 27.3 Structure of letrozole. both anastrozole (ATAC)126 and letrozole (BIG 1-98).127 Addition- ally, anastrozole has been studied in a sequential approach, and the sequence of tamoxifen followed by anastrozole is superior studies, letrozole caused a significant decline in plasma E and to 5 years of tamoxifen alone.128 Anastrozole has recently been 1 E within 24 hours of a single oral dose of 0.1 mg.139,140 After 2 compared to placebo in women at an increased risk of develop- 2 weeks of treatment, the blood levels of E , E , and estrone sul- ing breast cancer and was demonstrated to significantly reduce the 2 1 fate were suppressed 95% or more from baseline. This continued incidence of invasive breast cancer.129 over the 12 weeks of therapy. There was no apparent alteration The side effects of both anastrozole and letrozole are similar in plasma levels of cortisol and aldosterone with letrozole or and include arthralgias and myalgias in up to 50% of patients. Both after corticotropin stimulation.139 In postmenopausal women letrozole and anastrozole are associated with a higher rate of bone with advanced breast cancer, the drug did not have any effect on fracture, compared with the tamoxifen.130 At the present time, follicle-stimulating hormone (FSH), luteinizing hormone (LH), minimal long-term (longer than 5 years) clinical data regarding thyrotropin (previously thyroid-stimulating hormone), cortisol, the effect of aromatase inhibitors on bones are available. When 17-α-hydroxyprogesterone, androstenedione, or aldosterone blood offering anastrozole for extended periods of time to patients with CANCER THERAPEUTICS concentrations.141,142 early breast cancer, attention to bone health is paramount, and Anastrozole is a nonsteroidal aromatase inhibitor that is 200- bone density should be monitored in all patients. Prospective stud- fold more potent than aminoglutethimide.143 No effect on the ies have demonstrated that bisphosphonates prevent aromatase- adrenal glands has been detected. In human studies, the t is inhibitor–induced bone loss and a meta-analysis presented at the max 2 to 3 hours after oral ingestion.144 Elimination is primarily via 2013 San Antonio Breast Cancer Symposium demonstrated that hepatic metabolism, with 85% excreted by that route and only bisphosphonates reduce bone recurrences and prolong overall 10% excreted unchanged in urine. The main circulating metabo- survival. Therefore, bisphosphonates should be considered in AI- lite is triazole after cleavage of the two rings in anastrozole by N- treated patients, both in those with and without an increased risk dealkylation. Linear pharmacokinetics have been observed in the of bone fractures. dose range of 1 to 20 mg and do not change with repeat dosing. A meta-analysis of toxicities comparing aromatase inhibitors The terminal half-life is approximately 50 hours, and steady-state with tamoxifen has demonstrated a 30% increase in grade 3 and 4 concentrations are achieved in approximately 10 days with once-a- cardiac events with aromatase inhibitors.131 However, prospective day dosing and are three to four times higher than peak concentra- data demonstrate no differences in myocardial events comparing tions after a single dose. Plasma protein binding is approximately anastrozole with placebo, although an increase in hypertension 40%.145 In one study, anastrozole 1 mg and 10 mg daily, inhib- was observed.129 ited in vivo aromatization by 96.7% and 98.1%, respectively, and No impact has been seen with anastrozole on adrenal steroido- plasma E and E levels were suppressed 86.5% and 83.5%, respec- genesis at up to 10 times the clinically recommended dose.132 1 2 tively, regardless of dose.146 Thus, 1 mg of anastrozole achieves Although letrozole may decrease basal and adrenocorticotropic near maximal aromatase inhibition and plasma estrogen suppres- hormone–stimulated cortisol synthesis,133,134 the clinical effect sion in breast cancer patients. appears to be minimal. Aromatase inhibitors appear to have dif- A recent prospective study to evaluate the pharmacokinetics ferential effects on lipids. In a study of over 900 patients with of anastrozole (1 mg per day) demonstrated large interindividual metastatic disease, anastrozole showed no marked effect on lipid variations in plasma anastrozole and anastrozole metabolite con- profiles compared with baseline.135 Conversely, the administration centrations, as well as pretreatment and postdrug plasma E , E , of letrozole in women with advanced breast cancer resulted in sig- 1 2 and E conjugate and estrogen precursor (androstenedione and nificant increases in total cholesterol and low-density lipoprotein, 1 testosterone) concentrations.147 Further research is needed to de- from baseline, after 8 and 16 weeks of therapy.136 In the Breast In- termine the basis for the wide variability in the pharmacokinetics ternational Group 1-98 trial, more women who received letrozole of anastrozole and whether these findings are clinically relevant. experienced grade 1 hypercholesterolemia compared to women who received tamoxifen.127 Letrozole is a nonsteroidal aromatase inhibitor with a high Exemestane specificity for the inhibition of estrogen production (Fig. 27.3). Letrozole is 180 times more potent than aminoglutethimide as Exemestane has a steroidal structure and is classified as a type 1 an inhibitor of aromatase in vitro. Aldosterone production in vitro aromatase inhibitor, also known as an aromatase inactivator, be- is inhibited by concentrations 10,000 times higher than those re- cause it irreversibly binds with and permanently inactivates the quired for inhibition of estrogen synthesis.137,138 In a normal male enzyme.134 Exemestane has been compared to tamoxifen in both volunteer study, letrozole was shown to decrease E2 and serum the metastatic and adjuvant settings. In the setting of tamoxifen- E1 levels to 10% of baseline with a single 3-mg dose. In phase I refractory metastatic breast cancer, exemestane is superior to 284 Cancer Therapeutics megestrol acetate, as demonstrated in a phase III trial in which im- concentrations and include hot flashes, sweating, and nausea.164 provements in both median time to tumor progression and median These symptoms can be reversed with low doses of progesterone survival were observed.148 In the adjuvant setting, the international analogs.5 In males treated with GnRH analogs for prostate cancer, exemestane study compared 2 to 3 years of tamoxifen with 2 to an alternate strategy of intermittent schedule of GnRH administra- 3 years of exemestane in women who had previously competed tion may result in improved tolerability and quality of life, with 2 to 3 years of adjuvant tamoxifen. In this trial, a switch to ex- comparable efficacy compared with continuous GnRH analog emestane resulted in superior disease-free and overall survival in administration in well-selected advanced prostate cancer patient the hormone receptor–positive subtype. Furthermore, exemestane cohorts.165 However, in a recent trial comparing intermittent with has been compared with the nonsteroidal agent anastrozole in the continuous androgen ablation in newly diagnosed metastatic hor- adjuvant treatment of ER-positive breast cancer, and there were no mone sensitive prostate cancer patients, a greater risk for death differences in disease-free or overall survival.149 Finally, exemes- from an intermittent strategy could not be conclusively ruled out tane has been compared to placebo in patients at increased risk of although intermittent therapy resulted in small improvements in breast cancer, and a significant reduction in the risk of developing quality of life.166 invasive breast cancer was observed.150 GnRH analogs available for clinical use include goserelin167,168 and leuprolide.169 Both are available in depot intramuscular preparations to be given at monthly intervals. The recommended Side Effects of Exemestane monthly dose of leuprolide is 7.5 mg and of goserelin is 3.6 mg. There are also longer acting depot preparations to be administered Although preclinical studies have suggested that exemestane pre- every 3, 4, 6, and 12 months. vented bone loss in ovariectomized rats,151 the Intergroup Exemes- tane adjuvant trial still demonstrated a higher rate of bone fracture for patients randomized to the exemestane arm and there were no Pharmacology differences in fracture rates comparing anastrozole with exemes- 149 tane. Side effects, including arthralgias and myalgias, appear Analogs of the decapeptide GnRH167,169,170 have been synthesized to be similar to the other AIs. With regard to steroidogenesis, no by modifications of position 6 in which the l- has been ex- impact on either cortisol or aldosterone levels was seen in a small 152 changed for a d-amino acid and the C-terminal amino acid has study after the administration of exemestane for 7 days. Finally, been either replaced by an ethylamide or substituted for a modified exemestane has weak androgenic properties, and its use at higher amino acid. These changes increase the affinity of the analog for doses has been associated with steroidal-like side effects, such as the GnRH receptor and decrease the susceptibility to enzymatic weight gain and acne.153,154 However, these side effects have not 155 degradation. There is an amino acid structure of GnRH with the been observed with the FDA-approved dose (25 mg per day). substitutions for leuprolide and goserelin. Initial administration of these compounds results in stimulation of gonadotropin release. However, prolonged administration has led to profound inhibition Pharmacology 170 of the pituitary–gonadal axis. Plasma E2 and progesterone are consistently suppressed to postmenopausal or castrate levels after Exemestane is administered once daily by mouth, with the rec- 164,171 ommended daily dose being 25 mg. The time needed to reach 2 to 4 weeks of treatment with goserelin or leuprolide. These 156 157 drugs are administered intramuscularly or subcutaneously in a maximal E2 suppression is 7 days, and its half-life is 27 hours. At daily doses of 10 to 25 mg, exemestane suppresses estrogen con- parenteral sustained-release microcapsule preparation, because centrations to 6% to 15% of pretreatment levels. This activity is parenteral administration of the parent drug is otherwise associ- more pronounced than that produced by formestane and compa- ated with rapid clearance. The GnRH analogs are metabolized rable to that produced by the nonsteroidal AIs, anastrozole and in the liver, kidney, hypothalamus, and pituitary gland by neutral letrozole.158–160 Exemestane does not appear to affect cortisol or peptidase cleavage of the peptide bond between the tyrosine in the aldosterone levels when evaluated after 7 days of treatment based 5 position and the amino acid in position 6 and by a postproline- on dose-ranging studies, including doses from 0.5 to 800 mg.152 cleaving enzyme that cleaves the peptide bond between proline Exemestane is metabolized by CYP3A4.134 Although drug–drug in the 9 position and the glycine-NH2 in the 10 position. Substitu- interactions have not been formally reported for exemestane, there tions at the glycine 6 position and modification of the C-terminal is the potential for interactions with drugs that affect CYP3A4.134 make these analogs more resistant to this enzymatic cleavage. Leuprolide is approximately 80 to 100 times more potent than endogenous GnRH. It induces castrate levels of testosterone in men with prostate cancer within 3 to 4 weeks of drug administra- GONADOTROPIN-RELEASING HORMONE tion after an initial sharp increase in LH and FSH. The mecha- ANALOGS nisms of action include pituitary desensitization after a reduction in pituitary GnRH receptor binding sites and possibly a direct an- Gonadotropin-releasing hormone (GnRH) analogs result in a titumor effect in ER-positive human breast cancer cells.169 The medical orchiectomy in men and are used as a means of provid- depot form results in a dose rate of 210 μg per day of leuprolide. ing androgen ablation for hormone-sensitive and castration re- Peak concentrations of the depot form, achieved approximately 3 fractory metastatic prostate cancer.161 Because the initial agonist hours after drug administration, have been reported to range be- activity of GnRH analogs can cause a tumor flare from temporarily tween 13.1 and 54.5 μg/L. There appears to be a linear increase increased androgen levels, concomitant use of the antiandrogen in the area under the curve (AUC) for doses of 3.75, 7.5, and 15.0 flutamide or bicalutamide has been used to prevent this effect. mg in the depot form. The parenteral bioavailability of subcuta- GnRH analogs can also cause tumor regressions in hormonally re- neously injected leuprolide is 94%. The volume of distribution sponsive breast cancers162 and have received FDA approval for the ranges from 27.4 to 37.1 L. In human studies, leuprolide urinary treatment of metastatic breast cancer in premenopausal women. excretion as a metabolite was the primary route of clearance. Data suggest that these drugs may be useful as adjuvant therapy Goserelin is approximately 100 times more potent than the natu- of premenopausal women with resected breast cancer.163 The use rally occurring GnRH. Like leuprolide, it causes the stimulation of of these drugs in combination with tamoxifen or exemestane in LH and FSH acutely, and with subsequent administration, GnRH premenopausal women with primary breast cancer is the subject receptor numbers decrease, and the pituitary becomes desensitized of large, ongoing, international clinical trials. The primary toxici- with decreasing LH and FSH levels. Castrate levels of testosterone ties of GnRH analogs are secondary to the ablation of sex steroid are achieved within 1 month. In women, goserelin inhibits ovarian

tahir99 - UnitedVRG Chapter 27 Hormonal Agents 285 androgen production, but serum levels of dehydroepiandrosterone administration, or when the metastatic prostate cancer is unre- sulfate and, to a lesser extent, androstenedione, are preserved. In sponsive, despite androgen ablation therapy. The recommended vitro, goserelin has demonstrated antitumor activity in estrogen-de- dose is 250 mg by mouth three times a day. In patients whose pros- pendent MCF7 human breast cancer cells and LNCaP2 prostate tate cancer is growing despite flutamide use, stopping flutamide cancer cells. The drug is released at a continuous mean rate of can sometimes cause a flutamide-withdrawal response. 120 μg per day in the depot form, with peak concentrations in the The most common toxicity seen with flutamide is diarrhea, range of 2 to 3 μg/L achieved. The mean volume of distribution in with or without abdominal discomfort. Gynecomastia, which can six patients has been reported to be 13.7 L,172 which is consistent be tender, frequently occurs in men who are not receiving con- with extracellular fluid volume. Goserelin is principally excreted comitant androgen ablation therapy.178 Flutamide can rarely cause in the urine, with a mean total body clearance of 8 L per hour hepatotoxicity, a condition that is reversible if detected early, but in patients with normal renal function. The total body clearance this toxicity can also be fatal.179 There is no accepted, clinically is reduced by approximately 75%, with renal dysfunction and the recommended testing schedule to screen for flutamide-induced elimination half-life increased two- or threefold. However, dose ad- hepatotoxicity other than being aware of this phenomenon and justment for renal insufficiency does not appear to be necessary. testing for liver function if hepatic symptoms develop. The 5 to 10 hexapeptide and the 4 to 10 hexapeptide were detected in urine in animal studies.173 The terminal half-life of goserelin is approximately 5 hours after subcutaneous injection. Protein bind- Pharmacology ing is low, and no known drug interactions have been documented. Flutamide is a pure antiandrogen with no intrinsic steroidal activ- ity.180 Flutamide’s mechanism of action is as an androgen-receptor GONADOTROPIN-RELEASING HORMONE antagonist. This binding prevents dihydrotestosterone binding and ANTAGONISTS subsequent translocation of the androgen-receptor complex into the nuclei of cells. Because it is a pure antiandrogen, it acts only at the cellular level. The administration of flutamide alone leads to Modification to the structure of GnRH has resulted in the develop- increased LH and FSH production and a concomitant increase in ment of GnRH antagonist compounds that are currently being used plasma testosterone and E levels. Plasma protein binding ranges in the treatment of prostate cancer. Abarelix was initially approved 2 between 94% and 96% for flutamide and between 92% and 94% by the FDA in 2003 as the first depot-injectable GnRH antagonist, for 2-hydroxyflutamide, its major metabolite. When the drug is but was subsequently withdrawn in 2005. Degarelix is a syntheti- administered three times a day, steady state levels are achieved by cally modified compound with GnRH antagonist activity that was day 6. The elimination half-life at steady state is 7.8 hours, and approved for use by the FDA in 2008 for the management of pros- 2-hydroxyflutamide achieves concentrations 50 times higher than tate cancer.174 Its effect in prostate cancer treatment is to block the

the parent drug at steady state and has equal or greater potency CANCER THERAPEUTICS GnRH receptor, and thereby prevent the trigger for the production than that of flutamide.180 The elimination half-life for the metabo- of LH, which mediates androgen synthesis. In contrast to GnRH lite is 9.6 hours. The high plasma concentrations of 2-hydroxyflu- analogs, degarelix does not cause tumor flare symptoms secondary tamide, as compared with flutamide, suggest that the therapeutic to temporary increased androgen production. A large randomized benefits of flutamide are mediated primarily through its active clinical trial demonstrated that degarelix was associated with a rapid metabolite.181 and sustained reduction in serum testosterone, prostate-specific an- tigen (PSA), FSH, and LH levels, with a loading dose of 240 mg sub- cutaneously, followed by a monthly maintenance dose of 80 mg175 Bicalutamide with comparable efficacy to leuprolide.176 The most common side effects (greater than 10%) were hot flashes and pain at the injection Bicalutamide is another nonsteroidal antiandrogen that has been 176 site when patients were provided degarelix for a 12-month period. approved by the FDA for use in the United States. The recom- It is unknown if degarelix will have a similar chronic side effect mended dose is one 50-mg tablet per day. One randomized trial profile known to be associated with long-term GnRH analog use. reported that bicalutamide compared favorably with flutamide in patients with advanced prostate cancer.182 Bicalutamide appears to Pharmacology be relatively well tolerated and is associated with a lower incidence of diarrhea than is flutamide. The recommended loading dose of degarelix is 240 mg, adminis- tered as two injections of 120 mg each subcutaneously. Monthly Pharmacology maintenance doses of 80 mg as a 20 mg/mL solution is started 28 days after the loading dose. In an analysis of pharmacokinetic/ Bicalutamide has a binding affinity to the androgen receptor in pharmacodynamic (PK/PD) properties of degarelix in 60 healthy the rat prostate that is four times greater than that of 2-hydroxyflu- males, after a single subcutaneous dose, a terminal half life of 183,184 177 tamide. In vivo, bicalutamide caused a marked inhibition of 47 days was observed. PK properties of degarelix have been eval- growth of accessory sex organs in rats, with a potency 5 to 10 times uated when administered as a subcutaneous depot of drug as a gel greater than that of flutamide. Unlike flutamide, bicalutamide did in six different doses to 48 healthy males and when administered not cause a significant increase in LH or testosterone in rats. In hu- intravenously. Using data from several clinical trials, the rate of mans, the drug has a long plasma half-life of 5 to 7 days, so it may be drug diffusion from subcutaneous administration results in detect- administered on a weekly schedule. Pharmacokinetics of the drug able drug up to 60 days after a single dose compared to less than 4 showed a dose-dependent increase in mean peak plasma concentra- days when the drug is injected intravenously. tions, and the AUC increased linearly with the dose. The half-life of bicalutamide in humans was approximately 6 days, and the drug ANTIANDROGENS clearance was not saturable at plasma concentrations up to 1,000 ng/ mL. Daily dosing of the drug led to an approximately tenfold accu- mulation after 12 weeks of administration. In contrast to results in Flutamide rats, serum concentrations of testosterone and LH increased signifi- cantly from baseline at all dose levels tested in humans. Whereas The antiandrogen flutamide is used in men with metastatic pros- serum FSH concentrations remained essentially unchanged, the 185 tate cancer either as initial therapy, combined with GnRH analog median serum E2 concentrations increased significantly. 286 Cancer Therapeutics

Nilutamide binding to androgen response elements and the recruitment of coactivators.187 In early clinical trials, promising results have Nilutamide represents the third variation of an antiandrogen avail- been observed in castrate refractory and chemotherapy-resistant able for use in patients with prostate cancer. The observation of settings. The major metabolite of enzalutamide is N-desmethyl unique toxicities, night blindness, and pulmonary toxicity has lim- enzalutamide, and CYP2C8 is responsible for the formation of ited its use. the active metabolite, N-desmethyl enzalutamide. Enzalutamide pharmacokinetics, in the studied dose range between 30 mg to 480 mg, exhibited a linear, two-compartmental model with first- NOVEL ANTIANDROGENS order kinetics. In patients with mCRPC, the mean (% coefficient of variation [CV]) predose Cmin values for enzalutamide and Although testosterone depletion remains an unchallenged stan- N-desmethyl enzalutamide were 11.4 (25.9%) μg/mL and 13.0 dard for advanced stage hormone-sensitive disease, evidence has (29.9%) μg/mL, respectively. Enzalutamide is mainly metabo- emerged that castration-recurrent prostate cancer remains andro- lized by CYP2C8 and CYP3A4. Doses ranging from 30 to 600 gen receptor (AR) dependent and is neither hormone refractory nor mg daily have been evaluated, with dose-limiting toxicities includ- androgen independent, which were commonly used terms to de- ing fatigue, seizure, asthenia, anemia, and arthralgia occurring at fine the progression of advanced stage disease following androgen higher dose levels. At present, enzalutamide has been approved deprivation therapy. Recognition of AR functioning despite the for treating advanced castrate-recurrent prostate cancer188 after a paucity of circulating androgens is evidenced by the elevation of failure of docetaxel chemotherapy at a dose of 160 mg (four, 40 AR messenger RNA in castration-recurrent tumor tissue relative to mg oral capsules). Clinical trials are ongoing to evaluate the ef- androgen-dependent tumors and reexpression of some androgen- ficacy of enzalutamide in castrate-recurrent patients who are che- regulated genes during clinical castration resistance. Recently, the motherapy naïve. AR axis has been the focus of therapeutic targeting. Galeterone and Orteronel Abiraterone Acetate Novel CYP17 inhibitors that are more selective for 17,20- After the failure of initial androgen manipulation with GnRH over 17 α-hydroxylase are currently being developed. Orteronel analogs and peripheral antiandrogens, prostate cancer continues (TAK-700) is an example of a highly selective 17,20 lyase, which to respond to a variety of second- and third-line hormonal inter- is currently undergoing phase III clinical trials in a pre- and post- ventions. Based on this observation, CYP17, a key enzyme in an- chemotherapy castrate-recurrent setting after the failure of andro- drogen and estrogen synthesis, was targeted using ketoconazole, gen-deprivation therapy.189 Other novel agents being developed which is a weak, reversible, and nonspecific inhibitor of CYP17 include galeterone, which is an inhibitor of CYP 17 α-hydroxylase resulting in modest antitumor activity of short durability. More and C17,20 lyase. Survival mechanisms of prostate cancer cells recently, abiraterone, a more potent (i.e., 20 times more than ke- targeted by galeterone include its binding to AR, competitive in- toconazole), selective, and irreversible inhibitor of CYP17, has hibition of testosterone binding, and a reduction in the quantity been investigated in castration-recurrent prostate cancer, and sig- of AR protein within the prostate cancer cells. It can also enhance nificant objective responses have been observed.186 Chemically, it the degradation of constitutively active splice variants. Therefore, is a 3-pyridyl steroid pregnenolone–derived compound available taken together, it diminishes the ability of the cells to respond in an oral prodrug form of abiraterone acetate. Its main toxicity to the low levels of androgenic growth signals. This agent is cur- is from symptoms of mineralocorticoid excess (including hypo- rently in early clinical safety and efficacy testing for advanced stage kalemia, hypertension, and fluid overload), because continuous prostate cancer. CYP17 blockade results in raising adrenocorticotrophic hormone (ACTH) levels that increase steroid levels upstream of CYP17, including corticosterone and deoxycorticosterone. These adverse OTHER SEX STEROID THERAPIES effects are best avoided by the coadministration of steroids. The established dose of abiraterone is 1,000 mg a day (four 250 mg tablets). Following oral administration of abiraterone acetate, Fluoxymesterone the median time to maximum plasma abiraterone concentrations is 2 hours. At the dose of 1,000 mg daily, steady state values (mean Fluoxymesterone is an androgen that has been used in women ± standard deviation [SD]) of Cmax were 226 ± 178 ng/mL and with metastatic breast cancer who have hormonally responsive of AUC were 1173 ± 690 ng.hr/mL. Abiraterone is highly bound cancers and who have progressed on other hormonal therapies (>99%) to the human plasma proteins, albumin and alpha-1 acid such as tamoxifen, an aromatase inhibitor, or megestrol acetate. glycoprotein. The apparent steady state volume of distribution The usual dose is 10 mg given twice daily. Although the overall (mean ± SD) is 19,669 ± 13,358 L. No major deviation from response rate is low for fluoxymesterone used in this clinical situ- 190 dose proportionality was observed in the dose range of 250 mg to ation, there are some patients who have substantial antitumor 1,000 mg. However, the exposure was not significantly increased responses lasting for months or even years. when the dose was doubled from 1,000 to 2,000 mg (8% increase in Toxicities associated with fluoxymesterone are those that would mean AUC). The two main circulating metabolites of abiraterone be expected with an androgen: hirsutism, male-pattern baldness, in human plasma are abiraterone sulfate (inactive) and N-oxide voice lowering (hoarseness), acne, enhanced libido, and erythro- abiraterone sulfate (inactive), which each account for about 43% cytosis. Fluoxymesterone can also cause elevated liver function of exposure. CYP3A4 and SULT2A1 are enzymes involved in the test results in some patients and, rarely, has been associated with formation and conjugation of N-oxide abiraterone. hepatic neoplasms.

Enzalutamide Pharmacology

Enzalutamide is a new diarylthiohydantoin compound that binds Fluoxymesterone is a chlorinated synthetic analog of testosterone AR with an affinity that is several-fold greater than the antiandro- with potent androgenic and anabolic activity in humans. Limited gens bicalutamide and flutamide. This class of novel AR inhibitor pharmacologic information is available on this agent. Colburn,191 also disrupts the nuclear translocation of AR and impairs DNA using a radioimmunoassay, studied two patients after a single oral

tahir99 - UnitedVRG Chapter 27 Hormonal Agents 287 administration of a 50-mg dose. Peak serum concentrations were Furthermore, if patients who receive megestrol have a significant achieved between 1 and 3 hours after administration, with the infection, experience trauma, or undergo surgery, then cortico- average peak concentrations being 335 ng/mL. By 5 hours after steroid coverage should be administered. There appears to be drug administration, serum levels had declined to approximately a slightly increased incidence of thromboembolic phenomena 50% of the peak concentration. Urinary excretion of a 10-mg in patients receiving megestrol alone.202 This risk appears to be dose can be detected for 24 hours, and at least 6-hydroxy, 4-ene, higher if megestrol is administered with concomitant cytotoxic 3-β, and 11-hydroxy metabolites of fluoxymesterone have been therapy.205 There are conflicting reports regarding megestrol- detected.192 causing edema.206 If it does, the edema is generally minimal and easily handled with a mild diuretic. Megestrol may cause 207 Estrogens: Diethylstilbestrol and Estradiol impotence in some men. The incidence of this is controversial, although it is generally agreed that this is a reversible situation. Megestrol can cause menstrual irregularities, the most prominent Diethylstilbestrol (DES) had been the primary hormonal ther- of which is withdrawal menstrual bleeding within a few weeks of apy for postmenopausal metastatic breast cancer. Randomized drug discontinuation.5 Although nausea and vomiting have some- comparative trials demonstrated it had a similar response rate to times been attributed as a toxicity of this drug, there are data to that of tamoxifen.193,194 However, based on these trials, DES use demonstrate that this drug has antiemetic properties.200,201,205 In was supplanted by tamoxifen, primarily because DES has more terms of magnitude, megestrol appears to decrease both nausea toxicity. DES is occasionally used in metastatic breast cancer pa- and vomiting in advanced-stage cancer patients by approximately tients who have hormonally sensitive cancers that have failed to two thirds. respond to multiple other hormonal therapies. The usual dose in Medroxyprogesterone has many of the same properties, clinical this situation is 15 mg per day, either as a single dose or as divided uses, and toxicities as megestrol acetate. It has never been com- doses. DES was also used as androgen ablation therapy in men monly used in the United States for the treatment of breast cancer with metastatic prostate cancer.195 Doses of approximately 3 mg but has been used more in Europe. Medroxyprogesterone is avail- per day result in testosterone levels that are seen in an anorchid able in 2.5- and 10-mg tablets and in injectable formulations of state. 100 and 400 mg/L. Dosing for the treatment of metastatic breast DES toxicities include nausea and vomiting, breast tenderness, or prostate cancer has commonly been 400 mg per week or more and a darkening of the nipple–areolar complex. DES increases and 1,000 mg per week or more for metastatic endometrial cancer. the risk of thromboembolic phenomenon, which may result in Injectable or daily oral doses have been used for controlling hot life-threatening complications. Although DES is not clinically flashes. available in the United States, similar antitumor effects and toxici- ties are seen with estradiol, with a target dose of 10 mg by mouth three times a day. The pharmacology of E has been extensively 2 Pharmacology CANCER THERAPEUTICS described elsewhere.196 The exact mechanism of antitumor effect of medroxyprogester- Medroxyprogesterone and Megestrol one and megestrol is unclear. These drugs have been reported to suppress adrenal steroid synthesis,208 suppress ER levels,209 Medroxyprogesterone and megestrol are 17-OH-progesterone alter tumor hormone metabolism,210 enhance steroid metabo- derivatives differing in a double bond between C6 and C7 lism,211 and directly kill tumor cells.212 In addition, progestins positions in megestrol. Historically, megestrol was used as a hor- may influence some growth factors,213 suppress plasma estrone monal agent for patients with advanced breast cancer, usually sulfate formation, and, at high concentrations, inhibit P-glyco- at a total daily dose of 160 mg. Additionally, it is still used for protein. the treatment of hormonally responsive metastatic endometrial The oral bioavailability of these progestational agents is un- cancer, at a dose of 320 mg per day. In addition, doses of 160 mg known, although absorption appears to be poor for medroxypro- per day are occasionally used as a hormonal therapy for prostate gesterone relative to megestrol. cancer.197 Megestrol has also been extensively evaluated for the The terminal half-life for megestrol is approximately 14 198–201 214,215 216 treatment of anorexia/cachexia related to cancer or AIDS. hours, with a tmax of 2 to 5 hours after oral ingestion. The Various dosages ranging from 160 to 1,600 mg per day have been AUC for a single megestrol dose of 160 mg is between 2.5- and used. A prospective study has demonstrated a dose–response re- 8-fold higher than that for single-dose medroxyprogesterone at lationship with doses up to 800 mg per day.202 Low dosages of 1,000 mg with a radioactive dose of megestrol; 50% to 78% is megestrol (20 to 40 mg per day) have been shown to be an effec- found in the urine after oral administration, and 8% to 30% is tive means of reducing hot flashes in women with breast cancer found in the feces. and in men who have undergone androgen ablation therapy.5 Al- Metabolism and excretion of medroxyprogesterone have been though megestrol had historically been commonly administered incompletely characterized. In humans, 20% to 50% of a [3H]me- four times per day, the long terminal half-life supports once-per- droxyprogesterone dose is excreted in the urine and 5% to 10% day dosing. in the stool after intravenous administration.217–219 Metabolism Megestrol is a relatively well-tolerated medication, with its of medroxyprogesterone occurs via hydroxylation, reduction, de- most prominent side effects being appetite stimulation and re- methylation, and combinations of these reactions.220 The major sultant weight gain. Although these may be beneficial effects in urinary metabolite is a glucuronide. Less than 3% of the dose patients with anorexia/cachexia, they can be important problems is excreted as unconjugated medroxyprogesterone in humans. in patients with breast or endometrial cancers. Another side ef- Clearance of medroxyprogesterone has been reported to range be- fect of megestrol acetate is the marked suppression of adrenal tween 27 and 70 L per hour.219 The initial volume of distribution steroid production by suppression of the pituitary–adrenal is between 4 and 8 L in humans. The mean terminal half-life is 203 axis. Although this appears to be asymptomatic in the major- 60 hours. The tmax for medroxyprogesterone occurs 2 to 5 hours ity of patients, reports suggest that this adrenal suppression can after oral administration. Medroxyprogesterone appears to be con- cause clinical problems in some patients.204 This drug has been centrated in the small intestine, the colon, and in adipose tissue in abruptly stopped for decades without the recognition of untoward human autopsy studies.221 Drug interactions of medroxyprogester- sequelae in patients, and it seems reasonable to continue this prac- one have been reported with aminoglutethimide, which decreases tice. Nonetheless, if Addisonian signs or symptoms develop after plasma medroxyprogesterone levels.222 Medroxyprogesterone may drug discontinuation, corticosteroids should be administered. reduce the concentration of the N-desmethyltamoxifen metabo- 288 Cancer Therapeutics lite concentration. Progestational agents also may increase plasma Pharmacology warfarin levels.223 These reports are consistent with CYP3A being the site of interaction. Octreotide is an 8-amino acid synthetic analog of the 14-amino acid peptide somatostatin.227 Octreotide has a similar high affinity for somatostatin receptors, as does its parent compound, with a OTHER HORMONAL THERAPIES concentration that inhibits the receptor by 50% in the subnano- molar range. Octreotide inhibits insulin, glucagon, pancreatic polypeptide, gastric inhibitory polypeptide, and gastrin secretion. Octreotide It has a much longer duration of action than the parent compound because of its greater resistance to enzymatic degradation. Its ab- Octreotide is a somatostatin analog that is administered for the sorption after subcutaneous administration is rapid, and bioavail- treatment of carcinoid syndrome and other hormonal excess syn- ability is 100% after subcutaneous injection. Peak concentrations dromes associated with some pancreatic islet cell cancers and of 4 μg/L after a 100-μg dose occur within 20 to 30 minutes of acromegaly. Response rates (measured in terms of a reduction in subcutaneous injection and are 20% to 40% of the correspond- diarrhea and flushing) are high and can last for several months ing intravenous injection. Both peak concentration and AUC for to years. Occasionally, antitumor responses temporarily related to octreotide increase linearly with dose. The total body clearance octreotide are seen with these tumors. Octreotide may be useful to in healthy volunteers is 9.6 L per hour. Hepatic metabolism of alleviate 5-fluorouracil–associated diarrhea.224–226 octreotide accounts for 30% to 40% of the drug’s disposition, and Octreotide can be administered intravenously or subcutane- 11% to 20% is excreted unchanged in the urine. The volume of ously. Initial doses of 50 μg are given two to three times on the distribution ranges between 18 and 30 L, and the terminal half- first day. The dose is titrated upward, with a usual daily dose of life is reported to be between 72 and 98 minutes. Sixty-five per- 300 to 450 μg per day for most patients. A depot preparation is cent of the drug is protein bound primarily to the lipoprotein available, allowing doses to be administered at monthly intervals. fraction.227,228 Because of the short half-life, classic octreotide is Octreotide is generally well tolerated overall. It appears to cause administered subcutaneously two or three times per day.229 A slow- more toxicity in acromegalic patients, with such problems as bra- release form of octreotide, designed for once-per-month adminis- dycardia, diarrhea, hypoglycemia, hyperglycemia, hypothyroidism, tration, controls the symptoms of carcinoid syndrome at least as and cholelithiasis. well as three-times-per-day octreotide.230

SELECTED REFERENCES

The full reference list can be accessed at lwwhealthlibrary.com/oncology. 34. Toy W, Shen Y, Won H, et al. ESR1 ligand-binding domain mutations in hormone-resistant breast cancer. Nat Genet 2013;45:1439–1445. 1. Fisher B, Costantino JP, Wickerham DL, et al. Tamoxifen for the prevention 35. Robinson DR, Wu YM, Vats P, et al. Activating ESR1 mutations in hormone- of breast cancer: current status of the National Surgical Adjuvant Breast and resistant metastatic breast cancer. Nat Genet 2013;45:1446–1451. Bowel Project P-1 study. J Natl Cancer Inst 2005;97:1652–1662. 36. Merenbakh-Lamin K, Ben-Baruch N, Yeheskel A, et al. D538G mutation 2. Fisher B, Dignam J, Wolmark N, et al. Tamoxifen in treatment of intraductal in estrogen receptor-alpha: a novel mechanism for acquired endocrine resis- breast cancer: National Surgical Adjuvant Breast and Bowel Project B-24 ran- tance in breast cancer. Cancer Res 2013;73:6856–6864. domised controlled trial. Lancet 1999;353:1993–2000. 37. Fendl KC, Zimniski SJ. Role of tamoxifen in the induction of hormone- 4. Davies C, Pan H, Godwin J, et al. Long-term effects of continuing adjuvant independent rat mammary tumors. Cancer Res 1992;52:235–237. tamoxifen to 10 years versus stopping at 5 years after diagnosis of oestrogen 41. Mani C, Kupfer D. Cytochrome P-450-mediated activation and irreversible receptor-positive breast cancer: ATLAS, a randomised trial. Lancet 2013;381: binding of the antiestrogen tamoxifen to proteins in rat and human liver: pos- 805–816. sible involvement of flavin-containing monooxygenases in tamoxifen activa- 6. Loprinzi CL, Kugler JW, Sloan JA, et al. Venlafaxine in management of hot tion. Cancer Res 1991;51:6052–6058. flashes in survivors of breast cancer: a randomised controlled trial. Lancet 49. Paik S, Shak S, Tang G, et al. A multigene assay to predict recurrence of 2000;356:2059–2063. tamoxifen-treated, node-negative breast cancer. N Engl J Med 2004;351: 10. Stearns V, Beebe KL, Iyengar M, et al. Paroxetine controlled release in the 2817–2826. treatment of menopausal hot flashes: a randomized controlled trial. JAMA 52. Wu X, Hawse JR, Subramaniam M, et al. The tamoxifen metabolite, endoxi- 2003;289:2827–2834. fen, is a potent antiestrogen that targets estrogen receptor alpha for degrada- 11. Pandya KJ, Morrow GR, Roscoe JA, et al. Gabapentin for hot flashes in 420 tion in breast cancer cells. Cancer Res 2009;69:1722–1727. women with breast cancer: a randomised double-blind placebo-controlled 54. Desta Z, Ward BA, Soukhova NV, et al. Comprehensive evaluation of tamox- trial. Lancet 2005;366:818–824. ifen sequential biotransformation by the human cytochrome P450 system 12. Loprinzi CL, Qin R, Balcueva EP, et al. Phase III, randomized, double- in vitro: prominent roles for CYP3A and CYP2D6. J Pharmacol Exp Ther blind, placebo-controlled evaluation of pregabalin for alleviating hot flashes, 2004;310:1062–1075. N07C1. J Clin Oncol 2010;28:641–647. 55. Stearns V, Johnson MD, Rae JM, et al. Active tamoxifen metabolite plasma 13. Madlensky L, Natarajan L, Tchu S, et al. Tamoxifen metabolite concentra- concentrations after coadministration of tamoxifen and the selective sero- tions, CYP2D6 genotype, and breast cancer outcomes. Clin Pharmacol Ther tonin reuptake inhibitor paroxetine. J Natl Cancer Inst 2003;95:1758–1764. 2011;89:718–725. 56. Goetz MP, Rae JM, Suman VJ, et al. Pharmacogenetics of tamoxifen 14. Tamoxifen for early breast cancer: an overview of the randomised trials. Early biotransformation is associated with clinical outcomes of efficacy and hot Breast Cancer Trialists’ Collaborative Group. Lancet 1998;351:1451–1467. flashes.J Clin Oncol 2005;23:9312–9318. 15. Wickerham DL, Fisher B, Wolmark N, et al. Association of tamoxifen and 58. Schroth W, Goetz MP, Hamann U, et al. Association between CYP2D6 uterine sarcoma. J Clin Oncol 2002;20:2758–2760. polymorphisms and outcomes among women with early stage breast cancer 16. Dewar JA, Horobin JM, Preece PE, et al. Long term effects of tamoxifen on treated with tamoxifen. JAMA 2009;302:1429–1436. blood lipid values in breast cancer. BMJ 1992;305:225–226. 59. Rae JM, Drury S, Hayes DF, et al. CYP2D6 and UGT2B7 genotype and risk 17. Love RR, Mazess RB, Barden HS, et al. Effects of tamoxifen on bone mineral of recurrence in tamoxifen-treated breast cancer patients. J Natl Cancer Inst density in postmenopausal women with breast cancer. N Engl J Med 1992;326: 2012;104:452–460. 852–856. 60. Regan MM, Leyland-Jones B, Bouzyk M, et al. CYP2D6 genotype and tamox- 29. Pietras RJ, Arboleda J, Reese DM, et al. HER-2 tyrosine kinase pathway tar- ifen response in postmenopausal women with endocrine-responsive breast gets estrogen receptor and promotes hormone-independent growth in hu- cancer: the breast international group 1-98 trial. J Natl Cancer Inst 2012;104: man breast cancer cells. Oncogene 1995;10:2435–2446. 441–451. 30. Osborne CK, Bardou V, Hopp TA, et al. Role of the estrogen receptor coacti- 61. Goetz MP, Suman VJ, Hoskin TL, et al. CYP2D6 metabolism and patient vator AIB1 (SRC-3) and HER-2/neu in tamoxifen resistance in breast cancer. outcome in the Austrian Breast and Colorectal Cancer Study Group trial J Natl Cancer Inst 2003;95:353–361. (ABCSG) 8. Clin Cancer Res 2013;19:500–507. 33. Zhang QX, Borg A, Wolf DM, et al. An estrogen receptor mutant with strong 62. Province MA, Goetz MP, Brauch H, et al. CYP2D6 Genotype and adjuvant hormone-independent activity from a metastatic breast cancer. Cancer Res tamoxifen: meta-analysis of heterogeneous study populations. Clin Pharma- 1997;57:1244–1249. col Ther 2014;95:216–227.

tahir99 - UnitedVRG Chapter 27 Hormonal Agents 289

65. Kelly CM, Juurlink DN, Gomes T, et al. Selective serotonin reuptake inhibi- 129. Cuzick J, Sestak I, Forbes JF, et al. Anastrozole for prevention of breast can- tors and breast cancer mortality in women receiving tamoxifen: a population cer in high-risk postmenopausal women (IBIS-II): an international, double- based cohort study. BMJ 2010;340:c693. blind, randomised placebo-controlled trial. Lancet 2014;383:1041–1048. 67. Irvin WJ Jr., Walko CM, Weck KE, et al. Genotype-guided tamoxifen 130. Baum M, Budzar AU, Cuzick J, et al. Anastrozole alone or in combination dosing increases active metabolite exposure in women with reduced with tamoxifen versus tamoxifen alone for adjuvant treatment of postmeno- CYP2D6 metabolism: a multicenter study. J Clin Oncol 2011;29: pausal women with early breast cancer: first results of the ATAC randomised 3232–3239. trial. Lancet 2002;359:2131–2139. 69. Goetz MP, Suman VA, Reid JR, et al. A first-in-human phase I study of the 131. Amir E, Seruga B, Nira S, et al. Toxicity of adjuvant endocrine therapy in tamoxifen (TAM) metabolite, Z-endoxifen hydrochloride (Z-Endx) in wom- postmenopausal breast cancer patients: a systematic review and meta-analy- en with aromatase inhibitor (AI) refractory metastatic breast cancer (MBC) sis. J Natl Cancer Inst 2011;103:1299–1309. (NCT01327781). Cancer Res 2013;73(24 Suppl): Abstract nr PD3-4. . 147. Ingle JN, Buzdar AU, Schaid DJ, et al. Variation in anastrozole metabolism 70. Lien EA, Anker G, Lonning PE, et al. Decreased serum concentrations of and pharmacodynamics in women with early breast cancer. Cancer Res tamoxifen and its metabolites induced by aminoglutethimide. Cancer Res 2010;70:3278–3286. 1990;50:5851–5857. 149. Goss PE, Ingle JN, Pritchard KI, et al. Exemestane versus anastrozole in 74. Lien EA, Solheim E, Lea OA, et al. Distribution of 4-hydroxy-N-desmethyl- postmenopausal women with early breast cancer: NCIC CTG MA.27—a tamoxifen and other tamoxifen metabolites in human biological fluids dur- randomized controlled phase III trial. J Clin Oncol 2013;31:1398–1404. ing tamoxifen treatment. Cancer Res 1989;49:2175–2183. 150. Goss PE, Ingle JN, Ales-Martinez JE, et al. Exemestane for breast-cancer 110. Howell A, Osborne CK, Morris C, et al. ICI 182,780 (Faslodex): develop- prevention in postmenopausal women. N Engl J Med 2011;364:2381–2391. ment of a novel, “pure” antiestrogen. Cancer 2000;89:817–825. 151. Goss PE, Grynpas M, Qi S, et al. The effects of exemestane on bone and 116. Osborne CK, Pippen J, Jones SE, et al. Double-blind, randomized trial lipids in the ovariectomized rat. Breast Cancer Res Treat 2001;69:224. comparing the efficacy and tolerability of fulvestrant versus anastrozole in 163. Kaufmann M, Jonat W, Blamey R, et al. Survival analyses from the ZEBRA postmenopausal women with advanced breast cancer progressing on prior study. Goserelin (Zoladex) versus CMF in premenopausal women with endocrine therapy: results of a North American trial. J Clin Oncol 2002;20: node-positive breast cancer. Eur J Cancer 2003;39:1711–1717. 3386–3395. 166. Hussain M, Tangen CM, Berry DL, et al. Intermittent versus continuous 117. Howell A, Robertson JF, Abram P, et al. Comparison of fulvestrant versus androgen deprivation in prostate cancer. N Engl J Med 2013;368:1314–1325. tamoxifen for the treatment of advanced breast cancer in postmenopausal 168. Vogelzang NJ, Chodak GW, Soloway MS, et al. Goserelin versus orchiec- women previously untreated with endocrine therapy: a multinational, dou- tomy in the treatment of advanced prostate cancer: final results of a random- ble-blind, randomized trial. J Clin Oncol 2004;22:1605–1613. ized trial. Zoladex Prostate Study Group. Urology 1995;46:220–226. 118. Leo AD, Jerusalem G, Petruzelka L, et al. Final overall survival: fulvestrant 175. Van Poppel H, Tombal B, de la Rosette JJ, et al. Degarelix: a novel gonad- 500 mg vs 250 mg in the randomized CONFIRM trial. J Natl Cancer Inst otropin-releasing hormone (GnRH) receptor blocker—results from a 1-yr, 2014;106:djt337. multicentre, randomised, phase 2 dosage-finding study in the treatment of 121. Ma CX, Adjei AA, Salavaggione OE, et al. Human aromatase: gene rese- prostate cancer. Eur Urol 2008;54:805–813. quencing and functional genomics. Cancer Res 2005;65:11071–11082. 176. Klotz L, Boccon-Gibod L, Shore ND, et al. The efficacy and safety of degare- 124. Goss PE, Ingle JN, Martino S, et al. A randomized trial of letrozole in post- lix: a 12-month, comparative, randomized, open-label, parallel-group phase III menopausal women after five years of tamoxifen therapy for early-stage breast study in patients with prostate cancer. BJU Int 2008;102:1531–1538. cancer. N Engl J Med 2003;349:1793–1802. 186. Attard G, Reid AH, A’Hern R, et al. Selective inhibition of CYP17 with abi- 126. Howell A, Cuzick J, Baum M, et al. Results of the ATAC (Arimidex, Tamoxi- raterone acetate is highly active in the treatment of castration-resistant pros- fen, Alone or in Combination) trial after completion of 5 years’ adjuvant tate cancer. J Clin Oncol 2009;27:3742–3748. treatment for breast cancer. Lancet 2005;365:60–62. 188. Scher HI, Fizazi K, Saad F, et al. Increased survival with enzalutamide in CANCER THERAPEUTICS 127. Thurlimann B, Keshaviah A, Coates AS, et al. A comparison of letrozole and prostate cancer after chemotherapy. N Engl J Med 2012;367:1187–1197. tamoxifen in postmenopausal women with early breast cancer. N Engl J Med 224. Cascinu S, Fedeli A, Fedeli SL, et al. Control of chemotherapy-induced 2005;353:2747–2757. diarrhoea with octreotide in patients receiving 5-fluorouracil. Eur J Cancer 128. Jakesz R, Jonat W, Gnant M, et al. Switching of postmenopausal women with 1992;28:482–483. endocrine-responsive early breast cancer to anastrozole after 2 years’ adjuvant 230. Rubin J, Ajani J, Schirmer W, et al. Octreotide acetate long-acting formula- tamoxifen: combined results of ABCSG trial 8 and ARNO 95 trial. Lancet tion versus open-label subcutaneous octreotide acetate in malignant carci- 2005;366:455–462. noid syndrome. J Clin Oncol 1999;17:600–606. 28 Antiangiogenesis Agents

Cindy H. Chau and William Douglas Figg, Sr.

INTRODUCTION situ cancers never switch to the angiogenic phenotype during a nor- mal lifetime. Such incipient tumors are usually not neovascularized Blood vessels are indispensable for tumor growth and metastasis, and and can remain harmless to the host for long periods of time as mi- the formation of a new network of blood vessels from the existing croscopic lesions that are in a state of dormancy.9,10 These nonangio- vasculature, termed angiogenesis, is one of the essential hallmarks genic tumors cannot expand beyond the initial microscopic size and of cancer development.1 Indeed, it was over 70 years ago that the cannot become clinically detectable, lethal tumors until they have existence of tumor-derived factors responsible for promoting new switched to the angiogenic phenotype11–13 through neovasculariza- vessel growth was postulated,2 and that tumor growth is essentially tion and/or blood vessel cooption.14 Depending on the tumor type dependent on vascular induction and the development of a neovas- and the environment, this switch can occur at different stages of the cular supply.3 By the late 1960s, Dr. Judah Folkman and colleagues4 tumor progression pathway and ultimately depends on a net balance had begun the search for a tumor angiogenesis factor. In the 1971 of positive and negative regulators. Thus, the angiogenic phenotype landmark report, Folkman5 proposed that inhibition of angiogen- may result from the production of growth factors by tumor cells and/ esis by means of holding tumors in a nonvascularized dormant state or the downregulation of negative modulators. would be an effective strategy to treat human cancer, and hence Changes in this angiogenic balance affecting the levels of acti- laid the groundwork for the concept behind the development of vator and inhibitor molecules dictate whether an EC will be in a antiangiogenesis agents. This fostered the search for angiogenic fac- quiescent or an angiogenic state. Normally, the inhibitors predomi- tors, regulators of angiogenesis, and antiangiogenic molecules over nate, thereby blocking growth. Once the balance shifts in favor of the next few decades and shed light on angiogenesis as an important the angiogenic state, proangiogenic factors prompts the activation, therapeutic target for the treatment of cancer and other diseases. growth, and division of vascular ECs, resulting in the formation of A decade has passed since the regulatory approval of the first new blood vessels. Activated ECs produce and release matrix metal- antiangiogenic drug bevacizumab, and while initial results were loproteinases (MMP) into the surrounding tissue to break down the regarded as highly promising, clinical evidence indicated that extracellular matrix to allow the ECs to migrate and organize them- antiangiogenic therapy also had limitations. Successful develop- selves into hollow tubes that eventually evolve into a mature net- ment and clinical translation of this novel class of agents depends work of blood vessels. Proangiogenic factors or positive regulators of on the complete understanding of the biology of angiogenesis and angiogenesis include vascular endothelial growth factor (VEGF), the regulatory proteins that govern this angiogenic process, topics basic fibroblast growth factor (PlGF), platelet-derived growth fac- that have been covered in greater detail in another section of this tor (PDGF), placental growth factor, transforming growth factor-β, textbook. This chapter will briefly review the mechanisms underly- pleiotrophins, and others.15 Activation of the hypoxia-inducible ing tumor angiogenesis followed by an in-depth discussion of anti- factor 1 (HIF-1) via tumor-associated hypoxic conditions is also angiogenic therapy, the modes of action of angiogenesis inhibitors, involved in the upregulation of several angiogenic factors.16 The and the successes and challenges of this treatment modality. angiogenic switch also involves the downregulation of angiogenesis suppressor proteins, which include endostatin, angiostatin, throm- bospondin, and others.17,18 Most notably, however, is the link be- UNDERSTANDING THE ANGIOGENIC tween many oncogenes and angiogenesis and the significant role PROCESS oncogenes play in driving the angiogenic switch.19,20 These proan- giogenic oncogenes not only induce the expression of stimulators, 21 Angiogenic Switch and Regulatory Proteins but may also downregulate inhibitors of angiogenesis.

Tumor development and progression depend on angiogenesis. Endogenous Inhibitors of Angiogenesis Recruitment of new blood vessels to the tumor site is required for the delivery of nutrients and oxygen to the cancerous growths and for the The infrequency of microscopic in situ tumors that actually un- removal of waste products.6 Cancer cells promote angiogenesis at an dergo the angiogenic switch (<1%) suggests that naturally oc- early stage of tumorigenesis, beginning with the release of molecules curring endogenous inhibitors exist in the body to defend against that send signals to the surrounding normal host tissue and stimulate the angiogenic switch in pathologic conditions and to limit physi- the migration of microvascular endothelial cells (EC) in the direc- ologic angiogenesis.9 These circulating endogenous inhibitors tion of the angiogenic stimulus. These angiogenic factors not only could also prevent microscopic metastases from growing into vis- mediate EC migration, but also EC proliferation and microvessel ible tumors. Early studies by Langer et al.22,23 demonstrated the formation in tumors undergoing the switch to the angiogenic pheno- possible existence of such inhibitors through the extraction of a type.7 Experimental evidence for this angiogenic switch was observed functional inhibitor from cartilage, a tissue that is poorly vascular- when hyperplastic islets in transgenic mice (RIP-Tag model) switch ized. Since then, dozens of endogenous angiogenesis inhibitors from small (<1 mm), white microscopic dormant tumors to red, rap- have been identified, some of which are listed in Table 28.1.17,18,24 idly growing tumors.7 Dormant tumors have been discovered dur- Many of the endogenous inhibitors of angiogenesis that have been ing autopsies of individuals who died of causes other than cancer.8 discovered to date are proteolytically cleaved fragments of larger These autopsy studies suggest that the vast majority of microscopic in proteins that are members of either the clotting/coagulation system 290

tahir99 - UnitedVRG Chapter 28 Antiangiogenesis Agents 291

TABLE 28.1 raises the possibility that such inhibitors might also be able to slow tumor metastasis. Indeed, the inhibition of angiogenesis by angio- Examples of Endogenous Inhibitors of Angiogenesis statin significantly reduced the rate of metastatic spread. Alphastatin Angiostatin DRUG DEVELOPMENT OF ANGIOGENESIS Antithrombin III (cleaved) INHIBITORS Arrestin The first angiogenesis inhibitor was reported in 1980 and involved Canstatin the low-dose administration of interferon α (IFN-α).42–44 Over the Endostatin next decade, several compounds were discovered to have potent antiangiogenic activity, including protamine and platelet factor Interferon alpha/beta (IFN-α/β) 4,45 trahydrocortisol,46 and the fumagillin analog TNP-470.47 The 2-Methoxyestradiol (2-ME) proof of concept that targeting angiogenesis is an effective strategy for treating cancer came with the approval of the first angiogenesis Pigment epithelial-derived factor (PEDF) inhibitor, bevacizumab, by the U.S. Food and Drug Administration Platelet factor 4 (PF-4) (FDA). Since then, several antiangiogenic agents have received Tetrahydrocortisol-S FDA approval for cancer treatment (Table 28.2), and three addi- tional agents (pegaptanib, ranibizumab, and aflibercept) are ap- Thrombospondin 1 proved for the treatment of wet age-related macular degeneration. Tissue inhibitor of metalloproteinase 2 (TIMP-2) Tumstatin Rationale for Antiangiogenic Therapy Vasohibin Antiangiogenic therapy stems from the fundamental concept that tumor growth, invasion, and metastasis are angiogenesis depen- dent; thus, blocking blood vessel recruitment to starve primary and metastatic tumors is a rational approach. The microvascular EC re- or members of the extracellular matrix family of glycoproteins. cruited by a tumor has become an important second target in can- Endostatin is the most well-studied endogenous angiogenesis cer therapy. Unlike the cancer cell (the primary target of cytotoxic inhibitor.25,26 Other potent endogenous angiogenesis inhibitors chemotherapy), which is genetically unstable with unpredictable include thrombospondin-127 and tumstatin.28 The discovery of

mutations, the genetic stability of ECs may make them less suscep- CANCER THERAPEUTICS vasohibin, an endogenous inhibitor that is selectively induced in tible to acquired drug resistance.48 Moreover, ECs in the microvas- ECs by proangiogenic stimulatory growth factors such as VEGF, cular bed of a tumor may support 50 to 100 tumor cells. Coupling demonstrated the existence of an intrinsic and EC-specific feed- this amplification potential together with the lower toxicity of most back inhibitor control mechanism,29,30 whereas most endogenous angiogenesis inhibitors results in the use of antiangiogenic therapy, inhibitors of angiogenesis are extrinsic to ECs. More recently, a which should be significantly less toxic than conventional chemo- second endothelium-produced negative regulator of angiogenesis therapy. However, the variable responses of antiangiogenic therapy has been discovered, the Dll4-Notch signaling system.31,32 Both observed in different tumor types and the fact that angiogenesis intrinsic factors have since been shown to control tumor angiogen- inhibitors have not delivered the benefits initially envisaged suggest esis by an autoregulatory or negative-feedback mechanism. The that the precise mechanism of action of angiogenesis inhibitors is Dll4-Notch axis has emerged as a critical regulator of tumor an- complex and remains incompletely understood. giogenesis, and inhibitors of this pathway (e.g., demcizumab, the anti-Dll4 monoclonal antibody) are currently being investigated in early phase trials of solid tumors.33 Modes of Action of Antiangiogenic Agents Perhaps the most compelling genetic evidence that endogenous inhibitors suppress pathologic angiogenesis was observed in studies Various strategies for the development of antiangiogenic drugs using mice deficient in tumstatin, endostatin, or thrombospondin have been investigated over the years, with these agents being clas- 1 (TSP-1).34 These experiments demonstrate that normal physi- sified into several different categories depending on their modes ologic levels of the inhibitors can retard the tumor growth and that of action. Some inhibit ECs directly, whereas others inhibit the their absence leads to enhanced angiogenesis and increased tumor angiogenesis signaling cascade or block the ability of ECs to break growth by two- to threefold, strongly suggesting that endogenous in- down the extracellular matrix. Inhibitors may block one main hibitors of angiogenesis can act as endothelium-specific tumor sup- angiogenic protein, two or three angiogenic proteins, or have a pressors. The connection between a tumor suppressor protein and broad-spectrum effect by blocking a range of angiogenic regulators angiogenesis is best illustrated by the classic tumor suppressor p53. that can be located in both the tumor and ECs.49 In some cases, p53 inhibits angiogenesis by increasing the expression of TSP-135 by the antiangiogenic activity is discovered as a secondary function repressing VEGF36 and basic fibroblast growth factor–binding pro- after the drug has received regulatory approval for a different pri- tein,37 and by degrading HIF-1,38 which blocks the downstream in- mary function. For example, bortezomib is a proteasome inhibi- duction of VEGF expression. New evidence suggests that p53 also tor that is approved for multiple myeloma and was later found to indirectly downregulates VEGF expression via the retinoblastoma possess antiangiogenic activity via inhibiting VEGF. Some small- pathway in a p21-dependent manner during sustained hypoxia.39 molecule drugs may display their antiangiogenic activity through Furthermore, p53-mediated inhibition of angiogenesis may also inducing the expression of endogenous angiogenesis inhibitors occur in part via the antiangiogenic activity of endostatin and tum- such as , a cyclooxygenase-2 (COX-2) inhibitor, which statin.40 This landmark finding clearly demonstrates that p53 not inhibits angiogenesis by increasing levels of endostatin.25 only controls cell proliferation, but can also repress tumor angiogen- Some drugs possess antiangiogenic properties but with mech- esis through enzymatic mobilization of these endogenous angio- anisms that are not completely understood, such as thalidomide genesis inhibitor proteins to prevent ECs from being recruited into and its analogs, lenalidomide and pomalidomide, referred to as the dormant, microscopic tumors, thereby preventing the switch to immunomodulatory drugs. Thalidomide was originally shown the angiogenic phenotype.41 The discovery that these endogenous to inhibit angiogenesis by D’Amato et al.50 in 1994 and this was angiogenesis inhibitors can suppress the growth of primary tumors subsequently confirmed in several different in vitro and ex vivo 292 Cancer Therapeutics

TABLE 28.2 Antiangiogenic Agents that Have Received U.S. Food and Drug Administration Approval for Cancer Treatment

Mechanism Year of Drug Class (Cellular Targets) Approval Indications Dosages Bevacizumab Anti-VEGF VEGF 2004 First- and second-line metastatic 5 mg/kg IV q2wk + bolus IFL; 10 (Avastin) mAB CRC mg/kg IV q2wk + FOLFOX4 2006 First-line NSCLC 15 mg/kg IV q3wk + carboplatin/ paclitaxel 2009 Second-line GBM 10 mg/kg IV q2wk 2009 Metastatic RCC 10 mg/kg IV q2wk + IFN 2013 Second-line metastatic CRC 5 mg/kg IV q2wk or 7.5 mg/kg (after prior bevacizumab- IV q3wk + fluoropyrimidine– containing regimen) irinotecan or fluoropyrimidine- oxaliplatin–based regimen Ziv-aflibercept Anti-VEGF VEGFA, VEGFB, 2012 Metastatic CRC (after prior 4 mg/kg IV q2wk (1-hr infusion) (Zaltrap, VEGF mAB PlGF1, PlGF2 oxaliplatin-containing regimen) Trap) Sorafenib Small- VEGFR2, VEGFR3, 2005 Advanced RCC 400 mg PO bid (w/o food) (Nexavar, molecule PDGFR, FLT3, 2007 Unresectable HCC 400 mg PO bid (w/o food) BAY439006) TKI c-Kit 2013 RAI-refractory DTC 400 mg PO bid (w/o food) Sunitinib (Sutent, Small- VEGFR1, VEGFR2, 2006 Imatinib-resistant or -intolerant 50 mg PO qd, 4 wk on/2 wk off SU11248) molecule VEGFR3, GIST TKI PDGFR, FLT3, 2006 Advanced RCC 50 mg PO qd, 4 wk on/2 wk off c-Kit, RET 2011 Advanced pNET 37.5 mg PO qd Pazopanib Small- VEGFR1, VEGFR2, 2009 Advanced RCC 800 mg PO qd (w/o food) (Votrient) molecule VEGFR3, 2012 Advanced soft tissue sarcoma 800 mg PO qd (w/o food) TKI PDGFR, Itk, Lck, c-Fms Vandetanib Small RET, VEGFR, 2011 Advanced MTC 300 mg PO qd (Caprelsa) molecule EGFR, BRK, TKI TIE2 Axitinib (Inlyta) Small VEGFR1, VEGFR2, 2012 Advanced RCC (after failure of 5 mg PO bid molecule VEGFR3 prior therapy) TKI Cabozantinib Small MET, VEGFR2, 2012 Progressive, metastatic MTC 140 mg PO qd (w/o food) (XL184, molecule RET, KIT, AXL, Cometriq) TKI FLT3 Regorafenib Small RET, VEGFR1, 2012 Previously treated metastatic 160 mg PO qd × 21days (q28-day (Stivarga) molecule VEGFR2, CRC cycle) TKI VEGFR3, TIE2, 2013 GIST 160 mg PO qd × days 1–21 KIT, PDGFR (q28-day cycle) Temsirolimus mTOR mTOR 2007 Advanced RCC 25 mg IV qwk (infused over (Torisel) inhibitor 30–60 min) Everolimus mTOR mTOR 2009 Second-line advanced RCC (after 10 mg PO qd (Afinitor, RAD- inhibitor VEGFR TKI failure) 001)a 2010 SEGA associated w/TSC 4.5 mg/m2 PO qd 2011 pNET 10 mg PO qd 2012 Advanced HR+, HER2- breast 10 mg PO qd cancer 2012 AML associated w/TSC 10 mg PO qd a Afinitor Disperz (everolimus tablets for oral suspension) was approved in 2012 for children aged 1 and older who have SEGA+ TSC. mAB, monoclonal antibody; CRC, colorectal cancer; IV, intravenous; IFL, irinotecan, 5-fluorouracil, and leucovorin; FOLFOX4, 5-flourouracil, leucovorin, and oxaliplatin; NSCLC, non–small-cell lung cancer; GBM, glioblastoma multiforme; RCC, renal cell carcinoma; VEGFA, vascular endothelial growth factor A; PlGF, placental growth factor; TKI, tyrosine–kinase inhibitor; VEGFR, VEGF receptor; PDGFR, platelet-derived growth factor receptor; FLT, Fms-like tyrosine kinase; c-Kit, stem cell factor receptor; HCC, hepatocellular carcinoma; RAI, radioactive iodine; DTC, differentiated thyroid carcinoma; PO, orally; RET, glial cell line-derived neurotrophic factor receptor; pNET, pancreatic neuroendocrine tumor; GIST, gastrointestinal stromal tumor; qd, every day; Itk, interleukin-2 receptor inducible T-cell kinase; Lck, leukocyte-specific protein tyrosine kinase; c-Fms, transmembrane glycoprotein receptor tyrosine kinase; bid, twice daily; EGFR, epidermal growth factor receptor; BRK, protein tyrosine kinase 6; MTC, medullary thyroid cancer; mTOR, mammalian target of rapamycin; SEGA, subependymal giant cell astrocytoma; TSC, tuberous sclerosis complex; HR, hormone receptor; HER2, human epidermal growth factor receptor 2; AML, angiomyolipoma.

tahir99 - UnitedVRG Chapter 28 Antiangiogenesis Agents 293

assays. 51–54 Interestingly, unlike other mechanisms of action, the most cancer cell types. It stimulates EC proliferation, migration, antiangiogenic activity of thalidomide is believed to require enzy- and survival as well as induces increased vascular permeability. matic activation. The extent to which the antiangiogenic properties The different forms of VEGF bind to transmembrane recep- of thalidomide and its analogs play a role in its antimyeloma activity tor tyrosine kinases (RTK) on ECs: VEGFR1 (Flt-1), VEGFR2 is not clearly understood. Several mechanisms have been proposed (KDR/Flk-1 or kinase insert domain receptor/fetal liver kinase 1), that involve the downregulation of cytokines in EC, the inhibition or VEGFR3 (Flt-4).60 This results in receptor dimerization, acti- of EC proliferation, the decrease in the level of circulating ECs, vation, and autophosphorylation of the tyrosine–kinase domain, or the modulation of adhesion molecules between the multiple thereby triggering downstream signaling pathways. Other signal- myeloma cells and the endogenous bone marrow stromal cells, ing molecules that may represent attractive therapeutic targets thereby decreasing the production of VEGF and interleukin 6 include PDGF and the angiopoietins (Ang1, Ang2). PDGF-B/ (IL-6).55–59 The immunomodulatory agents are discussed in greater PDGF receptor (R)-β plays an important role in the recruitment detail in another section of this textbook. Examples of the various of pericytes and maturation of the microvasculature.61 Ang2, types of angiogenesis inhibitors are highlighted in Table 28.3. which binds the Tie-2 receptor, is mostly expressed in tumor-in- Drugs with antiangiogenic activity may be classified as either di- duced neovasculature, whereby its selective inhibition results in rect or indirect angiogenesis inhibitors. A direct angiogenesis inhibitor reduced EC proliferation.62 The angiopoietins are also involved in blocks vascular ECs from proliferating, migrating, or increasing their lymphangiogenesis, the formation of new lymphatic vessels, which survival in response to proangiogenic proteins. They target the acti- plays a key role in tumor metastasis. An increased Ang2/Ang1 ratio vated endothelium directly and inhibit multiple angiogenic proteins. correlates with tumor angiogenesis and poor prognosis in many Examples of direct angiogenesis inhibitors include many of the endog- cancers, thus making the angiopoietins an attractive therapeutic enous inhibitors of angiogenesis, such as endostatin, angiostatin, and target. Angiopoietin inhibitors are currently under investigation in TSP-1. Indirect angiogenesis inhibitors decrease or block expression the preclinical and clinical setting. of a tumor cell product, neutralize the tumor product itself, or block Other strategies for targeting angiogenesis involve the tumor mi- its receptor on ECs. The limitation to indirect inhibitors is that, over croenvironment. Breakdown of the extracellular matrix is required time, tumor cells may acquire mutations that lead to increased expres- to allow ECs to migrate into surrounding tissues and proliferate sion of other proangiogenic proteins that are not blocked by the indi- into new blood vessels; thus, drugs that target MMPs, enzymes that rect inhibitor. This may give the appearance of drug resistance and catalyze the breakdown of the matrix, can also inhibit angiogen- warrants the addition of a second antiangiogenic agent, one that would esis. However, clinical development of MMP inhibitors (MMPI) target the expression of these upregulated proangiogenic proteins. Ex- has yielded disappointing results.63–66 amples of drugs that interfere with the angiogenesis-signaling pathway Integrins are cell surface adhesion molecules that play an es- include the anti-VEGF monoclonal antibodies and small-molecule sential role in cell–cell and cell–matrix adhesion as well as in trans- tyrosine–kinase inhibitors. These drugs target the major signaling mitting signals important for cell migration, invasion, proliferation, pathways in tumor angiogenesis: VEGF, PDGF, and their respective and survival. The involvement of integrin in tumor angiogenesis CANCER THERAPEUTICS receptors, as well as other growth factors and/or signaling pathways. was demonstrated in studies that show the β-4 subunit of integrin VEGF (also known as vascular permeability factor) is a potent promoting endothelial migration and invasion.67 Agents that target proangiogenic growth factor and its expression is upregulated by integrins (inhibitors of αvβ3 and αvβ5) have been evaluated as poten- tial therapeutic options and include etaracizumab, cilengitide, and intetumumab. However, all three integrin inhibitors have proven to 68–73 TABLE 28.3 be largely ineffective in various early and late stage cancer trials. In summary, the downstream effects of antiangiogenic agents, in ad- Examples of Drugs that Possess Antiangiogenic dition to blocking angiogenesis, may involve inducing vessel regres- Activity or Inhibit Angiogenesis as a Secondary sion, promoting sensitization to radiotherapy and chemotherapy by Function depriving ECs of VEGF’s prosurvival signals, and inhibiting the recruitment of proangiogenic bone marrow–derived cells as well as Drug Class reducing the self-renewal capability of cancer stem cells.74 Cetuximab EGFR/HER monoclonal Panitumumab antibodies CLINICAL UTILITY OF APPROVED Trastuzumab ANTIANGIOGENIC AGENTS IN CANCER Gefitinib EGFR small-molecule tyrosine– THERAPY Erlotinib kinase receptor inhibitors Everolimus mTOR inhibitors The following section reviews the current FDA-approved angiogen- Temsirolimus esis inhibitors (Table 28.2). These agents include: (1) the mono- clonal anti-VEGF antibodies (bevacizumab and ziv-aflibercept); Thalidomide Immunomodulatory agents (2) small-molecule tyrosine–kinase inhibitors (TKI) (sorafenib, Lenalidomide sunitinib, pazopanib, vandetanib, axitinib, cabozantinib, and rego- Pomalidomide rafenib); and (3) the mammalian target of rapamycin (mTOR) Belinostat (PXD101) HDAC inhibitors inhibitors (temsirolimus and everolimus), as examples of drugs LBH589 that possess antiangiogenic activity. Other approved drugs that also Vorinostat (SAHA) inhibit angiogenesis as a secondary function, such as thalidomide, Celecoxib COX-2 inhibitors are discussed in greater detail in another section of this textbook and are presented in Table 28.3. Bortezomib Proteasome inhibitors Zoledronic acid Bisphosphonates Anti-VEGF Therapy Rosiglitazone PPAR-γ agonists Doxycycline Antibiotic Bevacizumab

EGFR, epidermal growth factor receptor; mTOR, mammalian target of Bevacizumab is a recombinant humanized anti–VEGF-A mono- rapamycin HDAC, histone deacetylase; COX-2, cyclooxygenase-2; PPAR, clonal antibody that received FDA approval in February 2004 for peroxisome proliferator–activated receptor. use in combination therapy with fluorouracil-based regimens for 294 Cancer Therapeutics metastatic colorectal cancer. Bevacizumab binds VEGF and pre- substantially greater than that of the chemotherapy alone.89 Side vents the interaction of VEGF to its receptors (Flt-1 and KDR) on effects included grade 3 hypertension, grade 1 or 2 proteinuria, a the surface of ECs. It is the first antiangiogenic agent clinically slight increase (less than two percentage points) in grade 3 or 4 proven to extend survival following a large, randomized, double- bleeding, and impaired surgical wound healing in patients who blind, phase III study in which bevacizumab was administered in underwent surgery during treatment with bevacizumab. However, combination with bolus irinotecan, 5-fluorouracil, and leucovorin potentially life-threatening events (e.g., arterial and venous throm- (IFL) as first-line therapy for metastatic colorectal cancer (CRC).75 boembolic events, gastrointestinal perforation, hemoptysis, risk of In 2006, its approval extended to first- or second-line treatment of ovarian failure) have occurred in some patients, thus requiring patients with metastatic carcinoma of the colon or rectum. This close patient monitoring in individuals who are at greater risk of recommendation is based on the demonstration of a statistically adverse events.90 In a recent meta-analysis of RCTs, bevacizumab significant improvement in overall survival (OS) in patients receiv- in combination with chemotherapy or biologic therapy, compared ing bevacizumab plus FOLFOX4 (5-flourouracil, leucovorin, and with chemotherapy alone, was associated with increased treat- oxaliplatin) when compared to those receiving FOLFOX4 alone. ment-related mortality.91 In January 2013, it was further approved to treat mCRC for sec- Although four phase III randomized studies have demonstrated ond-line treatment when used with fluoropyrimidine-based (com- improvements in PFS for ovarian cancer (OC)—two first-line bined with irinotecan or oxaliplatin) chemotherapy after disease trials (GOG 218 and ICON7) and two in recurrent OC [plati- progression following a first-line treatment with a bevacizumab- num-resistant (AURELIA Trial) or platinum-sensitive (OCEANS containing regimen based on clinical benefits observed in the ran- Trial)]—the role of bevacizumab in OC remains controversial. domized phase III study (ML18147).76 Despite the benefit in the Bevacizumab is approved for use in combination with chemo- metastatic setting, the addition of bevacizumab did not improve therapy in the first- and second-line treatment of advanced OC clinical outcomes in the adjuvant setting in CRC.77,78 In 2006, in Europe, but it is not currently licensed in the United States bevacizumab received an additional approval for use in combina- for this indication. Mature OS data and predictive biomarkers are tion with carboplatin and paclitaxel, and is indicated for first-line key to defining the subsets of patients who will most like benefit treatment of patients with unresectable, locally advanced, recur- from this therapy. More recently, a randomized, phase III trial rent, or metastatic nonsquamous, non–small-cell lung cancer (GOG240) has demonstrated for the first time that bevacizumab (NSCLC) based on the demonstration of a statistically significant can prolong OS and PFS for women with advanced, recurrent, improvement in OS in patients in the bevacizumab arm compared or persistent cervical cancer that was not curable with standard to those receiving chemotherapy alone.79 In February 2008, the chemotherapy. At the time of writing, there are currently over 400 FDA granted a conditional, accelerated approval for bevacizumab actively recruiting, ongoing trials investigating the clinical benefits to be used in combination with paclitaxel for the treatment of pa- of bevacizumab in combination with chemotherapeutic regimens tients who have not received chemotherapy for metastatic human or as adjuvant therapy in various stages and types of cancer (http:// epidermal growth factor receptor 2 (HER2)-negative breast cancer. clinicaltrial.gov). However, additional clinical trials were conducted and the new data showed only a small effect on progression free survival (PFS) Ziv-aflibercept without evidence of an improvement in OS or a clinical benefit to Ziv-aflibercept (previously known as aflibercept or VEGFTrap) is patients sufficient to outweigh the risks; thus, the FDA rescinded a recombinant humanized fusion protein of the extracellular do- its approval and removed the breast cancer indication from the mains of VEGF receptor 1 (VEGFR1) and VEGFR2 with the con- drug’s label in November 2011.80–82 This controversial decision stant region (Fc) of human immunoglobulin (Ig)G1 that binds to continues to be debated with ongoing subgroup analyses to iden- VEGF-A, VEGF-B, PlGF1, and PlGF2, thereby preventing these tify patients who would likely benefit from bevacizumab. ligands from binding to and activating their cognate receptors.92 Bevacizumab received another accelerated approval as a single Ziv-aflibercept has a higher VEGF-A binding affinity and more agent for patients with glioblastoma multiforme (GBM) with pro- potent blockade of VEGFR1 or VEGFR2 activation than bevaci- gressive disease following therapy in May 2009. The approval was zumab.93 In tumor models, ziv-aflibercept exerts its antiangiogenic based on the demonstration of durable objective response rates ob- effects through regressing tumor vasculature and size, remodeling served in two single-arm trials, AVF3708g and NCI 06-C-0064E.83 or normalizing surviving vasculature, and inhibiting ascites for- Currently, no data have shown whether bevacizumab improves mation.94 In August 2012, ziv-aflibercept received regulatory ap- disease-related symptoms or survival in people previously treated proval for use in combination with 5-fluorouracil, leucovorin, and for GBM. Moreover, phase III trials of bevacizumab in newly irinotecan (FOLFIRI) for the treatment of patients with metastatic diagnosed GBM (RTOG 8025 and AVAglio) have shown a 3- to CRC that is resistant to or that has progressed following treatment 4-month improvement of PFS, but no OS advantage over the stan- with an oxaliplatin-containing regimen. Results from the pivotal dard of care.84 The AVAglio trial improved patients’ quality of life, phase III VELOUR trial showed that ziv-aflibercept plus FOLFIRI whereas the RTOG 0825 did not and instead increased the burden statistically and significantly improved PFS (median PFS, 6.90 of symptoms with a negative impact on cognition. Although these versus 4.67 months, respectively), OS (median OS, 13.50 versus two studies showed that bevacizumab had a modest benefit as the 12.06 months, respectively), and overall response rates (19% versus initial therapy for GBM, it remained effective to treat recurrences 11.1%, respectively) relative to placebo plus FOLFIRI.95 Toxicities where treatment options are limited. In July 2009, bevacizumab related to ziv-aflibercept were consistent with those expected from was approved for use in combination with IFN-α for the treatment the anti-VEGF drug class. The frequency of vascular-related ad- of patients with metastatic renal cell carcinoma (RCC). Results verse events appeared to be higher with ziv-aflibercept than beva- from the AVOREN trial demonstrated a 5-month improvement in cizumab treatment when compared across trials. Current clinical median PFS in patients treated with bevacizumab plus IFN-α-2a data are insufficient to directly compare ziv-aflibercept and beva- versus IFN-α-2a plus placebo.85 Another phase III trial (CALGB cizumab in the first- or second-line setting for metastatic CRC. 90206) of bevacizumab plus IFN-α versus IFN-α monotherapy was conducted in patients with previously untreated, metastatic clear cell RCC. Median PFS was 8.4 months versus 4.9 months in Tyrosine–Kinase Inhibitor Therapy favor of the bevacizumab arm.86 Both studies did not demonstrate 87,88 a statistically significant advantage in OS. Sorafenib Clinical studies of bevacizumab in combination with oxalipl- atin-containing and 5-fluorouracil–based regimens have shown Sorafenib is a small-molecule Raf kinase and VEGF receptor that combination therapy is well tolerated with toxicity not being kinase (VEGFR2 and VEGFR3) inhibitor. It has been shown to

tahir99 - UnitedVRG Chapter 28 Antiangiogenesis Agents 295 exhibit broad-spectrum effects on multiple targets (PDGF recep- compared with 5.4 months for patients treated with placebo (HR, tor (PDGFR), stem cell factor (c-KIT) receptor, p38) that affect 0.427, 95% CI, 0.271 to 0.673], p <0.001).107 Common adverse the maintenance of the tumor vasculature and angiogenesis.96 In effects, including diarrhea, mucositis, asthenia, skin abnormalities, December 2005, the FDA granted approval for sorafenib, which and altered taste, were more common in patients receiving suni- is considered the first multikinase inhibitor, for the treatment of tinib. In addition, a decrease in left ventricular ejection fraction patients with advanced RCC. Safety and efficacy of sorafenib was and severe hypertension were also more commonly reported in the proven in the largest randomized phase III study conducted in ad- sunitinib arm. Grade 3 or 4 treatment-emergent adverse events were vanced RCC that showed prolong PFS in favor of sorafenib.97,98 reported in 56% versus 51% of patients on sunitinib versus placebo, In November 2007, sorafenib was approved for the treatment of respectively. patients with unresectable hepatocellular carcinoma (HCC) based on the study results in patients with advanced HCC who Pazopanib had not received previous systemic treatment. Median survival and Pazopanib is a second-generation, multitargeted TKI that binds to the time to radiologic progression were nearly 3 months longer for VEGFR-1, -2, -3, PDGFR-α and -β, c-KIT, and several other key patients treated with sorafenib than for those given placebo.99 In proteins responsible for angiogenesis, tumor growth, and cell sur- November 2013, sorafenib received a new indication under the vival. Pazopanib exhibited in vivo and in vitro activity against tumor FDA’s priority review program for the treatment of locally recur- growth, and early clinical trials demonstrated potent antitumor and rent or metastatic, progressive differentiated thyroid carcinoma antiangiogenic activity.108 A phase III clinical trial in treatment- (DTC) refractory to radioactive iodine (RAI) treatment based on naïve and cytokine-pretreated patients with advanced and/or meta- positive results from the phase III DECISION trial. Treatment static RCC showed a significant improvement in PFS and tumor with sorafenib improved PFS (the primary endpoint of the trial) by response compared with placebo,109 leading to the approval of pa- 41% compared with placebo (10.8 versus 5.8 months, respectively; zopanib in the United States in October 2009. A recent, random- hazard ratio [HR], 0.587, 95% confidence interval [CI] [0.454 to ized phase III trial (COMPARZ) compared the efficacy and safety 0.758]; p <0.0001).100 The overall response rates were 12% for of pazopanib and sunitinib as first-line therapy involving patients patients who received sorafenib versus 1% for the placebo arm. with metastatic RCC and demonstrated that both pazopanib and Although only about 5% to 15% of thyroid cancer patients become sunitinib have similar efficacy, but the safety and quality-of-life pro- refractory to RAI, no standard treatments are available and, thus, files favor pazopanib.110 In April 2012, pazopanib was approved for sorafenib is the first agent specifically approved for RAI-resistant the treatment of patients with metastatic nonadipocytic soft tissue DTC. Sorafenib was generally well tolerated with a predictable sarcoma who have received prior chemotherapy following a phase safety profile. Common adverse events include diarrhea, rash/des- III trial that demonstrated a statistically significant improvement in quamation, fatigue, hand–foot skin reaction, alopecia, and nausea/ PFS. The median PFS was 4.6 months for patients receiving pazo- vomiting. Grade 3/4 adverse events were 38% for sorafenib versus panib versus 1.6 months for the placebo arm.111 The drug is gener- 28% for placebo. Sorafenib-induced hypertension occurred in pa- CANCER THERAPEUTICS ally well tolerated, with the most common adverse events being tients with metastatic RCC. The treatment-related hypertension diarrhea, fatigue, anorexia, hypertension, and hair depigmenta- was noted to be a class effect observed not only with VEGFR in- tion, as well as laboratory abnormalities in elevated aspartate ami- hibitors, but also with the VEGF monoclonal antibody as well.90 notransferase and alanine aminotransferase. Pazopanib has shown No significant relationship between previously described media- clinical activity in a variety of tumors, including breast cancer, thy- tors of blood pressure and the magnitude of increase was found in roid cancer, HCC, and cervical cancer.112 Ongoing phase II and a study evaluating the mechanism of sorafenib-induced hyperten- III trials are further evaluating pazopanib in these malignancies. sion in patients.101 Vandetanib Sunitinib Vandetanib is an oral, small-molecule TKI that inhibits the ac- Sunitinib (SU11248) is a small-molecule, multitargeted TKI that tivity of RET kinase, VEGFR, epidermal growth factor receptor exhibits potent antitumor and antiangiogenic activity and inhibits (EGFR), protein tyrosine kinase 6 (BRK), TIE2, members of the VEGFR-1, -2, -3, c-KIT, PDGFR; FLT-3; colony-stimulating fac- ephrin (EPH) receptors kinase family, and members of the Src tor receptor type 1 receptor; and the glial cell line–derived neuro- family of tyrosine kinases.113 Vandetanib reduced endothelial cell trophic factor receptor. It was rationally designed and chosen for migration, proliferation, survival, and angiogenesis in vitro, and it its high bioavailability and its nanomolar-range potency against the decreased tumor vessel permeability and inhibited tumor growth antiangiogenic RTKs. Sunitinib received its first U.S. regulatory ap- and metastasis in vivo. In April 2011, vandetanib received U.S. proval in 2006 for the treatment of gastrointestinal stromal tumor regulatory approval for the treatment of symptomatic or progres- (GIST) after disease progression on, or intolerance to, imatinib and sive medullary thyroid cancer (MTC) in patients with unresect- accelerated approval for the treatment of advanced RCC.102 Suni- able, locally advanced, or metastatic disease. Until the approval tinib demonstrated significant efficacy (prolonged median time to of vandetanib, no systemic therapy was approved for the treatment progression) in imatinib-resistant or -intolerant GIST in a random- of unresectable MTC, making it the first molecularly targeted ized phase III trial.103 The accelerated approval for RCC was based agent approved for this disease. Results of a randomized phase III on durable partial responses, with a response rate of 26% to 37%, and trial of patients with unresectable, locally advanced, or metastatic a median duration of response of 54 weeks from two phase II, single- MTC demonstrated statistically significant and clinically meaning- arm trials of patients with cytokine-refractory RCC.104 The acceler- ful improvements in PFS for vandetanib compared with placebo ated approval was converted to regular approval in 2007 following (HR, 0.46; 95% CI, 0.31 to 0.69; p <0.001).114 Common grade 3 confirmation of an improvement in PFS and OS in a phase III trial and 4 toxicities (>5%) were diarrhea and/or colitis, hypertension of sunitinib for first-line treatment of patients with treatment-naïve, and hypertensive crisis, fatigue, hypocalcemia, rash, and corrected metastatic RCC.105,106 In May 2011, the drug received a new indi- QT interval (QTc) prolongation. Given the toxicity profile, which cation for the treatment of progressive, well-differentiated pancre- includes QTc prolongation and sudden death, vandetanib is only atic neuroendocrine tumors (pNET) in patients with unresectable, available through a restricted distribution program. Vandetanib is locally advanced, or metastatic disease. The randomized phase III also the first targeted drug to show evidence of efficacy in a random- trial was discontinued early after the independent data monitoring ized phase II trial in patients with locally advanced or metastatic committee observed more serious adverse events and deaths in the differentiated thyroid carcinoma,115 and a phase III trial is currently placebo group as well as a difference in PFS favoring sunitinib. The underway. Early phase studies are also being conducted in solid median PFS for patients treated with sunitinib was 10.2 months, tumors, including GIST and kidney and pancreatic cancers. 296 Cancer Therapeutics

Axitinib sorafenib and differs from the latter by the presence of a fluorine atom in the center phenyl ring, resulting in higher inhibitory po- Axitinib is a potent and selective second-generation inhibitor of tency against various proangiogenic receptors than sorafenib, in- VEGFR-1, -2, and -3. The in vitro half-maximal inhibitory con- cluding VEGFR2 and FGFR1. In September 2012, regorafenib centration (IC50) of axitinib is 10-fold lower for the VEGF fam- was approved for the treatment of patients with mCRC who have ily of receptors than for other TKIs such as pazopanib, sunitinib, 116 been previously treated with fluoropyrimidine-, oxaliplatin-, and or sorafenib. In January 2012, axitinib received approval for the irinotecan-based chemotherapy, with an anti-VEGF therapy, treatment of advanced RCC after the failure of one prior systemic and if KRAS wild type, with an anti-EGFR therapy. The phase therapy based on a phase III trial (AXIS) comparing the efficacy III CORRECT trial that resulted in approval of the drug demon- and safety of axitinib versus sorafenib as a second-line treatment for 117,118 strated a median OS of 6.4 months in the regorafenib group ver- metastatic RCC. The median PFS was 6.7 months with ax- sus 5.0 months in the placebo group (HR, 0.77; 95% CI, 0.64 to itinib compared to 4.7 months with sorafenib (HR, 0.67; 95% CI, = 127 < 0.94; one-sided p 0.0052). Regorafenib is the first TKI with 0.54, 0.81; one-sided p 0.0001). This improvement in PFS was survival benefits in mCRC that has progressed after all standard greater in the cytokine-pretreated subgroup in comparison with the therapies. In February 2013, it received another indication for sunitinib-pretreated subgroup. The most frequent adverse events the treatment of patients with locally advanced, unresectable, or with axitinib were diarrhea (all grade), hypertension (all grade), metastatic GIST who have been previously treated with imatinib fatigue, decreased appetite, nausea, and dysphonia. Moreover, and sunitinib. This was based on positive findings of the phase III hypertension, nausea, dysphonia, and hypothyroidism were more GRID trial that demonstrated a median PFS of 4.8 months for common with axitinib, whereas palmar–plantar erythrodysesthe- regorafenib and 0.9 months for placebo (HR, 0.27, 95% CI, 0.19 sia, alopecia, and rash were more frequent with sorafenib. A phase to 0.39; p <0.0001).128 In both studies, regorafenib provided sig- III trial (AGILE) comparing axitinib with sorafenib as first-line nificant improvements in PFS to highly refractory patient popu- therapy in patients with treatment-naïve metastatic RCC demon- lations who have progressed on standard treatments. The most strated no significant difference in median PFS between patients 119 common adverse events that were grade 3 or higher and related treated with axitinib or sorafenib. Additionally, axitinib is being to regorafenib were hand–foot skin reaction, fatigue, diarrhea, hy- studied as a single agent as well as in combination with chemo- pertension, and rash or desquamation. Its clinical development as therapy across several tumor types including HCC, NSCLC, and a single agent or in combination with standard chemotherapeu- pancreatic and thyroid cancers. tic agents in various malignant tumors is ongoing and includes a phase III trial in patients with HCC whose disease has progressed Cabozantinib after treatment with sorafenib. Cabozantinib (XL184) is a small-molecule TKI with potent activ- ity toward the MET receptor and VEGFR2, as well as a number mTOR Inhibitors of other receptor tyrosine kinases, including RET, KIT, AXL, and FLT-3. MET is the only known receptor for hepatocyte growth fac- The mTOR pathway is a central component of the PI3K/Akt tor (HGF), and its signaling activity plays a key role in tumorigenic signaling pathway and a regulator of many biologic processes growth, metastasis, and therapeutic resistance. The dysregulated that are essential for angiogenesis, cell proliferation, and metab- expression and/or activation of MET and HGF have been impli- olism.129 Inhibition of the mTOR kinase prevents downstream cated in the development of numerous human cancers including signaling via the Akt pathway, resulting in inhibition of protein glioma; melanoma; and hepatocellular, renal, gastric, pancreatic, translation and cell growth. mTOR plays a key role in angiogen- prostate, ovarian, breast, and lung cancers, and is often correlated esis and specifically regulates the expression of HIF-1, which with poor prognosis.120 Recent studies have determined that the is upregulated by the loss of the von Hippel–Lindau gene in MET pathway plays an important role in the development of re- RCC. In May 2007, temsirolimus was approved for the treat- sistance to VEGF pathway inhibition and that the use of VEGFR ment of advanced RCC. Efficacy and safety were demonstrated inhibitors, such as sunitinib, sorafenib, or a VEGFR2-targeting in a phase III study in previously untreated patients (n = 626) antibody, can result in the development of an aggressive tumor with poor risk features of metastatic RCC assigned to one of phenotype characterized by increased invasiveness and metasta- three treatment arms: IFN-α alone, temsirolimus 25 mg alone, sis.121–123 Thus, there is an advantage to targeting both the MET or the combination of temsirolimus (15 mg) and IFN-α.130 and VEGF pathways to disrupt angiogenesis, tumorigenesis, and Single-agent temsirolimus was associated with a statistically cancer progression. In November 2012, cabozantinib received significant improvement in OS when compared with IFN; the U.S. regulatory approval for progressive metastatic MTC based addition of temsirolimus to IFN did not improve OS. The re- on the phase III trial that demonstrated a statistically significant sults of the phase III INTORSECT trial compared the efficacy PFS prolongation for the cabozantinib-treatment arm.124 The es- of temsirolimus and sorafenib in the second-line treatment of timated median PFS was 11.2 months for cabozantinib versus 4.0 metastatic RCC after disease progression on sunitinib demon- months for placebo (HR, 0.28; 95% CI, 0.19 to 0.40; p <0.001). strated that temsirolimus did not improve survival over sorafenib Manageable toxicities included diarrhea, palmar–plantar erythro- in the second-line setting.131 The significant OS difference in dysesthesia, decreased weight and appetite, nausea, and fatigue. favor of sorafenib (stratified HR, 1.31; 95% CI, 1.05 to 1.63; Cabozantinib has been effective against several solid cancers, two-sided p = 0.01) suggested that VEGFR inhibition may be a including MTC, breast, NSCLC, melanoma, and liver cancer, better option than mTOR inhibitors for patients progressing on and is currently being studied in clinical trials in a number of sunitinib. The most common adverse reactions that occurred tumor types, with the most significant results observed in the re- were rash, asthenia, mucositis, nausea, edema, and anorexia. duction of bone metastatic lesions in castration-resistant prostate Rare, but serious adverse reactions associated with temsirolimus cancer.125 included interstitial lung disease, bowel perforation, and acute Regorafenib renal failure. Everolimus (RAD001) was approved in March 2009 for patients Regorafenib is a small-molecule TKI of multiple membrane-bound with advanced RCC whose disease had progressed on VEGFR-tar- and intracellular kinases including RET, VEGFR1, VEGFR2, geted therapy (sunitinib or sorafenib). Efficacy was demonstrated in VEGFR3, KIT, PDGFR-α, PDGFR-β, FGFR1, FGFR2, TIE2, a phase 3 trial that study met its primary endpoint with a median DDR2, TrkA, Eph2A, RAF-1, BRAF, BRAFV600E, SAPK2, PFS of 4.9 and 1.9 months in the everolimus and placebo arms, re- PTK5, and Abl pathways.126 Regorafenib is structurally related to spectively (HR, 0.33; p <0.0001).132 Everolimus is also indicated for

tahir99 - UnitedVRG Chapter 28 Antiangiogenesis Agents 297 subependymal giant cell astrocytoma (SEGA) associated with tuber- chemotherapy, sustains antiangiogenic activity and reduces acute 142 ous sclerosis complex (TSC), renal angiomyolipoma with TSC, pro- toxicity. Thus, the efficacy of metronomic chemotherapy may gressive neuroendocrine tumors of pancreatic origin, and advanced increase when administered in combination with specific antian- hormone receptor-positive, HER2-negative breast cancer in com- giogenic drugs. Another model addresses the use of antiangiogenic bination with exemestane.133 The most common adverse reactions drugs to slow down tumor cell repopulation between successive 143 were stomatitis, infections, asthenia, fatigue, cough, and diarrhea. cycles of cytotoxic chemotherapy. This model underscores the The most common grade 3/4 adverse reactions were infections, importance of timing and sequence in achieving the maximal ther- dyspnea, fatigue, stomatitis, dehydration, pneumonitis, abdominal apeutic benefit from combination therapies. In fact, a preclinical pain, and asthenia. Both temsirolimus and everolimus are currently study in murine tumor models demonstrated that the administration being evaluated in phase I through III studies of various cancer of sunitinib markedly reduced chemotherapy-induced bone marrow types. By downregulating HIF-1 in the tumor cell, mTOR inhibitors toxicity, suggesting that the sequential treatment regimen (delivery may complement the effects of TKIs at the level of the EC; thus, the of antiangiogenics followed by chemotherapy) showed superior sur- combination of mTOR inhibitors with other targeted agents such vival benefits compared with the simultaneous administration of two 144 as bevacizumab or sorafenib/sunitinib are also being investigated. drugs. Finally, other mechanisms that might also contribute to the synergism include angiogenesis inhibitor–induced tumor blood vessel regression, the prevention of tumor coopting of vessels from On the Horizon: Anti-VEGFR2 Monoclonal surrounding healthy tissues, and the formation of abnormal vessels Antibody in the tumor microenvironment.145 Nevertheless, it remains a chal- lenge to determine why bevacizumab has proved largely ineffective Ramucirumab (IMC-1121B) is a fully human IgG1 monoclonal as a single agent, whereas VEGF RTK inhibitors have repeatedly antibody that binds with high affinity to the extracellular VEGF- failed in randomized phase III trials when used in combination with binding domain of VEGFR-2. In a phase III trial (REGARD), chemotherapy. Furthermore, an additional challenge is to deter- ramucirumab monotherapy conferred a statistically significant mine the optimal dose and duration of antiangiogenic drugs as well benefit in OS and PFS compared to placebo in patients with ad- as the impact of drug sequencing in combination regimens. Studies vanced gastric or gastroesophageal junction adenocarcinoma in are warranted to delineate the discrepancy of bevacizumab’s efficacy the second-line setting with an acceptable safety profile.134 The in the macrometastatic versus micrometastatic disease settings.146,147 survival advantage is the first to be elicited by a single-agent bio- logic treatment in this setting and, based on these findings, the FDA has assigned a priority review designation for ramucirumab. BIOMARKERS OF ANTIANGIOGENIC An ongoing phase III trial (RAINBOW) of ramucirumab in com- THERAPY bination with chemotherapy as second-line treatment for patients with advanced gastric cancer is currently underway, and prelimi- Antiangiogenic therapy has created a need to develop effective CANCER THERAPEUTICS nary results demonstrated the trial met both its primary (OS) and biomarkers to assess the activity of these inhibitors. Biomarkers of secondary (PFS) endpoints. In April 2014, the U.S. FDA approved tumor angiogenesis activity are important to guide clinical devel- ramucirumab for use as a single agent for the treatment of patients opment of these agents and to select patients most likely to benefit with advanced or metastatic, gastric or gastroesophageal junction from this approach. Although there are currently no validated bio- adenocarcinoma with disease progression on or after prior treat- markers for clinically assessing the efficacy of or selecting patients ment with fluoropyrimidine- or platinum-containing chemo- who will respond to antiangiogenic therapies, a number of candi- therapy. The recommended ramucirumab dose and schedule is date markers, including tissue, imaging, and circulating biomark- 8 mg/kg administered as a 60-minute intravenous infusion every ers, are emerging that need to be prospectively validated.148,149 2 weeks. The drug also marginally improved survival in the second- Several avenues are currently being investigated and include tumor line treatment of NSCLC in an ongoing phase III (REVEL) trial. biopsy analysis, microvessel density, noninvasive vascular imaging modalities (positron-emission tomography, dynamic contrast-en- hanced magnetic resonance imaging), and measuring circulating COMBINATION THERAPIES biomarkers (levels of angiogenic factors in serum, plasma, urine, or circulating ECs and their precursors).150–152 Recent research ef- Tumor angiogenesis is a highly complex process involving multiple forts have focused on identifying genetic and toxicity biomarkers to growth factors and their receptor signaling pathways. Based on cur- predict which patients will benefit from anti-VEGF/VEGFR ther- rent evidence, with a few exceptions, effective therapy will probably apy and identify patients at risk of adverse events. The existence of rely on a combinatorial approach that involves targeting multiple VEGF single-nucleotide polymorphisms (SNP) and their associa- pathways simultaneously. However, a recent study has demonstrated tion with clinical outcomes may be predictive of patient response that simultaneous inhibition of the VEGF and EGF pathways in to bevacizumab. A recent study identified a locus in VEGFR1 that combination with chemotherapy shortens rather than prolongs PFS correlated with increased VEGFR1 expression and poor bevaci- as compared to inhibition of the VEGF pathway alone in combina- zumab treatment outcomes.153 Moreover, a breast cancer study tion with chemotherapy.135 Whether other targeted agents exhibit (E2100) reported the VEGF-2578 AA and VEGF-1154 AA geno- beneficial effects when combined with VEGF inhibitors remains types predicted an improved median OS, whereas the VEGF-634 to be investigated. Moreover, a number of studies have shown that CC and VEGF-1498 TT genotypes predicted protection from antiangiogenic agents in combination with chemotherapy or radio- grade 3/4 hypertension in the combination-treatment arm.154 The therapy result in additive or synergistic effects. Several models have degree of hypertension can serve as a predictive biomarker of sur- been proposed to explain the mechanism responsible for this poten- vival in patients after bevacizumab or TKI treatment. Although tiation, keying in on the chemosensitizing effects of antiangiogenic an association between hypertension and anti-VEGF therapy has therapy.136 One hypothesis is that antiangiogenic therapy may nor- been described, the clinical implications of this association and malize the tumor vasculature, thus resulting in improved oxygen- the predictive value of hypertension remains to be validated pro- ation, better blood perfusion, and consequently, improved delivery spectively. A retrospective analysis of hypertension and efficacy of chemotherapeutic drugs.137 A second model suggests that che- outcomes was conducted in seven large phase III trials (n = 6,486 motherapy delivered at low doses and at close, regular intervals with patients) and, in six of seven studies, early treatment-related blood no extended drug-free break periods preferentially damages ECs in pressure increase was neither predictive of clinical benefit from the tumor neovasculature. 138,139 and suppresses circulating endo- bevacizumab nor prognostic for the course of the disease.155 How- thelial progenitor cells.140,141 This regimen, also called metronomic ever, one study (AVF2107g) showed early increased blood pressure 298 Cancer Therapeutics was associated with longer PFS and OS. Because genetics play a However, resistance to these drugs eventually occurs, implicating significant role in modifying the risk of hypertension,156 it remains the existence of additional pathways mediating resistance to anti- to be determined whether polymorphisms in the VEGF/VEGFR angiogenic therapies. Moreover, tumor cells bearing genetic al- pathway may function as potential biomarkers to predict the terations of the p53 gene may display a lower apoptosis rate under association between treatment-related hypertension and response hypoxic conditions, which might reduce their reliance on vascular to anti-VEGF therapy, as previously implicated in the E2100 supply and, therefore, their responsiveness to antiangiogenic ther- trial.154 Other biomarkers of response include elevated VEGF and apy.159 The selection and overgrowth of tumor-variant cells that are placental growth factor levels,148,152 whereas biomarkers of resis- hypoxia resistant and, thus, less dependent159 on angiogenesis and tance, including circulating basic fibroblast growth factor, stromal vasculature remodeling, resulting in vessel stabilization,160 could cell-derived factor 1α, and viable circulating endothelial cells, in- also explain the resistance to antiangiogenic drugs. Other possible creased when tumors escaped treatment.157 A first prospective bio- mechanisms for acquired resistance include tumor vessels becom- marker study (MERiDiAN) in metastatic breast cancer is currently ing less sensitive to antiangiogenic agents, tumor regrowth via re- underway to evaluate the impact of bevacizumab in patients strati- bound revascularization, and vessel cooption.161–166 Perhaps one of fied for plasma short VEGF-A isoforms. If validated, these find- the most intriguing findings is that, although ECs are assumed to ings could help identify which subgroup of patients should receive be genetically stable, they may under some circumstances harbor antiangiogenic therapy and could lead the way to possible future genetic abnormalities and thus acquire resistance as well.167,168 tailoring of individualized antiangiogenic therapy. Recent studies report that VEGF-targeted therapies not only in- duce primary tumor shrinkage and inhibit tumor progression, but can also initiate mechanisms that increase malignancy to promote RESISTANCE TO ANTIANGIOGENIC tumor invasiveness and metastasis.122,123,169 These mechanisms of THERAPY resistance to antiangiogenic therapy involve tumor- and host-me- diated pathways and may allow for differential efficacy in different Despite a decade of trials with angiogenesis inhibitors, clinical ex- stages of disease progression.163 Specifically, antiangiogenic drug– perience reveals that VEGF-targeted therapy often prolongs the resistance mechanisms involve pathways mediated by the tumor, survival of cancer patients by only months because tumors elicit whether intrinsic or acquired in response to therapy or by the host, evasive resistance.145,158 Resistance to VEGF inhibitors may be which is either responding directly to therapy or indirectly to tu- observed in late-stage tumors when tumors regrow during treat- moral cues. Taken together, antiangiogenic therapy can enhance ment after an initial period of growth suppression from these an- tumor invasiveness and metastasis to facilitate and/or accelerate tiangiogenic agents. This resistance involves the reactivation of disease in microscopic tumors and, hence, reduce OS benefit. tumor angiogenesis and increased expression of other proangio- Understanding the mechanisms of resistance, whether intrinsic or genic factors. As the disease progresses, it is possible that redun- acquired, after exposure to antiangiogenic drug treatment is essen- dant pathways might be implicated, with VEGF being replaced tial for developing strategies that will allow for optimal exploitation by other angiogenic pathways, warranting the addition of a second of VEGF inhibitors. It is equally important to identify biomarkers angiogenesis inhibitor that would target these secondary growth of drug resistance and factors mediating this resistance because the factors and/or their activated receptor pathways, or the use of a development of reliable biomarkers can be invaluable to monitor multitargeted TKI antiangiogenic drug (e.g., sunitinib, sorafenib). the development of evasive resistance to angiogenesis inhibitors.

SELECTED REFERENCES

The full reference list can be accessed at lwwhealthlibrary.com/oncology. 33. Kuhnert F, Kirshner JR, Thurston G. Dll4-Notch signaling as a therapeutic target in tumor angiogenesis. Vasc Cell 2011;3:20. 1. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell 49. Folkman J. Angiogenesis: an organizing principle for drug discovery? Nat Rev 2011;144:646–674. Drug Discov 2007;6:273–286. 5. Folkman J. Tumor angiogenesis: therapeutic implications. N Engl J Med 60. Ferrara N, Gerber HP, LeCouter J. The biology of VEGF and its receptors. 1971;285:1182–1186. Nat Med 2003;9:669–676. 7. Hanahan D, Folkman J. Patterns and emerging mechanisms of the angio- 61. Lindahl P, Johansson BR, Leveen P, et al. Pericyte loss and microaneurysm genic switch during tumorigenesis. Cell 1996;86:353–364. formation in PDGF-B-deficient mice. Science 1997;277:242–245. 9. Folkman J, Kalluri R. Cancer without disease. Nature 2004;427:787. 69. Desgrosellier JS, Cheresh DA. Integrins in cancer: biological implications 11. Holmgren L, O’Reilly MS, Folkman J. Dormancy of micrometastases: bal- and therapeutic opportunities. Nat Rev Cancer 2010;10:9–22. anced proliferation and apoptosis in the presence of angiogenesis suppres- 74. Ellis LM, Hicklin DJ. VEGF-targeted therapy: mechanisms of anti-tumour sion. Nat Med 1995;1:149–153. activity. Nat Rev Cancer 2008;8:579–591. 12. Naumov GN, Bender E, Zurakowski D, et al. A model of human tumor 75. Hurwitz H, Fehrenbacher L, Novotny W, et al. Bevacizumab plus irinotecan, dormancy: an angiogenic switch from the nonangiogenic phenotype. J Natl fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med Cancer Inst 2006;98:316–325. 2004;350:2335–2342. 13. Udagawa T, Fernandez A, Achilles EG, et al. Persistence of microscopic hu- 79. Sandler A, Gray R, Perry MC, et al. Paclitaxel-carboplatin alone or with beva- man cancers in mice: alterations in the angiogenic balance accompanies loss cizumab for non-small-cell lung cancer. N Engl J Med 2006;355:2542–2550. of tumor dormancy. Faseb J 2002;16:1361–1370. 90. Chen HX, Cleck JN. Adverse effects of anticancer agents that target the 14. Holash J, Maisonpierre PC, Compton D, et al. Vessel cooption, regression, and VEGF pathway. Nat Rev Clin Oncol 2009;6:465–477. growth in tumors mediated by angiopoietins and VEGF. Science 1999;284: 91. Ranpura V, Hapani S, Wu S. Treatment-related mortality with bevacizumab 1994–1998. in cancer patients: a meta-analysis. JAMA 2011;305:487–494. 16. Carmeliet P, Dor Y, Herbert JM, et al. Role of HIF-1alpha in hypoxia- 94. Gaya A, Tse V. A preclinical and clinical review of aflibercept for the man- mediated apoptosis, cell proliferation and tumour angiogenesis. Nature agement of cancer. Cancer Treat Rev 2012;38:484–493. 1998;394:485–490. 97. Escudier B, Eisen T, Stadler WM, et al. Sorafenib in advanced clear-cell 19. Rak J, Yu JL. Oncogenes and tumor angiogenesis: the question of vascular renal-cell carcinoma. N Engl J Med 2007;356:125–134. “supply” and vascular “demand”. Semin Cancer Biol 2004;14:93–104. 99. Llovet JM, Ricci S, Mazzaferro V, et al. Sorafenib in advanced hepatocellular 20. Bottos A, Bardelli A. Oncogenes and angiogenesis: a way to personalize anti- carcinoma. N Engl J Med 2008;359:378–390. angiogenic therapy? Cell Mol Life Sci 2013;70:4131–4140. 101. Veronese ML, Mosenkis A, Flaherty KT, et al. Mechanisms of hypertension 24. Ribatti D. Endogenous inhibitors of angiogenesis: a historical review. Leuk associated with BAY 43-9006. J Clin Oncol 2006;24:1363–1369. Res 2009;33:638–644. 105. Motzer RJ, Hutson TE, Tomczak P, et al. Sunitinib versus interferon alfa in 31. Noguera-Troise I, Daly C, Papadopoulos NJ, et al. Blockade of Dll4 in- metastatic renal-cell carcinoma. N Engl J Med 2007;356:115–124. hibits tumour growth by promoting non-productive angiogenesis. Nature 107. Raymond E, Dahan L, Raoul JL, et al. Sunitinib malate for the treatment of 2006;444:1032–1037. pancreatic neuroendocrine tumors. N Engl J Med 2011;364:501–513. 32. Ridgway J, Zhang G, Wu Y, et al. Inhibition of Dll4 signalling inhibits tu- 110. Motzer RJ, Hutson TE, Cella D, et al. Pazopanib versus sunitinib in meta- mour growth by deregulating angiogenesis. Nature 2006;444:1083–1087. static renal-cell carcinoma. N Engl J Med 2013;369:722–731.

tahir99 - UnitedVRG Chapter 28 Antiangiogenesis Agents 299

112. Schutz FA, Choueiri TK, Sternberg CN. Pazopanib: Clinical development 147. Ebos JM, Kerbel RS. Antiangiogenic therapy: impact on invasion, disease of a potent anti-angiogenic drug. Crit Rev Oncol Hematol 2011;77:163–171. progression, and metastasis. Nat Rev Clin Oncol 2011;8:210–221. 120. Graveel CR, Tolbert D, Vande Woude GF. MET: a critical player in tumori- 148. Jain RK, Duda DG, Willett CG, et al. Biomarkers of response and resistance genesis and therapeutic target. Cold Spring Harb Perspect Biol 2013;5. to antiangiogenic therapy. Nat Rev Clin Oncol 2009;6:327–338. 123. Paez-Ribes M, Allen E, Hudock J, et al. Antiangiogenic therapy elicits malig- 152. Lambrechts D, Lenz HJ, de Haas S, et al. Markers of response for the antian- nant progression of tumors to increased local invasion and distant metastasis. giogenic agent bevacizumab. J Clin Oncol 2013;31:1219–1230. Cancer Cell 2009;15:220–231. 155. Hurwitz HI, Douglas PS, Middleton JP, et al. Analysis of early hypertension 126. Strumberg D, Schultheis B. Regorafenib for cancer. Expert Opin Investig and clinical outcome with bevacizumab: results from seven phase III studies. Drugs 2012;21:879–889. Oncologist 2013;18:273–280. 130. Hudes G, Carducci M, Tomczak P, et al. Temsirolimus, interferon alfa, or 158. Sennino B, McDonald DM. Controlling escape from angiogenesis inhibi- both for advanced renal-cell carcinoma. N Engl J Med 2007;356:2271–2281. tors. Nat Rev Cancer 2012;12:699–709. 133. Lebwohl D, Anak O, Sahmoud T, et al. Development of everolimus, a 161. Bergers G, Hanahan D. Modes of resistance to anti-angiogenic therapy. Nat novel oral mTOR inhibitor, across a spectrum of diseases. Ann N Y Acad Sci Rev Cancer 2008;8:592–603. 2013;1291:14–32. 162. Crawford Y, Ferrara N. Tumor and stromal pathways mediating refractori- 136. Kerbel RS. Antiangiogenic therapy: a universal chemosensitization strategy ness/resistance to anti-angiogenic therapies. Trends Pharmacol Sci 2009;30: for cancer? Science 2006;312:1171–1175. 624–630. 137. Jain RK. Normalization of tumor vasculature: an emerging concept in anti- 163. Ebos JM, Lee CR, Kerbel RS. Tumor and host-mediated pathways of resis- angiogenic therapy. Science 2005;307:58–62. tance and disease progression in response to antiangiogenic therapy. Clin 142. Kerbel RS, Kamen BA. The anti-angiogenic basis of metronomic chemo- Cancer Res 2009;15:5020–5025. therapy. Nat Rev Cancer 2004;4:423–436. 168. Streubel B, Chott A, Huber D, et al. Lymphoma-specific genetic aberra- 143. Hudis CA. Clinical implications of antiangiogenic therapies. Oncology (Wil- tions in microvascular endothelial cells in B-cell lymphomas. N Engl J Med liston Park) 2005;19:26–31. 2004;351:250–259. 145. Kerbel RS. Tumor angiogenesis. N Engl J Med 2008;358:2039–2049. 169. Loges S, Mazzone M, Hohensinner P, et al. Silencing or fueling metasta- 146. Mountzios G, Pentheroudakis G, Carmeliet P. Bevacizumab and microme- sis with VEGF inhibitors: antiangiogenesis revisited. Cancer Cell 2009;15: tastases: Revisiting the preclinical and clinical rollercoaster. Pharmacol Ther 167–170. 2014;141:117–124. CANCER THERAPEUTICS 29 Monoclonal Antibodies

Hossein Borghaei, Matthew K. Robinson, Gregory P. Adams, and Louis M. Weiner

INTRODUCTION IgG molecules to their cognate antigens and are composed of an intact light chain and half of a heavy chain. Each chain in the Fab Antibody-based therapeutics are important components of the domain is further divided into variable and constant regions, with cancer therapeutic armamentarium. Early antibody therapy stud- the variable region containing hypervariable, or complementarity ies attempted to explicitly target cancers based on the structural determining regions (CDR) in which the antigen-contact residues and biologic properties that distinguish neoplastic cells from their reside. The light and heavy chain variable regions each contain normal counterparts. The immunogenicity and inefficient effec- three CDRs (CDR1, CDR2, and CDR3). All six CDRs form the tor functions of the first-generation murine monoclonal antibodies antigen-binding pocket and are collectively defined in immuno- (MAb) that were evaluated in clinical trials limited their effective- logic terms as the idiotype of the antibody. In the majority of cases, ness.1–3 Patients developed human antimouse antibody (HAMA) the variable heavy chain CDR3 plays a dominant role in binding.10 responses against the therapeutic agents that rapidly cleared it The different isotypes of immunoglobulins are defined by the from the body and limited the number of times the therapy could structure and function of their Fc domains. The Fc domain, com- be administered. The development of engineered chimeric, hu- posed of the CH2 and CH3 regions of the antibody’s heavy chains, is manized, and fully human MAbs has identified a number of im- the critical determinant of how an antibody mediates effector func- portant and useful applications for antibody-based cancer therapy. tions, transports across cellular barriers, and persists in circulation.7,11 Currently, the U.S. Food and Drug Administration (FDA) has ap- proved 14 MAbs and MAb-conjugates for the treatment of cancer (Table 29.1) and many more are under evaluation in late-stage MODIFIED ANTIBODY-BASED MOLECULES clinical trials.4 Antibodies provide an important means by which to exploit the immune system by specifically recognizing and di- Advances in antibody engineering and molecular biology have recting antitumor responses. facilitated the development of many novel antibody-based struc- Antibodies are produced by B cells and arise in response to ex- tures with unique physical and pharmacokinetic properties (see posures to a variety of structures, termed antigens, as a result of a Fig. 29.1). These include chimeric human-murine antibodies series of recombinations of V, D, and J germline genes. Immuno- with human-constant regions and murine-variable regions,12 hu- globulin-G (IgG) molecules are most commonly employed as the manized antibodies in which murine CDR sequences have been working backbones of current therapeutic monoclonal antibod- grafted into human IgG molecules, and entirely human antibod- ies, although various other isotypes of antibodies have specialized ies derived from human hybridomas and, more recently, from functions (e.g., IgA molecules play important roles in mucosal im- transgenic mice expressing human immunoglobulin genes.13 munity, IgE molecules are involved in anaphylaxis). The advent of An accepted naming scheme based on “stems” was developed hybridoma technology by Kohler and Milstein5 made it possible to by the World Health Organization’s International Nonpropri- produce large quantities of antibodies with high purity and mono- etary Names (INN) for pharmaceuticals and is employed in the specificity for a single binding region (epitope) on an antigen. United States (Table 29.2). Engineering has also facilitated the The mechanisms that antibody-based therapeutics employ development of antibody-based fragments. In addition to the clas- ′ to elicit antitumor effects include focusing components of the sic, enzymatically derived Fab and F(ab )2 molecules, a plethora patient’s immune system to attack tumor cells6,7 and methods to of promising IgG-derivatives have been developed that retain alter signal transduction pathways that drive tumor progression.8,9 antigen-binding properties of intact antibodies (see Fig. 29.1; for Antibody-based conjugates employ the targeting specificity of an- review see Robinson et al.14). The basic building block for these tibodies to deliver toxic compounds, such as chemotherapeutics, molecules is the 25 kDa, monovalent single-chain Fv (scFv) that specifically to the tumor sites. is comprised of the variable domains (VH and VL) of an antibody fused together with a short peptide linker. Novel, bispecific anti- body-based structures can facilitate binding to two tumor antigens IMMUNOGLOBULIN STRUCTURE or bridge tumor cells with immune effector cells to focus anti- body-dependent cell-mediated cytotoxicity (ADCC) or killing by T cells. An example of the former is MM-111, a bispecific gene- Structural and Functional Domains fused molecule composed of an anti-HER2 scFv connected to an anti-HER3 scFv via a modified form of human serum albumin.15 An IgG molecule is typically divided into three domains consisting Examples of the latter mechanism include small scFv-based bi- of two identical antigen-binding (Fab) domains connected to an specific T-cell engagers (BiTE) such as the anti-CD3/anti-CD19 effector or Fc domain by a flexible hinge sequence. Figure 29.1 molecule blinatumomab16 and larger MAb-based antibodies such shows the structure of an IgG molecule. IgG antibodies are com- as catumaxomab, a rat/mouse anti-CD3/EpCAM bispecific MAb prised of two identical light chains and two identical heavy chains, produced via quadroma technology.17 Both classes of bispecifics with the chains joined by disulfide bonds, resulting in a bilaterally endow selectivity and targeting properties that are not obtainable symmetrical complex. The Fab domains mediate the binding of with natural antibody formats.

300

tahir99 - UnitedVRG Chapter 29 Monoclonal Antibodies 301

TABLE 29.1 FDA Approved Antibodies for the Treatment of Cancer

Isotype Generic Name (Trade Name) Origin (Conjugate) Indication Target Initial Approval Unconjugated MAbs Rituximab (Rituxan) Chimeric IgG1 NHL CD20 1997 Trastuzumab (Herceptin) Humanized IgG1 BrCa HER2 1998 Alemtuzumab (Campath-1H) Humanized IgG1 CLL CD52 2001 Cetuximab (Erbitux) Chimeric IgG1 CRC, SCCHN EGFR 2004 Bevacizumab (Avastin) Humanized IgG1 CRC, NSCLC, RCC, GBM VEGF 2004 Panitumumab (Vectibix) Human IgG2 CRC EGFR 2006 (XenoMouse) Ofatumumab (Arzerra) Human IgG1 CLL CD20 2009 (XenoMouse) Denosumab (Prolia/Xgeva) Human IgG2 Metastasis-related RANKL 2010 SREs, ADT/AI-associated osteoporosis, GCT Pertuzumab (Perjeta) Humanized IgG1 BrCa HER2 2012 Immunoconjugates Gemtuzumab ozogamicin Humanized IgG4 AML CD33 2000a (Mylotarg) (calicheamicin) Ibritumomab tiuxetan (Zevalin) Murine IgG1 (90Y) NHL CD20 2002 Tositumomab (Bexxar) Murine IgG2A (131I) NHL CD20 2003 Brentuximab vedotin (Adcetris) Chimeric IgG1 (MMAE) HL, sALCL CD30 2011 CANCER THERAPEUTICS Ado-trastuzumab emtansine Humanized IgG1 (DM1) BrCa HER2 2013 (Kadcyla) a Withdrawn from the US market in June 2010. NHL, non-Hodgkin lymphoma; BrCa, breast cancer; CLL, chronic lymphocytic leukemia; CRC, colorectal cancer; SCCHN, squamous cell carcinoma of head and neck; EGFR, epidermal growth factor receptor; NSCLC, non–small-cell lung cancer; RCC, renal cell carcinoma; GBM, glioblastoma multiforme; VEGF, vascular endothelial growth factor; SREs, skeletal-related events; ADT, androgen deprivation therapy; AI, aromatase inhibitor; GCT, giant cell tumor; RANKL, RANK ligand; AML, acute myelogenous leukemia; 90Y, yttrium-90; 131I, iodine-131; MMAE, Monomethyl auristatin E; HL, Hodgkin lymphoma; sALCL, systemic anaplastic large- cell lymphoma.

IgG

VH TABLE 29.2 Rules for Naming MAb for the Treatment of Cancer

C The International Nonproprietary Names (INN) for monoclonal 1 antibodies (MAbs) are composed of “stems” that indicate VL their origin, specificity, and modifications. The names include a random prefix to provide distinction from other names, a substem indicating the target specificity(-t[u]- for tumor), a substem indicating the species of origin (see the following) and a suffix (-mab), which indicates the presence of an immunoglobulin -S-S- C -S-S- L variable domain. Substem Indication of the Species on Which the Carbohydrate Immunoglobulin Sequence Is Based CH2 -o- mouse -xi- chimeric CH 3 -zu- humanized -xizu- chimeric/humanized Figure 29.1 Structure of an IgG. C, constant; V, variable; H, heavy chain; -u- human L, light chain. 302 Cancer Therapeutics

FACTORS REGULATING ANTIBODY-BASED required to add the bisecting GlcNAc residues.33 Defucosylation TUMOR TARGETING of antibody Fc domains is also associated with enhanced ADCC, and in a recently completed multicenter phase II trial of a defu- cosylated anti-CC chemokine receptor 4 (CCR4), MAb was as- Antibody Size sociated with meaningful antitumor activity, including complete responses and enhanced progression-free survival (PFS).35 Nonuniform distribution of systemically administered antibody is generally observed in biopsied specimens of solid tumors. Het- erogeneous tumor blood supply limits uniform antibody deliv- UNCONJUGATED ANTIBODIES ery to tumors, and elevated interstitial pressures in the center of tumors oppose inward diffusion.18 This high interstitial pressure The majority of monoclonal antibodies approved for clinical use slows the diffusion of molecules from their vascular extravasation display intrinsic antitumor effects that are mediated by one or site in a size-dependent manner.19,20 The relatively large transport more of the following mechanisms. distances in the tumor interstitium also substantially increase the 21 time required for large IgG macromolecules to reach target cells. Cell-Mediated Cytotoxicity

Tumor Antigens As components of the immune system, effector cells such as natu- ral killer (NK) cells and monocytes/macrophages represent natural Access to the target antigen is undoubtedly a critical determinant lines of defense against oncologically transformed cells. These ef- of therapeutic effect of antibody-based applications. Such access fector cells express Fcγ receptors (FcγR) on their cell surfaces, is regulated by the heterogeneity of antigen expression by tumor which interact with the Fc domain of IgG molecules. This family cells. Shed antigen in the serum, tumor microenvironment, or is comprised of three classes (type I, II, and III) that are further di- both may saturate the antibody’s binding sites and prevent bind- vided into subclasses (IIa/IIb and IIIa/IIIb).36 Recognition of trans- ing to the cell surface. Alternatively, a rapid internalization of an formed cells by immune effector cells leads to cell-mediated killing antibody/antigen complex, although critical for antibody–drug through processes such as ADCC and phagocytosis, as shown in conjugates (ADC), may deplete the quantity of cell surface MAb Figure 29.2, and can be mediated by FcγRI (CD64), a high affinity capable of initiating ADCC or cytotoxic signal transduction events. receptor capable of binding to monomeric IgG, or FcγRII (CD32) Finally, target antigens are normally tumor associated rather than and FcγRIII (CD16), which are low affinity receptors that pref- tumor specific. Tumor-specific antigens are both highly desirable erentially bind multimeric complexes of IgG. Signaling through and rare. Typically, such antigens arise as a result of unique tumor- type I, IIa, and IIIa receptors results in the activation of effector based genetic recombinations, such as clonal immunoglobulin cells due to associated immunoreceptor tyrosine-based activation 22 idiotypes expressed on the surface of B-cell lymphomas. motifs (ITAM), whereas the engagement of type IIb receptors in- Antibody affinity for its target antigen has complex effects on hibits cell activation through associated immunoreceptor tyrosine- tumor targeting. The binding-site barrier hypothesis postulates that based inhibitory motifs (ITIM).36 Clinical results support the idea antibodies with extremely high affinity for target antigen would that ADCC can play a role in the efficacy of antibody-based thera- bind irreversibly to the first antigen encountered upon entering pies. Naturally occurring polymorphisms in FcγRs alter their affin- the tumor, which would limit the diffusion of the antibody into the ity for human IgG1 and have been linked to clinical response.37,38 tumor and accumulate instead in regions surrounding the tumor A polymorphism in the FCGR3A gene results in either a valine 23,24 vasculature. Similarly, in tumor spheroids, the in vitro penetra- or phenylalanine at position 158 of FcγRIIIa. Human IgG1 binds tion of engineered antibodies is primarily limited by internaliza- more strongly to FcγRIIIa-158V than FcγRIIIa-158F, and likewise 25 tion and degradation. The valence of an antibody molecule can to NK cells from individuals that are either homozygous for 158F increase the functional affinity of the antibody through an avidity or heterozygous for this polymorphism.39 The FcγRIIIa-158v was a 26–28 effect. predictor of early response and was associated with improved PFS. Half-Life/Clearance Rate

The concentration of intact IgG in mammalian serum is main- Tumor tained at constant levels with half-lives of IgGs measured in Pore days. This homeostasis is regulated in part by the major histo- Cell formation compatibility complex (MHC)-class I–related Fc receptor, FcRn Tumor antigen (n = neonatal), a saturable, pH-dependent salvage mechanism Natural that regulates quality and quantity of IgG in serum. This mecha- Antibody killer cell cytotoxicity nism can be exploited via mutations in the Fc portion of an IgG to γ modulate IgGs pharmacokinetics.29,30 Indeed, multiple strategies Fc receptor have been developed to increase the serum persistence of anti- body-based fragments and other classes of protein therapeutics.14,31 Effector Cell Glycosylation

IgGs undergo N-linked glycosylation at the conserved Asn residue Macrophage at position 297 within the CH2 domain of the constant region. Gly- cosylation status of the residue has long been known to impact or neutrophil phagocytosis the ability of IgGs to bind effector ligands such as FcγR and C1q, which, in turn, affects their ability to participate in Fc-mediated Figure 29.2 Antibody-dependent cellular cytotoxicity. The antibody functions such as ADCC and complement-dependent cytotoxicity engages the tumor antigen and the Fc domain binds to cellular Fc 32–34 (CDC). The glycosylation of MAbs can be altered to increase receptors to bridge effector and target cells. This bridging induces effector ADCC by producing them in a cell line engineered to express cell activation, resulting in natural killer cell cytotoxicity or phagocytosis β(1,4)-N-acetylglucosaminyltransferase III (GnTIII), the enzyme by neutrophils, monocytes, or macrophages.

tahir99 - UnitedVRG Chapter 29 Monoclonal Antibodies 303

A second polymorphism, FcγRIIa-131H/R, did not predict early prevents the receptor from assuming the extended conformation 57 response but was an independent predictor of time to progression required for dimerization. Pertuzumab binds to the dimeriza- (TTP).38 Taken together, these data suggest that modulating the tion domain of HER-2, thereby sterically inhibiting subsequent affinity of MAbs for FcγRIIIa, FcγRIIa, or both may increase the receptor heterodimerization with other ligand-bound family mem- 58 efficacy of therapeutic MAbs. bers. Alternatively, signaling through growth factor receptors can Each class of FcγR exhibits a characteristic specificity for IgG be indirectly modified by MAbs that bind to activating ligands, as subclasses.40 Many groups have focused on modifying the Fc do- is seen with the anti–vascular endothelial growth factor (VEGF) 59 main of IgGs to optimize the engagement of subclasses of FcγR MAb, bevacizumab. and the induction of ADCC, based on the findings of Shields et 29 al., who performed a series of mutagenesis experiments to map IMMUNOCONJUGATES the residues required for IgG1-FcγR interaction. Antibodies such as ocrelizumab, a humanized version of rituximab, have increased MAbs that are not capable of directly eliciting antitumor effects, binding to low affinity FcγRIIIa variants and are now in clinical either by altering signal transduction or directing immune system trials. cells, can still be effective against tumors by delivering cytotoxic An alternative to modifying the Fc region of MAbs is to create payloads. MAbs have been employed to deliver a wide variety of bispecific antibodies (bsAbs) that recognize both a tumor-associ- agents, including chemotherapy, toxins, radioisotopes, and cy- ated antigen and a trigger antigen present on the surface of an im- tokines (for review see Adams and Weiner60). In theory, the ap- mune effector cell.43 Simultaneous engagement of both antigens propriate combination of toxic agents and MAbs could lead to a can redirect the cytotoxic potential of the effector cell against the synergistic effect. For example, delivery of a therapeutic radioiso- tumor.41–43 Such antibodies are capable of eliciting effector func- tope by a MAb would be significantly enhanced if, by binding to tion against tumor cell lines in vitro and in animal models. Two its target antigen, the MAb also activated a signaling event that HER-2 directed bispecific antibodies, 2B1 and MDX-H210, have increased the target cell’s sensitivity to ionizing radiation. been tested in phase I clinical trials.44,45 Catalytic toxins derived from plants catalytic toxins derived Bispecific antibodies have a number of distinctive proper- from plants (e.g., ricin) and microorganisms (e.g., Pseudomonas) ties, including flexible choices of cytotoxic trigger molecules,46 represent two classes of cytotoxic agent that have been investigated recruitment of effector function in the presence of excess IgG,42 for their utility in immunoconjugate strategies.61 Although there and custom tailoring of the affinity of the bsAb to match effector are promising preclinical studies,62 few successful clinical trials cell characteristics. These advantages have been facilitated by im- have been reported using this approach. In a phase I clinical trial proved methods of bsAb production.47 BiTE antibodies represent in hairy cell leukemia patients who were resistant to , a novel class of bispecific, single-chain Fv antibodies.48 Promising 11 of 16 patients exhibited complete remissions with minimal side results have been seen in early phase clinical trials with at least

effects with an anti-CD22 immunotoxin with a truncated form of CANCER THERAPEUTICS two BiTE antibodies, one of which, blinatumomab, targets CD19/ Pseudomonas exotoxin.63 Clinical trials with other immunotoxins CD3.49 Promising phase I results have also been reported in an have been associated with unacceptable neurotoxicity64 and life- interim analysis of an anti-EpCAM/anti-CD3 MT110 BiTE in the threatening vascular leak syndrome.65 setting of advanced lung and gastrointestinal tumors.50 Immunocytokine fusions have also been investigated as an ap- proach to direct the patient’s immune response to his or her own Complement-Dependent Cytotoxicity tumor.66 A number of cytokines have been incorporated into an- tibody-based constructs, including interleukin-2 (IL-2),67,68 inter- In addition to cell-mediated killing (see previous), MAbs can re- feron γ (IFN-γ),69 tumor necrosis factor α (TFN-α),69 VEGF,70 cruit the complement cascade to kill cells via CDC. Although and IL-12.71 IgM is the most effective isotype for complement activation, it is not widely used in clinical oncology. Similar to ADCC, the human IgG subclass used to construct a therapeutic MAb dic- Antibody–Drug Conjugates tates its ability to elicit CDC; IgG1 is extremely efficient at fixing complement, in contrast to IgG2 and IgG4.51 Antibodies activate The first ADC, gemtuzumab ozogamicin (Mylotarg), was ap- complement through the classical pathway, by engaging multiple proved by the FDA in 2000 for the treatment of patients with C1q to trigger activation of a cascade of serum proteases, which relapsed CD33-positive acute myeloid leukemia, but was vol- kill the antibody-bound cells.52,53 The anti-CD20 MAb rituximab untarily withdrawn from the US market by its manufacturer in has been found to depend in part on CDC for its in vivo efficacy.54 2010 after a confirmatory phase III trial (SWOG S0106) recom- Antibody engineering approaches have identified residues in the mended, based on results of a planned interim analysis, that My- lotarg randomizations be terminated due to a lack of efficacy in CH2 domain of the Fc region that either suppress or enhance the 72 55 the presence of enhanced toxicity. Although two additional ran- ability of rituximab to bind C1q and activate CDC. The ability to 73,74 manipulate complement fixation through engineering approaches domized trials suggested that some patient populations may warrants in vivo testing to determine the impact of these changes benefit from Mylotarg therapy, the drug remains off the market on the efficacy and toxicity of MAbs. in the United States. The majority of ADCs under development employ potent cy- totoxic agents that block the polymerization of tubulin (e.g., au- ALTERING SIGNAL TRANSDUCTION ristatins or maytansines) or damage DNA (e.g., calicheamicins or pyrrolobenzodiazepines) by employing a variety of linkers and Growth factor receptors represent a well-established class of tar- conjugation strategies.75 gets for therapeutic intervention. Normal signaling through these A variety of ADCs specific for a wide range of oncology targets receptors often leads to mitogenic and prosurvival responses. Un- are currently in clinical evaluation, with the majority of the more regulated signaling, as seen in a number of common cancers due advanced agents being tested in the setting of diffuse malignan- to receptor overexpression, promotes tumor cell growth and insen- cies.76 The majority of these employ auristatins or maytansines as sitivity to chemotherapeutic agents. Clinically relevant MAbs can their payloads. Early observations suggest that cumulative, dose-re- modulate signaling through their target receptors to normalize lated peripheral sensory neuropathy can result when auristatins are cell growth rates and sensitize tumor cells to cytotoxic agents. The conjugated to an antibody via a cleavable linker, and dose-limiting binding of cetuximab or panitumumab to the epidermal growth thrombocytopenia can result when auristatins and maytansinoids factor receptor (EGFR) physically blocks ligand binding56 and are conjugated to the antibody via an uncleavable linker.76,77 304 Cancer Therapeutics

Two ADCs are now approved for use in clinical practice. Ado- Pertuzumab trastuzumab emtansine (T-DM1, Kadcyla), an ADC composed of the anti-HER2 MAb trastuzumab linked to DM1,78 is now ap- Pertuzumab (Perjeta) is a humanized IgG1 MAb that binds to proved for the treatment of patients with refractory HER2/neu ex- domain II of HER2 and blocks ligand-dependent dimerization of pressing breast cancers. The other, brentuximab vedotin (SGN-35, HER2 with other members of the EGFR family.96 Pertuzumab, in Adcetris), is an ADC consisting of the anti-CD30 chimeric MAb combination with trastuzumab and docetaxel, is approved for use cAC10 that is linked to three to five molecules of the microtubule- as first-line therapy in HER2-positive metastatic breast cancer pa- disrupting agent Monomethyl auristatin E. At this point, this drug tients. Use of the combination is also approved for the treatment of is approved for use in patients with recurrent systemic anaplas- HER2-positive, locally advanced, inflammatory, or high-risk early tic large cell lymphoma. The clinical data associated with both breast cancer (>2 cm node negative or node positive) in the neo- of these ADCs will be discussed in subsequent sections of this adjuvant setting. chapter. FDA-approval of pertuzumab was based on results of a phase Antibodies also can be used to target liposome-encapsulated III trial (CLEOPATRA) of 808 patients with locally recurrent, drugs79 and other cytotoxic agents, such as antisense RNA80 or ra- unresectable, or metastatic breast cancer randomized to receive dionuclides to tumors. trastuzumab plus docetaxel with or without the addition of per- tuzumab. Inclusion of pertuzumab increased the independently Radioimmunoconjugates assessed PFS by 6.1 months from 12.4 to 18.5 (hazard ratio [HR], 0.62 (95% confidence interval [CI], 0.51, 0.75), p< 0.0001], with 97 Two anti-CD20 radioimmunoconjugates have been FDA ap- a trend toward improved overall survival that reached statistical = 98 proved for radioimmunotherapy (RIT) of non-Hodgkin lymphoma significance (p 0.0008) after an additional year of follow-up. (NHL). Ibritumomab (Zevalin) and tositumomab (Bexxar) are mu- The addition of pertuzumab did increase rates of grade 3 adverse rine MAbs labeled with yttrium-90 (90Y) and iodine-131 (131I), re- events (AE), but it did not adversely affect cardiac function. Accel- spectively. Both are associated with impressive clinical efficacy.81,82 erated approval was granted for use of pertuzumab in combination Although these radioimmunoconjugates are effective therapeutics, with trastuzumab and docetaxel for the neoadjuvant treatment of cumbersome logistics surrounding their administration have sig- high-risk early-stage breast cancer. This approval was based on re- nificantly limited their use. Despite significant preclinical evi- sults from a four-arm, open-label phase II study of 417 patients dence supporting the use of RIT for solid malignancies, clinical randomized to receive trastuzumab plus docetaxel, pertuzumab results have not demonstrated consistent antitumor activity.60 plus docetaxel, pertuzumab plus trastuzumab, or the triple combi- nation. The triple combination improved the pathologic complete response (pCR) rate by 17.8% over the trastuzumab plus docetaxel ANTIBODIES APPROVED FOR USE IN arm (39.3% versus 21.5%) in the pertuzumab arm.99 Follow-up SOLID TUMORS studies to confirm a correlation between pCR and long-term clini- cal benefit are ongoing. Trastuzumab Cetuximab Trastuzumab (Herceptin) is a humanized IgG183 that targets do- main IV of the HER2/ErbB2 member of the EGFR/ErbB fam- Cetuximab (Erbitux) targets the EGFR. This chimeric IgG1 binds ily of receptor tyrosine kinases. Gene amplification as judged by to domain III of the EGFR, with roughly a tenfold higher affinity fluorescence in situ hybridization (FISH) with concomitant over- than either EGF or transforming growth factor α (TGF-α) ligands expression of HER2 protein measured by immunohistochemistry and thereby inhibits ligand-induced activation of this tyrosine (IHC) is seen in approximately 25% of breast cancers.84,85 HER2 kinase receptor. Cetuximab may also function to downregulate amplification and overexpression is now recognized to also be a EGFR-dependent signaling by stimulating EGFR internaliza- critical driver in a subset (7% to 34%) of gastric cancers.86 Trastu- tion.100 Cetuximab is approved for the treatment of colorectal can- zumab inhibits tumor cell growth by binding to HER2 and block- cer (CRC) and, more recently, for the treatment of squamous cell ing the unregulated HER2 signaling that is associated with its high cancer of the head and neck (SCCHN). level overexpression. The efficacy and safety of cetuximab against CRC was dem- Trastuzumab became the first FDA-approved monoclonal an- onstrated alone and in combination with irinotecan in a phase II, tibody for the treatment of solid tumors based on a series of stud- multicenter, randomized, and controlled trial of 329 patients.101 ies carried out in the setting of HER2-positive metastatic breast The combination of irinotecan plus cetuximab increased both cancer.87,88 A subsequent phase III trial investigating trastuzumab the overall response and the median duration of response as com- in combination with cytotoxic chemotherapy demonstrated an pared to cetuximab alone. Additionally, patients with irinotecan improved response rate compared to chemotherapy alone, from refractory disease responded to treatment with the combination 25.0% to 57.3% with a taxane regimen.89 regimen. Recent studies in patients with colorectal cancers have Trastuzumab is also approved for use in the adjuvant setting indicated that patients with KRAS mutations in codon 12 or 13 based on an approximately 50% reduction in recurrence after should not receive anti-EGFR therapy.101,102 1 year in multiple phase III trials.90–92 Myocardial dysfunction, An international, multicenter, phase III trial comparing de- seen with anthracycline therapy, was observed with increased fre- finitive radiotherapy to radiotherapy plus cetuximab in SCCHN quency in patients receiving antibody alone93 or with doxorubicin demonstrated that EGFR blockade with radiotherapy signifi- or epirubicin. cantly reduced the risk of locoregional failure by 32% and the risk Recognition of HER2 as a driver in a subset of gastric cancers of death by 26%. In advanced stage non–small-cell lung cancer led to an open-label, randomized, phase III trial (ToGA) that inves- (NSCLC) expressing EGFR, the combination of cetuximab and tigated the addition of trastuzumab to standard of care chemother- standard doublet chemotherapy (cisplatin plus vinorelbine) was apy94 and showed increased median overall survival with higher studied in a prospective randomized phase III trial.103 The addition levels of HER2 expression. A study by Gomez-Martin et al.95 in 99 of cetuximab was associated with a slight, but statistically signifi- patients with metastatic gastric cancer being treated with first-line cant, benefit in overall survival over chemotherapy alone (median trastuzumab plus chemotherapy identified a mean HER2/CEP17 overall survival 10.1 versus 11.3 months). A similar study using ratio of 4.7 to be an optimal cut-off to discriminate between trastu- the carboplatin plus paclitaxel backbone in combination with zumab-sensitive and refractory patients. cetuximab did not meet its primary endpoint of improved PFS,

tahir99 - UnitedVRG Chapter 29 Monoclonal Antibodies 305 although cetuximab-treated patients exhibited higher objective re- improvement in overall survival between the two groups and the sponse rates.104 Therefore, the benefit of adding cetuximab to stan- rate of invasive disease-free survival was also not significantly differ- dard chemotherapy for patients with advanced NSCLC is unclear. ent between the treatment groups. Bevacizumab is also approved for the management of recurrent 117 Panitumumab glioblastomas based on results of phase II studies.

Panitumumab (Vectibix) is a fully human IgG2 monoclonal an- Ado-Trastuzumab Emtansine tibody that binds to EGFR. Similar to cetuximab, panitumumab inhibits EGFR activation by blocking the binding of EGF and Ado-trastuzumab emtansine (T-DM1, Kadcyla) is an ADC com- TGF-α. However, it does so by binding to EGFR with a higher posed of the anti-HER2 MAb trastuzumab linked to DM1, a highly affinity than cetuximab (5 × 10-11 M versus 1 × 10-10 M). As previ- potent derivative of maytansine, through a stable thioether linker.78 ously mentioned, the IgG2 class of antibodies does not induce ac- Based on two single-agent phase II trials of T-DM1118,119 that dem- tivation of the immune system cell via the Fc-receptor mechanism, onstrated single-agent activity in the setting of metastatic breast can- so panitumumab’s primary action appears to be interference with cer, two separate phase III studies were conducted. The 991 patient EGFR–ligand interactions. EMILIA trial demonstrated that T-DM1 significantly prolongs both A phase III trial of 463 patients with metastatic colorectal can- PFS and overall survival as compared to a regimen of lapatinib plus cer compared panitumumab plus best supportive care (BSC) to capecitabine when used in the setting of metastatic breast cancer BSC alone.105 A partial-response rate of 8% and a stable-disease that had progressed after treatment with trastuzumab plus a taxane.120 rate of 28% were reported for the panitumumab arm compared Grade 3 and worse AEs were lower in the T-DM1 arm (200, 40.8%) with a 10% stable-disease rate in the best supportive care arm of as compared to the lapatinib plus capecitabine arm (278, 57%). Re- the study. As with cetuximab, patients with metastatic colorectal sults are still awaited from the ongoing MARIANNE trial that is as- cancers who have KRAS mutations in codons 12 or 13 are not sessing first-line efficacy and safety of T-DM1 alone and T-DM1 plus routinely offered therapy with panitumumab.106 pertuzumab versus trastuzumab plus taxane (NCT01120184).

Bevacizumab Denosumab

Bevacizumab (Avastin or rhuMAb VEGF) is a humanized mono- Denosumab (Xgeva) is a fully human IgG2 RANK ligand (RANKL) clonal antibody targeting VEGF. VEGF is a critical determinant neutralizing antibody. Denosumab is FDA-approved for use in of tumor angiogenesis, a process that is a necessary component of adults and skeletally mature adolescents who have either surgically tumor invasion, growth, and metastasis. VEGF expression by inva- unsalvageable giant cell tumors of the bone (GCTB) or where re- sive tumors has been shown to correlate with vascularity and cellu- section is anticipated to result in severe morbidity. Approval was CANCER THERAPEUTICS lar proliferation and is prognostic for several human cancers.107–109 based in part on two open-label, phase II trials examining subcu- Interestingly, the inhibition of VEGF signaling via bevacizumab taneous administration of 120 mg q4 week with additional load- treatment may normalize tumor vasculature, promoting a more ing doses on days 8 and 15 of the first cycle.121,122 Serious adverse effective delivery of chemotherapy agents.110 Bevacizumab is events were seen in 9% of patients (n = 25). Of 187 patients, 47 approved for use as a first-line therapy for metastatic colorectal (25%) exhibited partial objective responses based on modified Re- cancer and NSCLC when given in combination with appropriate sponse Evaluation Criteria in Solid Tumors (RECIST) criteria. cytotoxic chemotherapy regimens. Phase III clinical trials leading Denosumab is also approved in for use in two supportive to the approval of bevacizumab for the treatment of colorectal care settings based on three randomized, double-blind, placebo- cancer demonstrated improved response rates from 35% to 45% controlled phase III trials evaluating its efficacy versus zoledronic compared to fluorouracil (5-FU)–based chemotherapy alone. En- acid123–125 to reduce bone metastasis-related skeletal-related events hanced response durations and improved patient survival were (SRE). Based on data from two phase III trials, a second formula- seen in patients treated with chemotherapy plus bevacizumab as tion and dosing schedule of denosumab is approved to increase compared to patients receiving chemotherapy alone.111 A survival bone mass in prostate cancer126 and breast cancer127 patients at benefit was also seen in the setting of NSCLC. A randomized high risk for bone fracture due to hormone-ablation therapies. phase III trial (ECOG 4599) of paclitaxel and carboplatin with or without bevacizumab in patients with advanced nonsquamous NSCLC led to a significant improvement in median survival (12.5 ANTIBODIES USED IN HEMATOLOGIC months versus 10.2 months; p = 0.0075) for patients in the beva- MALIGNANCIES cizumab arm,112 with significantly higher response rates. A higher incidence of bleeding was associated with bevacizumab (4.5% ver- sus 0.7%). Five of 10 treatment-related deaths occurred as a result Rituximab of hemoptysis, all in the bevacizumab arm. A phase III trial randomized 722 patients with metastatic breast Rituximab (Rituxan) is a chimeric anti-CD20 monoclonal antibody cancer with no prior chemotherapy for advanced disease to either that was the first MAb to be approved by the FDA for use in human paclitaxel or paclitaxel and bevacizumab.113 PFS was significantly malignancy.128,129 Studies have shown that multiple doses can be better in the paclitaxel plus bevacizumab arm (median, 11.8 versus safely administered, and in vitro studies have demonstrated mul- 5.9 months; HR for progression, 0.60; p <0.001) with an increased tiple mechanisms by which anti-CD20 antibodies can lead to cell response rate (36.9% versus 21.2%, p <0.001). Overall survival, death.130 Efficacy of rituximab monotherapy is well established.131 however, was similar. Rituximab has been tested in conjunction with chemotherapy In contrast,114 in a randomized phase III trial, capecitabine/bev- based on supportive preclinical data.132,133 The combination of ritux- acizumab increased response rates compared with capecitabine imab with cyclophosphamide, doxorubicin, vincristine, and prednis- alone in 462 anthracycline and taxane pretreated metastatic breast olone (CHOP) resulted in a 95% overall response rate (55% complete cancer patients but did not meet its primary endpoint of improved response, 40% partial response) among 40 patients with low-grade or PFS. Overall survival and time to deterioration in quality of life follicular B-cell non–Hodgkin lymphoma, with molecular complete were comparable in both treatment groups. remissions observed.134 A long-term study of elderly patients with pre- Bevacizumab has not demonstrated activity in the adju- viously untreated diffuse large-cell lymphoma randomized to either vant colorectal and breast cancer settings.115,116 There was no CHOP chemotherapy plus rituximab (R-CHOP) or CHOP alone 306 Cancer Therapeutics demonstrated a significant improvement in event-free survival, PFS, tested as a therapeutic agent for CLL and promyelocytic l eukemias, disease-free survival, and overall survival for the combination arm.135 as well as other non–Hodgkin lymphomas. No significant differences in long-term toxicity were noted. Low-grade B-cell lymphoma patients possessing the 158V/V polymorphism in FcγRIII experience superior response rates and Brentuximab Vedotin outcomes when treated with rituximab.37,38 These findings signify that antibody Fc domain::Fc receptor interactions underlie at least Brentuximab vedotin (SGN-35, Adcetris) is an ADC consisting of the some of the clinical benefit of rituximab, and indicate a possible anti-CD30 chimeric MAb cAC10 that is linked to three to five mol- role for ADCC that depends on such interactions. ecules of the microtubule-disrupting agent Monomethyl auristatin E (MMAE). MMAE is a highly potent derivative of dolastatin. Linkage A combination of active agents (such as lenalidomide and thalid- 151 omide) that are also immune modulating may be additive with ritux- of MMAE to cAC10 occurs through a protease-cleavable linter. imab,136 and perhaps synergize by increasing ADCC.137 Cytokines Brentuximab vedotin is approved for treating systemic, chemother- such as interleukin-2 (IL-2), IL-12, or IL-15 and myeloid growth fac- apy-refractory anaplastic large-cell lymphomas (sALCL). It is also ap- tors may also enhance therapeutic antibody activity as suggested by proved to treat patients with Hodgkin lymphoma who have progressed preclinical data demonstrating that IL-2 can promote NK cell pro- after an autologous stem cell transplant (ASCT). Patients ineligible for liferation and activation and can enhance rituximab activity138 and ASCT must have failed two prior multidrug chemotherapy regimens. clinical efficacy.139,140 Myeloid growth factors, in combination with Brentuximab vedotin received accelerated approval in 2011 141 based in part on the results of two phase II trials. In a multicenter rituximab, may also activate ADCC. Alternative approaches to 152 induce effector cell activity by combining Toll-like receptors (TLR) trial conducted by Pro et al., 58 patients with relapsed or refrac- agonists, such as CpG oligonucleotides, have been investigated.142 tory sALCL received brentuximab vedotin (1.8 mg per kilogram per Altering the balance of proapoptotic and antiapoptotic signals could week), and 86% of patients achieved objective response. Complete generate more rituximab-induced cytotoxicity. BCL-2 downregula- responses occurred in 57% of patients, with a median duration of tion by antisense oligonucleotides was found to enhance rituximab 13.2 months. An additional 17 patients (29%) had partial responses. efficacy in preclinical testing.143,144 However, small molecules that Median overall response was 12.6 months. Most common grade 3 bind to the BH-3 domain common to many members of the BCL-2 and 4 adverse events (AE) were neutropenia (21%), thrombocytope- family of proteins may be better therapeutic agents.145–147 nia (14%), and peripheral sensory neuropathy (12%). A similar trial, in Hodgkin lymphoma, was reported by Younes et al.153 Patients (n = 102) that had failed ASCT received brentuximab vedotin on Ofatumumab the same schedule as listed previously and were assessed for the objective response rate. In this setting, 75% of patients had objective 148 The anti-CD20 ofatumumab is a fully human antibody that responses, with 34% being complete remissions. The median dura- binds an epitope on CD20 distinct from that bound by rituximab tion of complete responses was 20.5 months, and 31 patients were and is engineered for better complement activation, although it in- progression free after a median follow-up of 1.5 years. Phase III trials duces less ADCC. Ofatumumab has received regulatory approval to assess the known risk of neuropathy (AETHERA) and to confirm for the treatment of patients with -refractory chronic overall clinical benefit seen in the phase II trials (ECHELON-2, or lymphocytic leukemia (CLL). In a recently reported, planned ClinicalTrials.gov Identifier NCT01712490) are ongoing. interim analysis that included 138 CLL patients with treatment- refractory disease or bulky (>5 cm) lymphadenopathy, treatment with ofatumumab led to an overall response rate (primary end- CONCLUSION point) of 47% in patients with bulky disease and 5% in patients refractory to both alemtuzumab and fludarabine.149 150 In the 35 years since Kohler and Milstein first developed the hy- Additional humanized anti-CD20 antibodies (veltuzumab bridoma technology that enabled antibody-based therapeutics, the and ocrelizumab) are under development. field has made remarkable progress. Numerous antibody-based molecules are currently in clinical trials and many more are in de- Alemtuzumab velopment. Multiple therapeutic antibodies have a proven clinical benefit and have been licensed by the FDA. The thoughtful ap- Alemtuzumab (Campath-1H) targets the CD52 glycopeptide, plication of advances in cancer biology and antibody engineering which is highly expressed on T and B lymphocytes. It has been suggest that this progress will continue.

SELECTED REFERENCES

The full reference list can be accessed at lwwhealthlibrary.com/oncology. 38. Weng WK, Levy R. Two immunoglobulin G fragment C receptor polymor- phisms independently predict response to rituximab in patients with follicu- 4. Reichert JM, Dhimolea E. The future of antibodies as cancer drugs. Drug lar lymphoma. J Clin Oncol 2003;21:3940–3947. Discov Today 2012;17:954–963. 48. Mack M, Riethmuller G, Kufer P. A small bispecific antibody construct ex- 5. Kohler G, Milstein C. Continuous cultures of fused cells secreting antibody pressed as a functional single-chain molecule with high tumor cell cytotoxic- of predefined specificity. Nature 1975;256:495–497. ity. Proc Natl Acad Sci U S A 1995;92:7021–7025. 10. Komissarov AA, Calcutt MJ, Marchbank MT, et al. Equilibrium bind- 49. Bargou R, Leo E, Zugmaier G, et al. Tumor regression in cancer patients by ing studies of recombinant anti-single-stranded DNA Fab. Role of heavy very low doses of a T cell-engaging antibody. Science 2008;321:974–977. chain complementarity-determining regions. J Biol Chem 1996;271: 54. Di Gaetano N, Cittera E, Nota R, et al. Complement activation determines 12241–12246. the therapeutic activity of rituximab in vivo. J Immunol 2003;171:1581–1587. 11. Ghetie V, Popov S, Borvak J, et al. Increasing the serum persistence of an IgG 57. Li S, Schmitz KR, Jeffrey PD, et al. Structural basis for inhibition of the fragment by random mutagenesis. Nat Biotechnol 1997;15:637–640. epidermal growth factor receptor by cetuximab. Cancer Cell 2005;7: 19. Jain RK. Physiological barriers to delivery of monoclonal antibodies and 301–311. other macromolecules in tumors. Cancer Res 1990;50:814s–819s. 59. Presta LG, Chen H, O’Connor SJ, et al. Humanization of an anti-vascular 29. Shields RL, Namenuk AK, Hong K, et al. High resolution mapping of the endothelial growth factor monoclonal antibody for the therapy of solid tu- binding site on human IgG1 for Fc gamma RI, Fc gamma RII, Fc gamma mors and other disorders. Cancer Res 1997;57:4593–4599. RIII, and FcRn and design of IgG1 variants with improved binding to the Fc 63. Kreitman RJ, Wilson WH, Bergeron K, et al. Efficacy of the anti-CD22 gamma R. J Biol Chem 2001;276:6591–6604. recombinant immunotoxin BL22 in chemotherapy-resistant hairy-cell 31. McDonagh CF, Huhalov A, Harms BD, et al. Antitumor activity of a novel leukemia. N Engl J Med 2001;345:241–247. bispecific antibody that targets the ErbB2/ErbB3 oncogenic unit and inhibits 66. Lode HN, Xiang R, Becker JC, et al. Immunocytokines: a promising ap- heregulin-induced activation of ErbB3. Mol Cancer Ther 2012;11:582–593. proach to . Pharmacol Ther 1998;80:277–292.

tahir99 - UnitedVRG Chapter 29 Monoclonal Antibodies 307

74. Castaigne S, Pautas C, Terre C, et al. Effect of gemtuzumab ozogamicin 106. Amado RG, Wolf M, Peeters M, et al. Wild-type KRAS is required for pani- on survival of adult patients with de-novo acute myeloid leukaemia (ALFA- tumumab efficacy in patients with metastatic colorectal cancer. J Clin Oncol 0701): a randomised, open-label, phase 3 study. Lancet 2012;379:1508–1516. 2008;26:1626–1634. 76. Lambert JM. Drug-conjugated antibodies for the treatment of cancer. Br J 111. Hurwitz H, Fehrenbacher L, Novotny W, et al. Bevacizumab plus irinotecan, Clin Pharmacol 2013;76:248–262. fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 82. Witzig TE, White CA, Wiseman GA, et al. Phase I/II trial of IDEC-Y2B8 2004;350:2335–2342. radioimmunotherapy for treatment of relapsed or refractory CD20(+) B-cell 112. Sandler A, Gray R, Perry MC, et al. Paclitaxel-carboplatin alone or with non-Hodgkin’s lymphoma. J Clin Oncol 1999;17:3793–3803. bevacizumab for non-small-cell lung cancer. N Engl J Med 2006;355: 83. Carter P, Presta L, Gorman CM, et al. Humanization of an anti-p185HER2 an- 2542–2550. tibody for human cancer therapy. Proc Natl Acad Sci U S A 1992;89:4285–4289. 113. Miller K, Wang M, Gralow J, et al. Paclitaxel plus bevacizumab versus pacli- 90. Piccart-Gebhart MJ, Procter M, Leyland-Jones B, et al. Trastuzumab af- taxel alone for metastatic breast cancer. N Engl J Med 2007;357:2666. ter adjuvant chemotherapy in HER2-positive breast cancer. N Engl J Med 120. Verma S, Miles D, Gianni L, et al. Trastuzumab emtansine for HER2-posi- 2005;353:1659–1672. tive advanced breast cancer. N Engl J Med 2012;367:1783–1791. 91. Romond EH, Perez EA, Bryant J, et al. Trastuzumab plus adjuvant chemo- 122. Chawla S, Henshaw R, Seeger L, et al. Safety and efficacy of denosumab therapy for operable HER2-positive breast cancer. N Engl J Med 2005;353: for adults and skeletally mature adolescents with giant cell tumour of bone: 1673–1684. interim analysis of an open-label, parallel-group, phase 2 study. Lancet Oncol 92. Smith I, Procter M, Gelber RD, et al. 2-year follow-up of trastuzumab after 2013;14:901–908. adjuvant chemotherapy in HER2-positive breast cancer: a randomised con- 128. Maloney D, Grillo-López A, Bodkin D, et al. IDEC-C2B8: results of a phase trolled trial. Lancet 2007;369:29–36. I multiple-dose trial in patients with relapsed non-Hodgkin’s lymphoma. J 94. Bang YJ, Van Cutsem E, Feyereislova A, et al. Trastuzumab in combination Clin Oncol 1997;15:3266–3274. with chemotherapy versus chemotherapy alone for treatment of HER2-posi- 131. Coiffier B, Haioun C, Ketterer N, et al. Rituximab (anti-CD20 monoclonal tive advanced gastric or gastro-oesophageal junction cancer (ToGA): a phase antibody) for the treatment of patients with relapsing or refractory aggressive 3, open-label, randomised controlled trial. Lancet 2010;376:687–697. lymphoma: a multicenter phase II study. Blood 1998;92:1927–1932. 97. Baselga J, Cortes J, Kim SB, et al. Pertuzumab plus trastuzumab plus 135. Feugier P, Van Hoof A, Sebban C, et al. Long-term results of the R-CHOP docetaxel for metastatic breast cancer. N Engl J Med 2012;366:109–119. study in the treatment of elderly patients with diffuse large B-cell lymphoma: 98. Swain SM, Kim SB, Cortes J, et al. Pertuzumab, trastuzumab, and docetaxel a study by the Groupe d’Etude des Lymphomes de l’Adulte. J Clin Oncol for HER2-positive metastatic breast cancer (CLEOPATRA study): overall 2005;23:4117–4126. survival results from a randomised, double-blind, placebo-controlled, phase 140. Khan KD, Emmanouilides C, Benson DM Jr., et al. A phase 2 study of ritux- 3 study. Lancet Oncol 2013;14:461–471. imab in combination with recombinant interleukin-2 for rituximab-refracto- 101. Van Cutsem ELI, D’haens G. KRAS status and efficacy in the first-line treat- ry indolent non-Hodgkin’s lymphoma. Clin Cancer Res 2006;12:7046–7053. ment of patients with metastatic colorectal cancer (metastatic CRC) treated 149. Wierda WG, Kipps TJ, Mayer J, et al. Ofatumumab as single-agent CD20 with FOLFIRI with or without cetuximab: The CRYSTAL experience. Ab- immunotherapy in fludarabine-refractory chronic lymphocytic leukemia. J stract 2. J Clin Oncol 2008;26:5s. Clin Oncol 2010;28:1749–1755. 103. Pirker R, Pereira JR, Szczesna A, et al. Cetuximab plus chemotherapy in 152. Pro B, Advani R, Brice P, et al. Brentuximab vedotin (SGN-35) in patients patients with advanced non-small-cell lung cancer (FLEX): an open-label with relapsed or refractory systemic anaplastic large-cell lymphoma: results of randomised phase III trial. Lancet 2009;373:1525–1531. a phase II study. J Clin Oncol 2012;30:2190–2196. 105. Gibson TB, Ranganathan A, Grothey A. Randomized phase III trial of panitu- 153. Younes A, Gopal AK, Smith SE, et al. Results of a pivotal phase II study mumab, a fully human anti-epidermal growth factor receptor monoclonal an- of brentuximab vedotin for patients with relapsed or refractory Hodgkin’s tibody, in metastatic colorectal cancer. Clin Colorectal Cancer 2006;6:29–31. lymphoma. J Clin Oncol 2012;30:2183–2189. CANCER THERAPEUTICS Assessment of Clinical 30 Response

Antonio Tito Fojo and Susan E. Bates

INTRODUCTION and medicine advanced. Table 30.1 compares the WHO criteria with those of RECIST 1.0 and RECIST 1.1 and three modifica- Approaches to response assessments have become increasingly tions of RECIST, whereas Figure 30.1 provides a visual presenta- important over the past decade as the drug development pipeline tion of the RECIST threshold required to qualify as response or has steadily increased in volume. In 2012, an estimated 981 medi- progression.3–9 cines were in development for cancer, and the number is certainly higher today.1 The challenge is, first, how to measure the activity of an agent in the research setting, and, second, how to measure ASSESSING RESPONSE activity in the standard of care setting. The “modern era” of drug development began in 1976 when RECIST 1.1 16 experienced oncologists treating lymphoma gathered to decide what would be considered a reliable measure of response to a The RECIST 1.0 guidelines were updated as RECIST 1.1 in therapy.2 Each oncologist measured 12 simulated tumor masses 2009, with a number of differences between the two response cri- employing usual clinical methods (i.e., calipers or rulers). A prin- teria highlighted. RECIST 1.1 preserves the same categories of cipal goal was to identify the amount of shrinkage that could not response found in RECIST 1.0: be ascribed to operator error and that would not be found if a pla- cebo was administered. Moertel and Hanley recommended that ■ Complete response: Complete disappearance of all disease to avoid error, a 50% reduction in the product of perpendicular ■ Partial response: ≥30% reduction in the sum of the longest diameters be employed as the criterion for efficacy.2 It was from this diameter of target lesions beginning that our current methodologies of response assessment ■ Stable disease: Change not meeting criteria for response or evolved. The important point to note is that the decision to use progression a 50% reduction in the product of perpendicular diameters as a ■ Progression: ≥20% increase in the sum of the longest diameter measure of efficacy was made so as to reduce error andnot because of target lesions it represented a value that conferred clinical benefit. However, a decade of experience with RECIST identified several problems with the criteria, some of which could be cor- From Calipers and Rulers in Lymphoma to rected. In RECIST 1.0, minimum size varied between 1 and 2 cm the Bidimensional World Health Organization depending on technique; in RECIST 1.1, a 1-cm lesion is the minimum measurable. In RECIST 1.0, 10 lesions were to be mea- Criteria sured, 5 per organ; RECIST 1.1 reduced that to 5 lesions, 2 per

2 organ. Response criteria in RECIST 1.0 did not address lymph In 1981, five years after the Moertel and Hanley report, a World nodes; in RECIST 1.1, lymph nodes decreasing to <1 cm in their Health Organization (WHO) initiative developed standardized ap- short axis could constitute a complete response. Disease progres- proaches for the “reporting of response, recurrence and disease- 3 sion in nontarget disease was further defined to indicate that in free interval.” The WHO criteria, like Moertel and Hanley, addition to a 20% increase in target lesions over the smallest sum recommended that malignant disease be measured in two dimen- on study, there must be an absolute increase of 5 mm, and that an sions. Complete response (CR) was defined as the disappearance increase of a single nontarget lesion should not trump an overall of all known disease, and a partial response (PR) was scored if there disease status assessment based on target lesions. occurred a “50% decrease in the sum of the products of the perpen- dicular diameters of the multiple lesions.” Thus, the 50% reduction initially chosen as an operationally optimal value became institu- Variations of the RECIST Criteria tionalized as the threshold for declaring efficacy in the majority of cancers. This measure of efficacy was perpetuated in 2000 with the The RECIST criteria have been widely used for standardizing now widely used Response Evaluation Criteria in Solid Tumors the reporting of clinical trial results and have improved repro- (RECIST), but shifting to one dimension.4 The authors noted “the ducibility. However, the increasing precision and codification of definition of a partial response, in particular, is an arbitrary con- RECIST has led to recognition of its limitations. For example, vention—there is no inherent meaning for an individual patient of there are unique challenges in central nervous system (CNS) a 50% decrease in overall tumor load.” Nevertheless, the threshold disease, relating response to tumor size measurements based on chosen—a 30% reduction in one dimension—was comparable in contrast enhancement. Pseudoprogression refers to an increase volume to the 50% decrease in the sum of the products of the in contrast enhancement due to a transient increase in vascu- perpendicular diameters and thus perpetuated the 1976 standard. lar permeability after irradiation, whereas pseudoresponse is a In spite of its arbitrary origins, the 50% reduction has held up decrease in contrast enhancement that may occur due to a re- over time. But the major impact of the WHO criteria was that it duction in vascular permeability following corticosteroids or an marked the beginning of a common language of response. These antiangiogenic agent such as bevacizumab.10–12 The McDonald criteria have been revisited and refined over time, as technology criteria, traditionally used in determining glioma response based

308

tahir99 - UnitedVRG Chapter 30 Assessment of Clinical Response 309

TABLE 30.1 Key Features of Response Criteria

CNS RANO RECIST RECIST WHO3 RECIST 1.04 RECIST 1.15 Criteria7 Mesothelioma8 Immunotherapy9 Dimension Uni- and Unidimensional Unidimensional Bidimensional Unidimensional Bidimensional bidimensional Measurable Not definedongest L diameter, Longest Two Tumor thickness Longest Lesion ≥20 mm with diameter ≥10 perpendicular perpendicular perpendicular most modalities; mm on CT or diameters to chest wall or diameters ≥10 mm with on skin if using of contrast mediastinum, spiral CT calipers; ≥20 enhancing measured in mm if using CXR lesions ≥10 mm two positions at three levels on transverse cuts of CT scan Measurable Not defined Not defined ≥15 mm short —— — Lymph Nodes axis Disease Burden All (not Measurable Measurable Two to five Pleural disease 5 lesions per to be Assessed at specified) target lesions target lesions lesions in in perpendicular organ, up to 10 Baseline up to 10 total (5 up to 5 total patients with diameter; nodal, visceral lesions per organ); other (2 per organ); several lesions subcutaneous, and five cutaneous lesions nontarget other lesions and other lesions nontarget bidimensional lesions measured unidimensionally as per the RECIST criteria CANCER THERAPEUTICS Sum Sum of the Sum of longest Sum of the Sum of the Sum of the six SPD with products of diameters of longest products of measurements new lesions bidimensional all measurable diameters of perpendicular defines a pleural incorporated into diameters or lesions target lesions diameters of unidimensional baseline; tumor sum of linear with only all measurable measure burden =

unidimensional exception use enhancing target SPDindex lesions + diameters of short axis for lesions SPDnew lesions lymph nodes Complete Disappearance Disappearance all Disappearance — Disappearance Disappearance Response all known known disease all known all target lesions all lesions in disease disease; lymph with no evidence two consecutive nodes <10 mm of tumor observations elsewhere Partial Response ≥50% ≥30% decrease; ≥30% decrease; ≥50% reduction; ≥30% reduction ≥50% decrease decrease all other no all other disease, stable or in total tumor compared with evidence of no evidence of decreased measurement baseline in two progression progression steroid use observations compared to baseline Response ≥4 weeks apart ≥4 weeks apart ≥4 weeks apart ≥4 weeks apart Repeat on two ≥4 weeks apart Confirmation? (if response occasions ≥4 primary end weeks apart point); no, if secondary endpoint

(continued) 310 Cancer Therapeutics

TABLE 30.1 Key Features of Response Criteria (continued)

CNS RANO RECIST RECIST WHO3 RECIST 1.04 RECIST 1.15 Criteria7 Mesothelioma8 Immunotherapy9 Progressive ≥25% increase ≥20% increase, ≥20% increase, ≥25%, or any ≥20% increase ≥25% increase Disease in size of taking as with absolute new lesions in the total tumor compared with one or more reference increase ≥5 measurement nadir confirmed≥ 4 measurable smallest sum mm, taking over the nadir weeks apart; up to lesions or in study; or as reference measurement, or five new lesions appearance of appearance of smallest sum the appearance (≥5 × 5 mm) per new lesions new lesions in study; or of one or more organ incorporated appearance of new lesions into tumor burden new lesions Nonmeasurable Nonmeasurable Nonmeasurable Nonmeasurable —New, disease: disease: disease: disease: >5 nonmeasurable Estimated unequivocal unequivocal mm increase lesions (i.e., <5 increase of progression progression in maximal × 5 mm) do not ≥25% diameter; ≥25% define progression increase in SPD; or significant increase in nonenhancing lesions on same or lower dose of corticosteroids Stable Disease Stable disease Non-PR, non-PD; Non-PR, non-PD; — Non-PR, non-PD Non-irPR, non-irPD or non-PR and minimum time minimum time non-PD ≥4 defined by defined by weeks protocol protocol

CXR, Chest X-ray; SPD, sum of products of two largest perpendicular diameters; PD, progressive disease; irPR, immune-related partial response; irPD, immune- related progressive disease.

on two-dimensional measurements, have been recently updated strategies have emerged to quantify these diseases, including modi- as part of the Response Assessment in Neuro-Oncology (RANO) fications of RECIST, quantifying positron-emission tomography response criteria and extended to include a response assessment (PET) imaging, and biomarker criteria, as will be discussed. The for metastatic CNS disease.7,13 RECIST adaptation for mesothelioma, growing along the pleural Other examples where RECIST is limited include mesothe- surface, is to measure the diameter perpendicular to the chest wall lioma, gastrointestinal stromal tumors (GIST), hepatocellular can- or mediastinum, and to measure at three levels.8 The adaptation cers, among others. The pleural disease of mesothelioma increases for hepatocellular cancer following local therapy is measurement in depth while following the pleural surface. GIST tumors may of the longest diameter of the tumor that shows enhancement on remain unchanged in size after treatment, whereas the center of the arterial phase of the scan, bypassing the dense, homogeneous the tumor mass undergoes necrosis, and progression may occur Lipiodol-containing necrotic area.15 in the remaining rim.14 Hepatocellular cancers are often treated Investigators have also observed that following immunotherapy, with local–regional therapy in which the goal is tumor necrosis tumor lesions may increase in size due to the increased infiltration and treatment failure occurs in surviving viable tumor.15 Different of T cells, even meeting criteria for RECIST-defined progressive

PD Figure 30.1 RECIST thresholds in three Progressive parameters: diameter, product of diameters, disease and volume. In the figure, spheres meeting Diameter: 120% RECIST criteria for progressive disease (PD) Initial Product of diameters: 144% and for PR are shown with the percentage Volume: 173% relative to the baseline calculated for each parameter. To meet the threshold for PD, the Diameter: 100% longest diameter must increase to 120%, Product of diameters: 100% PR PD which is equivalent to a 144% increase in Volume: 100% Response the product of the perpendicular diameters progression and a 173% increase in the volume of a Diameter: 70% Diameter: 84% sphere. Although PR definitions are almost Product of diameters: 49% Product of diameters: 70% identical to those employed with WHO, Volume: 34% Volume: 59% RECIST has a higher threshold to meet PD.6

tahir99 - UnitedVRG Chapter 30 Assessment of Clinical Response 311 disease (PD). Previously radiographically undetectable lesions may pooled analysis of 11,955 patients enrolled on 12 neoadjuvant tri- appear. Departing from conventional RECIST, which defines any als, individual patients with pCR had improved EFS and OS.21 new lesion as PD, the immune response criteria allow the appear- However, at the trial level, pCR rates did not correlate with EFS ance of new lesions, adding them to the total tumor burden.9 An or OS, a problem likely due to heterogeneity of breast cancer sub- increase in total tumor burden of >25% relative to baseline or types among the trials. Despite this, pCR rates were recently used nadir is required to define PD. to support the approval of pertuzumab and trastuzumab in the neoadjuvant setting.21,22 International Working Group Criteria for Computed Tomography-Based Tumor Density Lymphoma One approach, often called the Choi criteria, advocates assess- Revised guidelines for lymphoma assessment were promulgated ing tumor response in GIST, renal cell cancer, or hepatocellular by the International Working Group (IWG) in 2007.16 These cancer based on density on computed tomography (CT) scans guidelines incorporated 18F-fluorodeoxyglucose (FDG)-PET as- (Table 30.2). This variation was prompted by the evident response sessments in metabolically active lymphomas.16 Although a CR to treatment with imatinib but with minimal tumor shrinkage.23 requires the complete disappearance of detectable disease, a post- The Choi criteria are still considered exploratory in GIST,24,25 and treatment residual mass is permitted if it is negative on FDG-PET it is too soon to know of benefits in other histologies.26,27 Further and was positive at baseline. For lymphomas that are not consis- study should determine its utility, although it will likely be con- tently FDG avid, or if FDG avidity is unknown, a CR requires fined to specific tumor types with specific drugs. that nodes >1.5 cm before therapy regress to <1.5 cm, and nodes that were 1.1 to 1.5 cm in long axis and >1.0 cm in the short axis FDG-PET shrink to ≤1.0 cm in short axis. The definition of PR resembles the WHO criteria, in that a ≥50% decrease in the sum of the Although widely used in clinical practice, FDG-PET has become product of the diameters in up to six nodal masses or in hepatic part of standardized response criteria for clinical trials only in lym- or splenic nodules must be documented. Although RECIST 1.1 phoma (see Table 30.2). In solid tumors, FDG-PET can aid in the now includes lymph node assessment, the IWG criteria remain detection of new or recurrent sites of disease, and can be used as the assessment method typically used in lymphoma clinical trials. an adjunct during assessments for disease progression when using RECIST criteria.5 Although FDG uptake is a powerful diagnostic tool and its uptake reflects a tumor’s metabolic activity, it has some ALTERNATE RESPONSE CRITERIA limitations: Some tumors have variable FDG avidity; differences can occur due to variations in patient activity, carbohydrate intake, blood glucose, and timing; and there are several benign sources

The previous examples represent attempts to more accurately mea- CANCER THERAPEUTICS sure tumor burden. Evolving imaging technology enabling volu- of uptake, including inflammatory and postsurgical sites. Multiple metric measurements of tumor masses may eventually resolve some methods of quantitating FDG-PET and assessing response have been proposed, but to date there is no consensus, particularly of these problems, but effective therapeutic agents are required to 28–33 enable validation and utilization of response assessment tools. The regarding the definition of a metabolic response. lack of an agent that can mediate substantial tumor shrinkage un- The two most widely used response criteria—the European derlies the concept of clinical benefit response (CBR) as an endpoint Organisation for the Research and Treatment of Cancer (EORTC) in pancreatic cancer. Clinical benefit was defined as a combination criteria and PET Response Criteria in Solid Tumors (PERCIST) of improvement in pain, performance status, and weight; the assess- (see Table 30.2)—have been evaluated in specific disease types, but unifying FDG-PET response criteria remains a challenge in ment of CBR supported the U.S. Food and Drug Administration 28,30 (FDA) approval of gemcitabine in pancreatic cancer.17,18 Better anticancer drug development. We would note that, as shown therapies for pancreatic cancer that result in tumor shrinkage or in Figure 30.1, a 30% reduction in the diameter of a sphere—the eradication should include and then eclipse clinical benefit. magnitude of change required to score a response according to Response criteria may be specific to a particular disease or clini- RECIST—represents a 65% decrease in volume. If an standard- cal setting. Some diseases by their nature require specific strategies ized uptake value (SUV) decrease is directly equated to a volume for response assessment. decrease, a reduction of 25% translates to a 10% reduction in diam- eter, a value that likely constitutes an insufficient response. Severity-Weighted Assessment Tool Score in Cutaneous T-Cell Lymphoma Serum Biomarkers of Response The ideal response assessment method is an assay that could mea- Cutaneous T-cell lymphoma (CTCL) is a disease that can involve sure tumor quantity by a simple blood test (see Table 30.2). Cir- the entire epidermis, or comprise individual skin lesions varying culating protein biomarkers have been identified and studied for widely in severity rather than size. The severity-weighted assessment several decades for screening, early detection of recurrent disease, tool (SWAT) assigns a factor for skin lesion severity—patch, plaque, determining prognosis, selecting therapy, and monitoring response or tumor—multiplies this factor by the percent of skin involved with to therapy. These serum tumor markers are to be distinguished each lesion type and then adds these together. This complex system from the assays determining the presence of an overexpressed or formed the basis of the FDA approval of vorinostat for CTCL.19 mutated molecular target. With the successful launch of therapies Pathologic Complete Response in Breast Cancer against such molecular targets, there has been increased inter- est in the assays needed to select therapy for individual patients One unique response endpoint is the assessment of breast cancer (predictive biomarkers). The analytical and clinical validation of treated in the neoadjuvant setting. The purpose of neoadjuvant such assays, along with determination of their clinical utility, has therapy is to improve survival, render locally advanced cancer created a new regulatory paradigm known as companion diagnos- amenable to surgery, or to aid in breast conservation. In that set- tics.34,35 This investment in the development of predictive markers ting, the absence of cancer cells in resected breast tissue has been for companion diagnostics has reduced the focus on protein bio- used to define a pathologic complete response (pCR). The rate markers of treatment response relative to older literature. of pCR has been proposed as a surrogate endpoint for event-free As a result, there are few clinically validated biomarkers of re- survival (EFS) or overall survival (OS) to support approval of new sponse.36 In addition to issues regarding sensitivity and specificity, agents or combinations of agents tested in clinical trials.20 In a their use and development has also been hindered by the often 312 Cancer Therapeutics

TABLE 30.2 Alternate Response Criteria: Biomarkers

Criterion Baseline Response Progression CA-125 in ovarian cancer (GCIG CA 125 >2× ULN CA 125 decline ≥50% 2× nadir OR criteria)43 confirmed at 28 days 2× ULN if normalized on therapy on two occasions 1 wk apart PSA in prostate cancer (PSA PSA ≥5 ng/mL and PSA decline of 50% from After decrease from baseline, a WG1)45,a documentation of two baseline (measured twice 3–4 50% increase AND an increase consecutive increases in PSA wks apart) ≥5 ng/mL, or back to baseline, 1 wk apart whichever is lower PSA in prostate cancer PSA ≥2.0 ng/mL; estimate Report percent change from PSA increase ≥25% and absolute (PCWG2)46,a pretreatment PSA-DT: Need baseline (rise or fall) at 12 increase by ≥2 ng/mL above ≥3 values ≥4 wks apart weeks, and separately, the the nadir, confirmed by a maximal change (rise or fall) at second value ≥3 wks later (i.e., any time using a waterfall plot confirmed rising trend) OR PSA increase ≥25% and ≥2 ng/mL above baseline >12 wks hCG and AFP in testicular Decrease consistent with Rising levels usually indicate need cancer50–51 marker half-life: 2–3 d for to change therapy hCG, 5–7 d for AFP Choi Criteria for CT Imaging Choi criteria24–27 ≥10% decrease in tumor size An increase in tumor size ≥10% OR and does not meet criteria of ≥15% reduction in tumor PR by tumor attenuation on CT density

FDG-PET Criteria EORTC criteria29–31 ROI should be drawn, SUV CMR: Complete resolution of PMD: SUV increase >25% in calculated uptake regions defined on baseline, or PMR: SUV reduction ≥25% appearance of new FDG-avid SMD: SUV increase <25% and lesions decrease <15% PERCIST criteria29 SUL peak >1. 5 × normal liver CMR: Complete resolution of PMD: SUL increase >30% in uptake regions defined on baseline, or PMR: SUL reduction ≥30% appearance of new FDG avid lesions a Guidelines for PSA assessment have evolved from those of the PSAWG1, where responses were dichotomized based on the percent decline, to those in the PCWG2 where PSA response is considered a continuous variable. Recently, emphasis has shifted to assessing PSA doubling time. CA-125, cancer antigen 125; GCIG, Gynecologic Cancer InterGroup; ULN, upper limit of normal; PSA, prostate-specific antigen; PSAWG1, PSA Working Group 1; PCWG2, Prostate Cancer Working Group 2; PSA-DT, PSA-doubling time; hCG, human chorionic gonatropin; AFP, alpha-fetoprotein; EORTC, European Organisation for the Research and Treatment of Cancer; ROI, regions of interest; SUV, standardized uptake value; CMR, complete metabolic response; PMR, partial metabolic response; SMD, stable metabolic disease; PMD, progressive metabolic disease; PERCIST, PET response criteria in solid tumors; SUL, SUV normalized to lean body mass. limited efficacy of therapies; response biomarkers are of little value 1999, established PSA criteria, particularly for use in patients without highly effective primary and salvage therapies. For exam- with disease that was difficult to quantify.44 There followed a ple, a recent clinical trial indicates that in asymptomatic patients second working group (PCWG2) that recommended plotting with ovarian cancer whose only evidence of disease progression is the percent PSA change for each patient in a waterfall plot so an isolated rising CA-125, nothing is gained by instituting treat- as to avoid creating a dichotomous variable from the changes ment before there is other evidence of progression.37,38 in PSA.45 PCWG2 also recommended keeping patients on trial until evidence of a change in clinical status—either sympto- ■ Cancer Antigen 125 (CA-125): Despite recognized limita- matic or radiographic progression. The latter addressed con- tions, CA-125 is widely used. For example, the Gynecologic cerns with patients in whom PSA changes did not reflect clinical Cancer InterGroup (GCIG) criteria have evolved to help deter- status, particularly those with transient increases in the first 12 mine whether a patient’s tumor has responded to therapy.39–41 weeks of a new therapy. Response is defined as a 50% decline from an elevated base- ■ Human Chorionic Gonadotropin (hCG) and alpha feto- line value, whereas progression is defined as a doubling over protein (AFP): Because testicular cancer is a highly curable the nadir or the upper limit of normal.42 In clinical practice, disease with validated biomarkers, outcome assessment has CA-125 levels are followed as part of standard management, focused on the rapid detection of patients whose tumors have but making clinical decisions on marker changes alone is not a poor response to therapy. Because both markers have rela- recommended.43 tively short half-lives—2 to 3 days for hCG and 5 to 7 days ■ Prostate-Specific Antigen (PSA): Similar issues have con- for serum AFP—the rate of decline can be determined. fronted investigators caring for patients with prostate cancer. The Various methods have demonstrated that a rapid decline or PSA Working Group 1 (PCWG1) guidelines, first published in early normalization of marker levels is indicative of a good

tahir99 - UnitedVRG Chapter 30 Assessment of Clinical Response 313

outcome, without any one method achieving widespread benefits.56–58 Equally important, however, is the duration of re- acceptance.46–48 Nonetheless, the 2010 American Society sponse, a value that is measured from the time of initial response of Clinical Oncology (ASCO) guidelines on serum tumor until documented tumor progression, and which assumes added markers concluded there was still insufficient evidence to importance when ORR is the endpoint for regulatory approval. recommend changing therapy solely on the basis of a slow Unlike PR and CR, the FDA has generally not been willing marker decline.49 Rising levels after two cycles of therapy to include stable disease (SD), defined as shrinkage that qualifies (outside the first week of treatment when rises can be due to as neither response nor progression, as part of the ORR, feeling it tumor lysis) can be considered an indication to change the is often indicative of the underlying disease biology rather than a treatment plan.49,50 drug’s therapeutic effect.55,59 Nevertheless, in reporting data, in- vestigators are increasingly using the term CBR, which includes CR + PR + SD and which is a misuse of the term clinical benefit Circulating Tumor Cells and Circulating because neither CR, PR, or SD are objective tumor findings that Tumor DNA address the true clinical benefit of a therapy.58,60 In the absence of standardized definitions for SD that are shown to effect meaning- Two response endpoints under recent investigation show a poten- ful changes in a clinical outcome, SD should not be used as a tial to detect the impact of therapy. One is the measurement of cir- response endpoint. A better approach is to use nondichotomized culating tumor cells (CTC) in the bloodstream, enriched by one response assessments, such as the waterfall plot or one of the ki- or more capture strategies, including one that has received FDA netic analyses, discussed later. approval.51 The number of CTCs in the blood has been shown to be prognostic, with higher levels conferring a poor prognosis, and to correlate with a response to therapy. A second approach is Progression-Free Survival, Time to the determination of levels of circulating tumor DNA (ctDNA) in Progression, and Time to Treatment Failure the blood. This is detected by quantitating the number of DNA molecules carrying a given mutation or gene rearrangement in the In cancer drug development, one usually finds ORR assessed as an blood, typically detected through targeted sequencing of common indicator of activity in phase II trials, whereas randomized phase mutations, or of a previously identified mutation signature or gene III trials rely on other endpoints such as progression-free survival rearrangement. The amount of ctDNA appears to correlate with (PFS) and time to progression (TTP) (see Table 30.3). Although tumor burden, increases with stage, and in one study, was deemed PFS and TTP attempt to assess efficacy in close proximity to a more sensitive than CTC detection.52–54 Whether these tests will therapy, they score outcomes differently and are not interchange- ultimately prove to be more sensitive and accurate than the serum able. TTP is defined as the time from randomization tothe time of biomarkers discussed previously remains to be determined. Be- disease progression.55 In TTP analyses, deaths are censored either at cause targeted sequencing can be very sensitive, one concern is the time of death or at an earlier visit. In contrast, PFS is defined CANCER THERAPEUTICS that false-positive ctDNA detection may occur after treatment, or from the time of randomization to the time of disease progression or intermittently in the setting of enlarging tumor masses. At the least, death. Although patients who discontinue trial participation for ad- detection of CTCs and ctDNA is advancing our understanding of verse events might be censored in both analyses, patients who die cancer biology, as studies reveal evidence of metastatic heteroge- while on study are censored only in the TTP analysis. Those who neity, clonal heterogeneity, and emergence of resistance mutations favor TTP argue that if a patient dies without their tumor meet- in clinical samples. ing criteria for progression, one cannot accurately estimate when progression might have occurred, so the data should be censored. However, those who favor PFS argue that, in some cases, death DETERMINING OUTCOME might be an adverse effect of the therapy. High-dose therapies rep- resent an example of why PFS might be a preferable (regulatory) The response measures described previously represent differ- endpoint. If in a given tumor there is evidence of a dose-response ent approaches to quantitate tumor burden. What happens after relationship for an active drug, then high doses may have a greater those data are obtained varies depending on the clinical setting. response. However, such high doses may also be responsible for In the community, less emphasis is placed on strict criteria. In a greater number of deaths. Assessing only those who survive the the setting of a clinical trial, tumor size is measured and the re- high dose therapy and ignoring those who die (i.e., TTP) may lead sponse categorized. For FDA submission, these are but factors in to the conclusion that the high-dose therapy is more effective. The the risk-benefit equation needed for drug approvals. The FDA balance sheet that includes death (i.e., PFS) would clearly demon- conveys full approval to new agents based on true clinical benefit strate this efficacy came at too great a price. (i.e., an improvement in a survival endpoint or symptom relief).55 Although many have argued that PFS and TTP should be ac- Surrogates for clinical benefit, such as response rate, may sup- ceptable endpoints for cancer clinical trials, in the majority of tu- port either regular approval or accelerated approval, depending mors there is no convincing evidence PFS is a surrogate for OS, on the setting. and in those where there is some evidence, its value is arguable.61 Table 30.3 presents the attributes and drawbacks of PFS and TTP. Note that the definition of progression is often difficult, particu- Overall Response Rate, Duration of Response, larly in some tumor types, and that investigator bias can influence and Stable Disease PFS and TTP. Problems with ascertainment bias and censoring, depicted in Figure 30.2, can also impact outcomes. Overall response rate (ORR) is the proportion of patients with Alternate endpoints include time to treatment failure (TTF), a tumor size reduction of a predefined amount for a minimum defined as a composite endpoint measuring time from random- time period. The FDA has generally defined ORR as the sum of ization to discontinuation of treatment for any reason, includ- PRs and CRs. Although OS remains the gold standard, ORR is ing disease progression, treatment toxicity, and death. The FDA often used both in drug development and in clinical practice to has not recommended TTF as a regulatory endpoint for drug indicate antitumor efficacy of a given therapy. Table 30.3 sum- approval. However, the high rates of censoring due to toxicity marizes the attributes and drawbacks of using ORR as a method seen in phase III clinical trails may lead to a reassessment of this of assessment. Using standardized definitions of response, it has position given that most can agree that not only is efficacy im- been shown that ORR often correlates with OS, although ORR portant, but so too is tolerability, and TTF can capture both of usually explains only a fraction of the variability of the survival these attributes. 314 Cancer Therapeutics

TABLE 30.3 A Comparison of Important Cancer Approval Endpoints

Regulatory Evidence Endpoints Advantages Disadvantages Clinical benefit used for Overall survival (OS) ■ Universally accepted direct measure of ■ May involve larger studies regular approvals clinical benefit ■ May require long follow-up ■ Easily measured ■ May be affected by crossover and/or ■ Includes treatment-related mortality sequential therapies that can obscure benefit in a subset ■ Includes noncancer deaths ■ Precisely measured; unambiguous ■ Not dependent on assessment intervals Symptom endpoints ■ Patient perspective of direct clinical ■ Blinding is often difficult (patient-reported benefit ■ Data are frequently missing or outcomes) incomplete ■ Clinical significance of small changes is unknown ■ Multiple analyses ■ Lack of validated instruments Surrogates used for Disease-free survival ■ Smaller sample size and shorter ■ Not statistically validated as surrogate accelerated approvals or (DFS) follow-up necessary compared with for survival in all settings regular approvals survival studies ■ Not precisely measured; subject to assessment bias, particularly in open- label studies ■ Definitions vary among studies Objective response ■ Can be assessed in single-arm studies ■ Not a direct measure of benefit rate (ORR) ■ Assessed earlier and in smaller studies ■ Not a comprehensive measure of drug compared with survival studies activity ■ Effect attributable to drug, not natural ■ Only a subset of patients who benefit history Complete response ■ Can be assessed in single-arm studies ■ Not a direct measure of benefit in all (CR) ■ Durable complete responses can cases represent clinical benefit ■ Not a comprehensive measure of drug ■ Assessed earlier and in smaller studies activity compared with survival studies ■ Small subset of patients with benefit ■ Definition of progressive disease ■ Requires prospective, consistent (PD) identifies uniform time to end definition. Meaningful response treatment and data capture durations not standardized ■ Definition of PD is arbitrary without evidence it actually represents end of benefit period Progression-free ■ Smaller sample size and shorter ■ Statistically validated as surrogate for survival (PFS) or time follow-up necessary compared with survival only in some settings to progression (TTP)a survival studies ■ Not precisely measured; subject to ■ Measurement of stable disease assessment bias particularly in open- included label studies ■ Not confounded by crossover or ■ Definitions vary among studies; little subsequent therapies agreement on magnitude of difference ■ Generally based on objective and that constitutes clinical benefit quantitative assessment ■ Requires frequent and consistent radiological or other assessments ■ Involves balanced timing of assessments among treatment arms a Progression-free survival includes all deaths; time to progression censors deaths that occur before progression. Adapted from U.S. Department of Health and Human Services, Food and Drug Administration, Center for Drug Evaluation and Research (CDER), Center for Biologics Evaluation and Research (CBER). Guidance from Industry. Clinical Trial Endpoints for the Approval of Cancer Drugs and Biologics. 2007. http://www.fda.gov/downloads/Drugs/.../Guidances/ucm071590.pdf.

tahir99 - UnitedVRG Chapter 30 Assessment of Clinical Response 315

Prespecified evaluation interval Ideally, disease progression is reported at a prespecified evaluation interval

Ascertainment (evaluation) bias An earlier evaluation leads to earlier scoring of progression (e.g., concern for symptoms prompt earlier evaluation)

Ascertainment (evaluation) bias A later evaluation leads to delay in scoring progression (e.g., evaluation delayed by toxicity or treatment delays)

Censoring bias Patient whose disease would have progressed quickly is censored early. Here censoring is “beneficial.”

Censoring bias Patient whose disease would have progressed late is censored early. Here censoring is “detrimental.”

Informative censoring Central review cannot score progression with available data. Although progression had been scored, the data is instead censored centrally. This is usually “beneficial.” = Time of actual progression = Evaluation interval;

evaluate for response/progression CANCER THERAPEUTICS = Time when censored = Median PFS/TTP = Time progression scored

Figure 30.2 The potential problems encountered when PFS is used as an endpoint. Ideally, as depicted at the top, response assessment will be conducted at a prespecified time. owever,H the date at which progression is scored may suffer from either ascertainment or censoring bias. Ascertainment bias can occur if either an evaluation occurs before the prespecified date or if it is delayed. For example, a clinician concerned about a patient who is not experiencing side effects and has likely been randomized to placebo may be more inclined to investigate symptoms early and document progression before the prespecified time, while delaying the evaluation of a patient randomized to the experimental arm who experiences some toxicity. Similarly, censoring—an increasing problem in randomized trials—may impact the outcome of a given study arm by either censoring patients who would experience early progression (beneficial impact) or censoring those who would have remained progression free for a long time (detrimental impact). Finally, informative censoring can occur when independent radiologic review cannot concur with an investigator’s assessment of progression and censors the patient. This outcome is usually beneficial, because a patient who is very close to experiencing progression is censored. (Adapted from Villaruz LC, Socinski MA. The clinical viewpoint: definitions, limitations of RECIST, practical considerations of measurement.Clin Cancer Res 2013;19:2629–2636.)

Overall Survival subsequently.65 An ITT analysis avoids the bias introduced by omitting dropouts and noncompliant patients that can negate ran- Defined as the time from randomization to death, OS has been domization and overestimate clinical effectiveness. considered the gold standard of clinical trial endpoints (see Table 30.3). In part, this is so because it is unambiguous and does not suffer from interpretation bias. An additional advantage Kaplan–Meier Plots of the survival endpoint is that it can balance the effect of thera- pies with high treatment-related mortality even if tumor control In a typical clinical trial, data are often presented as a Kaplan– is substantially better with the new treatment. However, some Meier plots. In discrete time intervals, the number of patients in worry that because patients may receive multiple lines of therapy each group who are progression free and alive (PFS analysis) or following the clinical trial, the results may be confounded by alive (OS analysis) at the end of the interval are counted and di- those subsequent therapies. The latter concern is often cited as vided by the total number of patients in that group at the beginning the reason why an advantage in PFS/TTP disappears when one of the time interval. One excludes from this calculation patients looks to OS. But as a review of clinical trials confirms,62 the mag- censored for a reason other than progressive disease or death dur- nitude of the difference does not disappear, only the statistical ing the same interval. This has the advantage that it allows one validity (Fig. 30.3).63,64 to include censored patients in estimates of the probability of When evaluating a randomized controlled trial, it is important PFS or OS up to the point when they were censored (i.e., they that the OS as well as the PFS analyses are always by intention are excluded only beyond the point of censoring). In most clinical to treat (ITT). In an ITT analysis, often described as once ran- trials, a fraction of patients are typically censored. domized, always analyzed, all patients assigned to a group at the In constructing the Kaplan–Meier plot, probabilities are cal- time of randomization are analyzed regardless of what occurred culated for each interval of time. The probability of surviving 316 Cancer Therapeutics

Figure 30.3 Hypothetical distribution of PFS and OS data demonstrating the disappearance of PFS benefit. Because chemotherapy does not exert a lasting effect on the underlying tumor biology and because PFS is a shorter interval (measured in increments, not daily as is OS) PFS differences often disappear. The hypothetical example shown illustrates this phenomenon. The left panel shows a histogram of PFS distributions with a difference of 0.34 months that nevertheless achieves statistical significance over the short interval when PFS is measured. The right panel depicts similar histograms for OS captured over a longer time period. Despite a larger absolute difference of 0.5 months, the OS difference does not reach statistical significance. For these hypothetical curves, random number generated data sets (with normal distribution), histograms, and density plots were generated using R version 2.11.1 (2010-05-31).75 The differences were deliberately chosen to be small, but a similar disappearance can also occur with larger differences. As can be seen, what disappears is not the absolute benefit, but the statistical validity.

progression free or being counted as a survivor to the end of any in- hazard ratio, the better the experimental therapy. To determine terval of assessment is the product of the probabilities of surviving whether the hazard ratio has statistical significance, one can (1) in all the preceding assessment intervals multiplied by the prob- use a log-rank test to show that the null hypothesis that the two ability for the interval of interest. One might ask to what extent treatments lead to the same survival probabilities is wrong, or (2) the two curves in each study differ. One measure that is of value is use a parametric approach writing a regression model and fitting the median PFS or OS—a value calculated in most studies from a the data to the model so that one can establish the hazard ratio for Kaplan–Meier plot. the whole trial and its statistical significance. In many cases, the Cox proportional hazard model is used. Although the ideal hazard Hazard Ratios ratio would capture the differential benefit throughout the period of study, in practice, the extremes depicted in a Kaplan–Meier plot may not be analyzed. Increasingly, however, hazard ratios are cited in preference to the more traditional measures of efficacy such as the median PFS and median OS. However, because a hazard ratio is a value that has no Forest Plots dimensions, it has very limited value, informing the reader only with regard to the reliability and uniformity of the data. It does not quan- Interest in determining whether there is heterogeneity in a tify the magnitude of the benefit. A physician and, especially, a pa- treatment effect, such that better outcomes occur in some sub- tient want to know the magnitude of the benefit (i.e., the extent to groups, has led to the use of Forest plots to display treatment which a life will be prolonged), not what a dimensionless hazard effects across subgroups. Although simple in concept, these ratio is. By definition, the hazard ratio is a ratio of the hazard plots are subject to error because subgroups are composed of rates. The hazard rate quantifies the likelihood that a patient will smaller numbers and the confidence intervals are therefore experience a hazardous event or a hazard during a defined inter- wider than those for the entire group. The most common pre- val of observation, and this is expressed as a rate or percent. For sentation includes a vertical line at the no effect point (e.g., a example, if during a given period of observation 20 of 100 patients hazard ratio of 1.0), with symbols of varying size representing receiving a reference or control therapy experience progression the subgroups, each with its confidence interval depicted by a or death, their hazard rate during this interval is 0.2 (20/100). If line that stretches from the symbol to both sides (the symbol during this same interval, only 10 of the 100 patients receiving the size is usually proportional to the size of the subgroup). If the experimental therapy experience progression or death, their haz- confidence interval for a subgroup crosses the no effect point, ard rate is 0.1 (10/100). In this simple example, the hazard ratio this is commonly interpreted (not necessarily correctly) as a lack for the interval, calculated as the ratio of the hazard rates is 0.5 of effect in the subgroup. The information one seeks from a For- (0.1/0.2) and indicates the likelihood of experiencing a hazardous est plot is whether the effect size for different subgroups varies event is reduced by 50% in the experimental arm. As commonly significantly from the main effect, which is determined by a test presented, and as this simple example illustrates, the lower the for heterogeneity.66

tahir99 - UnitedVRG Chapter 30 Assessment of Clinical Response 317

0.40 PD 0.20 0.00 SD –0.20 –0.40

Percent Change –0.60 PR –0.80 –1.00

Figure 30.4 Example of a waterfall plot demonstrating for each patient the maximum benefit obtained with the study therapy. Those to the left represent patients whose tumors increased, and those on the right represent patients whose tumors regressed. The vertical red lines at +20% and -30% define the boundaries of stable disease according to RECIST. Ideally, all responses should be confirmed after a period of at least 4 weeks. The example shown is of patients with renal cell carcinoma treated with the microtubule targeting agent ixabepilone. (From Huang H, Menefee M, Edgerly M, et al. A phase II clinical trial of ixabepilone [Ixempra; BMS-247550; NSC 710428], an epothilone B analog, in patients with metastatic renal cell carcinoma. Clin Cancer Res 2010;16:1634–1641.)

Beyond Dichotomized Data related outcomes, often neglecting assessments of QOL. Although a QOL assessment in clinical settings is possible with currently Quality of Life available instruments, there must be continued development and refinement of these instruments. Such development must focus not The assessment of cancer patients enrolled on a clinical trial can be only on extracting valuable information in an unbiased manner, but said to consist of two sets of endpoints: cancer outcomes and patient also and equally important, developing an instrument that is user outcomes. Cancer outcomes measure the response of the tumor to friendly and will be completed in a high percentage of encounters. treatment, the duration of the response, the symptom-free period, and the early recognition of relapse. In contrast, patient outcomes Waterfall Plots CANCER THERAPEUTICS assess the benefit achieved with a given therapy bymeasuring the increase in survival and the quality of life (QOL) before and after The arbitrary nature of the 50% cutoff set by Moertel and Han- therapy. Unfortunately, physicians tend to concentrate on cancer- ley and its evolution to the current RECIST threshold of 30%

AB

Figure 30.5 The effect of the growth rate constant, g, on two commonly reported clinical values: maximum tumor shrinkage and PFS. Tumor measurements obtained in patients can be analyzed mathematically. (A–E) The black CDline depicts idealized clinical data using tumor quantities measured as patients received chemotherapy. Actually, clinical measurements comprise concurrent tumor regression (dashed red line) and growth (dashed blue line) that can be described by a rate constant and a first order kinetic equation, f(t) = exp(-d∙t) + exp(g∙t) − 1, where exp is the base of the natural logarithm, e = 2.7182. . . , and f is the percent change in tumor measurement at time t, normalized to the value when treatment began. The rate constant d accounts for exponential decrease, whereas the rate constant g accounts for exponential growth occurring during EFtreatment.68,69 To demonstrate the correlation between the growth rate, tumor shrinkage and PFS, the same regression rate (d ) has been modeled in panels A through E, whereas the growth rate constant, g, increases in each successive panel. The black triangles depict the point at which tumor size is 20% above the nadir (RECIST definition of PD). As the growth rate increases (i.e., faster tumor growth) from A to E, the nadir is reached sooner, and the depth of the nadir is less. (F) The correlation between PFS and maximum tumor shrinkage (nadir) is shown, plotting the correlation between PFS and response fraction, which is defined as the ratio of nadir to initial value.68,69 Although idealized plots are given, the curves are based firmly on data obtained from patients enrolled on clinical trials. 318 Cancer Therapeutics reduction in the size of the maximum diameter raises valid que- exponential tumor shrinkage after treatment, followed by tumor ries as to why 30% is valuable and not 29% or 25%. On this back- regrowth that is either exponential or linear and have been shown ground, waterfall plots such as the one shown in Figure 30.4 have to correlate with OS and to discriminate effective therapies as well become increasingly popular because they depict the benefit or as individual patients within trials.68–73 A major advantage is that lack thereof in all patients as a continuum of response, rather than more of the data are used, relative to dichotomized response as- a dichotomized response rate.67 Waterfall plots can be generated sessment, and regression or growth rates can be determined even from any quantitative assessment. If ctDNA or tumor cells prove to in patients who are censored in a Kaplan–Meier analysis. Equa- be as quantitative as hoped, the maximum decline could be plot- tions that model both regression and growth rates confirm the ted as a waterfall plot. clinical intuition that resistant disease is emerging even as overall tumor volume is reduced. Further, “the strategy of studying tumor Growth Kinetics growth kinetics circumvents one weakness of ‘progression crite- ria,’ which is that they inherently dichotomize a complex biologi- Efforts to quantify tumor kinetic parameters from clinical data cal process that may be better characterized using a continuous have been investigated in recent years. Different equations have function.”74 As shown in Figure 30.5, the response of a tumor to a been applied to describe the two-phase curve based on tumor therapy is exemplified by the nadir, the time to the nadir, and the size as observed in most solid tumor trials, where there is first time to progression or PFS, and these are all are all dependent on shrinkage followed by regrowth (Fig. 30.5). These models show the growth rate.

REFERENCES

1. America’s Biopharmaceutical Research Companies. Medicines in Develop- 20. von Minckwitz G, Untch M, Blohmer JU, et al. Definition and impact ment for Cancer. PhRMA Web site. http://www.phrma.org/sites/default/files/ of pathologic complete response on prognosis after neoadjuvant chemo- pdf/phrmamedicinesindevelopmentcancer2012.pdf. therapy in various intrinsic breast cancer subtypes. J Clin Oncol 2012;30: 2. Moertel CG, Hanley JA. The effect of measuring error on the results of thera- 1796–1804. peutic trials in advanced cancer. Cancer 1976;38:388–394. 21. Cortazar P, Zhang L, Untch M, et al. Pathological complete response and 3. Miller AB, Hoogstraten B, Staquet M, et al. Reporting results of cancer treat- long-term clinical benefit in breast cancer: the CTNeoBC pooled analysis. ment. Cancer 1981;47:207–214. Lancet 2014 [Epub ahead of print]. 4. Therasse P, Arbuck SG, Eisenhauer EA, et al. New guidelines to evaluate the 22. Bardia A, Baselga J. Neoadjuvant therapy as a platform for drug development response to treatment in solid tumors. European Organization for Research and approval in breast cancer. Clin Cancer Res 2013;19:6360–6370. and Treatment of Cancer, National Cancer Institute of the United States, 23. Choi H, Charnsangavej C, Faria SC, et al. Correlation of computed National Cancer Institute of Canada. J Natl Cancer Inst 2000;92:205–216. tomography and positron emission tomography in patients with metastatic 5. Eisenhauer EA, Therasse P, Bogaerts J, et al. New response evaluation crite- gastrointestinal stromal tumor treated at a single institution with imatinib ria in solid tumours: revised RECIST guideline (version 1.1). Eur J Cancer mesylate: proposal of new computed tomography response criteria. J Clin 2009;45:228–247. Oncol 2007;25:1753–1759. 6. Mazumdar M, Smith A, Schwartz LH. A statistical simulation study finds dis- 24. Schramm N, Englhart E, Schlemmer M, et al. Tumor response and clini- cordance between WHO criteria and RECIST guideline. J Clin Epidemiol cal outcome in metastatic gastrointestinal stromal tumors under sunitinib 2004;57:358–365. therapy: comparison of RECIST, Choi and volumetric criteria. Eur J Radiol 7. Wen PY, Macdonald DR, Reardon DA, et al. Updated response assessment 2013;82:951–958. criteria for high-grade gliomas: response assessment in neuro-oncology work- 25. Dudeck O, Zeile M, Reichardt P, et al. Comparison of RECIST and Choi ing group. J Clin Oncol 2010;28:1963–1972. criteria for computed tomographic response evaluation in patients with 8. Byrne MJ, Nowak AK. Modified RECIST criteria for assessment of response advanced gastrointestinal stromal tumor treated with sunitinib. Ann Oncol in malignant pleural mesothelioma. Ann Oncol 2004;15:257–260. 2011;22:1828–1833. 9. Wolchok JD, Hoos A, O’Day S, et al. Guidelines for the evaluation of 26. Ronot M, Bouattour M, Wassermann J, et al. Alternative response criteria immune therapy activity in solid tumors: immune-related response criteria. (Choi, European Association for the Study of the Liver, and Modified Re- Clin Cancer Res 2009;15:7412–7420. sponse Evaluation Criteria in Solid Tumors [RECIST]) versus RECIST 1.1 10. Quant EC, Wen PY. Response assessment in neuro-oncology. Curr Oncol in patients with advanced hepatocellular carcinoma treated with sorafenib. Rep 2011;13:50–56. Oncologist 2014. http://prostatecancer.theoncologist.com/article/alternative- 11. Hawkins-Daarud A, Rockne RC, Anderson AR, et al. Modeling tumor- response-criteria-choi-european-association-study-liver-and-modified- associated edema in gliomas during anti-angiogenic therapy and its impact response. on imageable tumor. Front Oncol 2013;3:66. 27. van der Veldt AA, Meijerink MR, van den Eertwegh AJ, et al. Choi response 12. Fink J, Born D, Chamberlain MC. Pseudoprogression: relevance with re- criteria for early prediction of clinical outcome in patients with metastatic spect to treatment of high-grade gliomas. Curr Treat Options Oncol 2011; renal cell cancer treated with sunitinib. Br J Cancer 2010;102:803–809. 12:240–252. 28. Wahl RL, Jacene H, Kasamon Y, et al. From RECIST to PERCIST: evolv- 13. Lin NU, Lee EQ, Aoyama H, et al. Challenges relating to solid tumour brain ing considerations for PET response criteria in solid tumors. J Nucl Med metastases in clinical trials, part 1: patient population, response, and progres- 2009;50:122S–150S. sion. A report from the RANO group. Lancet Oncol 2013;14:e396–e406. 29. Shankar LK, Hoffman JM, Bacharach S, et al. Consensus recommendations 14. Mabille M, Vanel D, Albiter M, et al. Follow-up of hepatic and peritoneal for the use of 18F-FDG PET as an indicator of therapeutic response in pa- metastases of gastrointestinal tumors (GIST) under Imatinib therapy requires tients in National Cancer Institute Trials. J Nucl Med 2006;47:1059–1066. different criteria of radiological evaluation (size is not everything!!!). Eur J 30. Young H, Baum R, Cremerius U, et al. Measurement of clinical and sub- Radiol 2009;69:204–208. clinical tumour response using [18F]-fluorodeoxyglucose and positron emis- 15. Liu L, Wang W, Chen H, et al. EASL- and mRECIST-evaluated responses sion tomography: review and 1999 EORTC recommendations. European to combination therapy of sorafenib with transarterial chemoembolization Organization for Research and Treatment of Cancer (EORTC) PET Study predict survival in patients with hepatocellular carcinoma. Clin Cancer Res Group. Eur J Cancer 1999;35:1773–1782. 2014; 20:1623–1631. 31. Kramer-Marek G, Capala J. Can PET imaging facilitate optimization of 16. Cheson BD, Pfistner B, Juweid ME, et al. Revised response criteria for malig- cancer therapies? Curr Pharm Des 2012;18:2657–2669. nant lymphoma. J Clin Oncol 2007;25:579–586. 32. Niederkohr RD, Greenspan BS, Prior JO, et al. Reporting guidance for onco- 17. Bernhard J, Dietrich D, Scheithauer W, et al. Clinical benefit and quality logic 18F-FDG PET/CT imaging. J Nucl Med 2013;54:756–761. of life in patients with advanced pancreatic cancer receiving gemcitabine 33. Liu Y, Litière S, de Vries EG, et al. The role of response evaluation criteria plus capecitabine versus gemcitabine alone: a randomized multicenter phase in solid tumour in anticancer treatment evaluation: results of a survey in the III clinical trial—SAKK 44/00-CECOG/PAN.1.3.001. J Clin Oncol 2008; oncology community. Eur J Cancer 2014;50:260–266. 26:3695–3701. 34. Rubin EH, Allen JD, Nowak JA, et al. Developing precision medicine in a 18. Burris HA, Moore MJ, Andersen J, et al. Improvements in survival and clini- global world. Clin Cancer Res 2014;20:1419–1427. cal benefit with gemcitabine as first-line therapy for patients with advanced 35. Parkinson DR, McCormack RT, Keating SM. Evidence of clinical utility: an pancreas cancer: a randomized trial. J Clin Oncol 1997;15:2403–2413. unmet need in molecular diagnostics for cancer patients. Clin Cancer Res 19. Mann BS, Johnson JR, He K, et al. Vorinostat for treatment of cutaneous 2014;20:1428–1444. manifestations of advanced primary cutaneous T-cell lymphoma. Clin 36. Buyse M, Sargent DJ, Grothey A, et al. Biomarkers and surrogate end points— Cancer Res 2007;13:2318–2322. the challenge of statistical validation. Nat Rev Clin Oncol 2010;7:309–317.

tahir99 - UnitedVRG Chapter 30 Assessment of Clinical Response 319

37. Karam AK, Karlan BY. Ovarian cancer: the duplicity of CA125 measure- 57. Bruzzi P, Del Mastro L, Sormani MP, et al. Objective response to chemo- ment. Nat Rev Clin Oncol 2010;7:335–339. therapy as a potential surrogate end point of survival in metastatic breast can- 38. Rustin GJ, van der Burg ME, Griffin CL, et al. Early versus delayed treatment cer patients. J Clin Oncol 2005;23:5117–5125. of relapsed ovarian cancer (MRC OV05/EORTC 55955): a randomised trial. 58. Vidaurre T, Wilkerson J, Simon R, et al. Stable disease is not preferentially Lancet 2010;376:1155–1163. observed with targeted therapies and as currently defined has limited value 39. Vergote I, Rustin GJ, Eisenhauer EA, et al. Re: new guidelines to evaluate the in drug development. Cancer J 2009;15:366–373. response to treatment in solid tumors [ovarian cancer]. Gynecologic Cancer 59. McKee AE, Farrell AT, Pazdur R, et al. The role of the U.S. Food and Drug Intergroup. J Natl Cancer Inst 2000;92:1534–1535. Administration review process: clinical trial endpoints in oncology. Oncolo- 40. Guppy AE, Rustin GJ. CA125 response: can it replace the traditional gist 2010;15:13–18. response criteria in ovarian cancer? Oncologist 2002;7:437–443. 60. Ohorodnyk P, Eisenhauer EA, Booth CM. Clinical benefit in oncology 41. Rustin GJ, Quinn M, Thigpen T, et al. Re: New guidelines to evaluate the trials: is this a patient-centred or tumour-centred end-point? Eur J Cancer response to treatment in solid tumors (ovarian cancer). J Natl Cancer Inst 2009;45:2249–2252. 2004;96:487–488. 61. Buyse M. Use of meta-analysis for the validation of surrogate endpoints and 42. Rustin GJ, Vergote I, Eisenhauer E, et al. Definitions for response and pro- biomarkers in cancer trials. Cancer J 2009;15:421–425. gression in ovarian cancer clinical trials incorporating RECIST 1.1 and CA 62. Wilkerson J, Fojo T. Progression-free survival is simply a measure of a drug’s 125 agreed by the Gynecological Cancer Intergroup (GCIG). Int J Gynecol effect while administered and is not a surrogate for overall survival. Cancer J Cancer 2011;21:419–423. 2009;15:379–385. 43. Eisenhauer EA. Optimal assessment of response in ovarian cancer. Ann On- 63. Reck M, von Pawel J, Zatloukal P, et al. Overall survival with cisplatin-gem- col 2011;22:viii49–viii51. citabine and bevacizumab or placebo as first-line therapy for nonsquamous 44. Bubley GJ, Carducci M, Dahut W, et al. Eligibility and response guidelines non-small-cell lung cancer: results from a randomised phase III trial (AVAiL). for phase II clinical trials in androgen-independent prostate cancer: recom- Ann Oncol 2010;21:1804–1809. mendations from the Prostate-Specific Antigen Working Group. J Clin Oncol 64. Hortobagyi GN, Gomez HL, Li RK, et al. Analysis of overall survival from 1999;17:3461–3467. a phase III study of ixabepilone plus capecitabine versus capecitabine in pa- 45. Scher HI, Halabi S, Tannock I, et al. Design and end points of clinical tri- tients with MBC resistant to anthracyclines and taxanes. Breast Cancer Res als for patients with progressive prostate cancer and castrate levels of testos- Treat 2010;122:409–418. terone: recommendations of the Prostate Cancer Clinical Trials Working 65. Hennekens C, Buring J. Epidemiology in Medicine. 1st ed. Boston: Little, Group. J Clin Oncol 2008;26:1148–1159. Brown and Co.; 1987. 46. Mazumdar M, Bajorin DF, Bacik J, et al. Predicting outcome to chemo- 66. Cuzick J. Forest plots and the interpretation of subgroups. Lancet 2005; therapy in patients with germ cell tumors: the value of the rate of decline 365:1308. of human chorionic gonadotrophin and alpha-fetoprotein during therapy. 67. Huang H, Menefee M, Edgerly M, et al. A phase II clinical trial of ixa- J Clin Oncol 2001;19:2534–2541. bepilone (Ixempra; BMS-247550; NSC 710428), an epothilone B analog, 47. Fizazi K, Culine S, Kramar A, et al. Early predicted time to normalization of in patients with metastatic renal cell carcinoma. Clin Cancer Res 2010;16: tumor markers predicts outcome in poor-prognosis nonseminomatous germ 1634–1641. cell tumors. J Clin Oncol 2004;22:3868–3876. 68. Stein WD, Gulley JL, Schlom J, et al. Tumor regression and growth rates 48. Toner GC. Early identification of therapeutic failure in nonseminomatous determined in five intramural NCI prostate cancer trials: the growth rate germ cell tumors by assessing serum tumor marker decline during che- constant as an indicator of therapeutic efficacy. Clin Cancer Res 2011;17: motherapy: still not ready for routine clinical use. J Clin Oncol 2004;22: 907–917. 3842–3845. 69. Stein WD, Wilkerson J, Kim ST, et al. Analyzing the pivotal trial that com- 49. Gilligan TD, Seidenfeld J, Basch EM, et al. American Society of Clinical pared sunitinib and IFN-α in renal cell carcinoma, using a method that as- CANCER THERAPEUTICS Oncology Clinical Practice Guideline on uses of serum tumor markers in sesses tumor regression and growth. Clin Cancer Res 2012;18:2374–2381. adult males with germ cell tumors. J Clin Oncol 2010;28:3388–3404. 70. Maitland ML, Wu K, Sharma MR, et al. Estimation of renal cell carcinoma 50. Albers P, Albrecht W, Algaba F, et al. EAU guidelines on testicular cancer: treatment effects from disease progression modeling. Clin Pharmacol Ther 2011 update. Eur Urol 2011;60:304–319. 2013;93:345–351. 51. Yap T, Lorente D, Omlin A, et al. Circulating tumor cells: a multifunctional 71. Claret L, Girard P, Hoff PM, et al. Model-based prediction of phase III over- biomarker. Clin Cancer Res 2014;20:2553–2568. all survival in colorectal cancer on the basis of phase II tumor dynamics. 52. Dawson SJ, Tsui DW, Murtaza M, et al. Analysis of circulating tumor DNA J Clin Oncol 2009;27:4103–4108. to monitor metastatic breast cancer. N Engl J Med 2013;368:1199–1209. 72. Claret L, Gupta M, Han K, et al. Evaluation of tumor-size response met- 53. Punnoose EA, Atwal S, Liu W, et al. Evaluation of circulating tumor cells rics to predict overall survival in Western and Chinese patients with first-line and circulating tumor DNA in non-small cell lung cancer: association with metastatic colorectal cancer. J Clin Oncol 2013;31:2110–2114. clinical endpoints in a phase II clinical trial of pertuzumab and erlotinib. 73. Wang Y, Sung C, Dartois C, et al. Elucidation of relationship between tumor Clin Cancer Res 2012;18:2391–2401. size and survival in non-small-cell lung cancer patients can aid early deci- 54. Bettegowda C, Sausen M, Leary RJ, et al. Detection of circulating tu- sion making in clinical drug development. Clin Pharmacol Ther 2009;86: mor DNA in early- and late-stage human malignancies. Sci Transl Med 167–174. 2014;6:224ra24. 74. Oxnard GR, Morris MJ, Hodi FS, et al. When progressive disease does not 55. Pazdur R. Endpoints for assessing drug activity in clinical trials. Oncologist mean treatment failure: reconsidering the criteria for progression. J Natl 2008;13:19–21. Cancer Inst 2012;104:1534–1541. 56. Buyse M, Thirion P, Carlson RW, et al. Relation between tumour response to 75. Team RDC. R: A language and environment for statistical computing. R first-line chemotherapy and survival in advanced colorectal cancer: a meta- Foundation for Statistical Computing. Vienna, Austria: R Foundation for analysis. Meta-Analysis Group in Cancer. Lancet 2000;356:373–378. Statistical Computing, 2010. http://www.r-project.org. tahir99 - UnitedVRG PART IV

Cancer Prevention and Screening Tobacco Use and the Cancer 31Patient

Graham W. Warren, Benjamin A. Toll, Irene M. Tamí-Maury, and Ellen R. Gritz

INTRODUCTION system,8,9 resulting in nicotine’s rewarding effect experienced by tobacco users.10–12 Dopaminergic neurotransmission may also be Tobacco is commonly described as the largest preventable cause of involved in the assignment of incentive salience, or stimulus for cancer. Over 50 years ago, tobacco was increasingly recognized as the a pleasure based reward, to tobacco use–related environmental primary cause of lung cancer, with definitive recognition for tobacco cues13,14 that may become conditioned reinforcers of tobacco use use as a causative factor in the seminal 1964 U.S. Surgeon General’s behaviors. For example, an individual who smokes while drinking Report (SGR) on Smoking and Health.1 Recent editions of the SGR their morning coffee may associate coffee, or even holding a coffee have described the widespread adverse health effects of tobacco on cup in their hand, with the reward from smoking. Thus, cigarette a spectrum of diseases, including as a causative agent for a spectrum smoking is directly linked to external nontobacco-based behavioral of cancers.2,3 Tobacco use is an addiction usually initiated in youth stimuli. Activation of the nucleus accumbens has further been im- prior to the age of 18 and is driven by the highly addictive drug, nico- plicated in drug reinstatement or relapse.15,16 Individuals who have tine.4 As related to the cancer patient, considerable work has been quit tobacco use for years have restarted a tobacco habit simply by conducted to associate tobacco use with the risk of developing can- sitting next to a smoker and being exposed to secondhand smoke. cer and how tobacco cessation can substantially reduce cancer risks. Substantial work has been conducted on the addictive nature of However, there is a relative paucity of effort that has been put forth to tobacco and nicotine, and readers are referred to several compre- identify the effects of smoking on outcomes for cancer patients or to hensive reviews on this topic.9,12,17 establish methods to help cancer patients quit smoking. Fortunately, in recent years, the importance of tobacco use by the cancer patient has been increasingly recognized as an important health behavior, TOBACCO USE PREVALENCE AND THE including a National Cancer Institute (NCI)–sponsored conference EVOLUTION OF TOBACCO PRODUCTS on tobacco use in 2010, a joint sponsored NCI–American Associa- tion of Cancer Research (AACR)–sponsored workshop at the In- Much of the discussion on tobacco use epidemiology and carci- stitute of Medicine in 2012, and recent recommendations by the nogenesis is presented in Chapter 4. In brief, the prevalence of AACR and the American Society of Clinical Oncology (ASCO) to cigarette smoking among adults in the United States decreased to 5,6 address tobacco use in cancer patients. The recently released 2014 19.0% as compared with 22.8% in 2001, but it did not meet the SGR now provides substantial evidence behind the effects of smok- Healthy People 2010 objective to reduce smoking prevalence to 7 ing by cancer patients with the following conclusions : 12%.18,19 There have been substantial changes in the landscape of 1. In cancer patients and survivors, the evidence is sufficient to tobacco use over time as a direct consequence of cigarette-centered policies and regulations aiming to reduce the harmful effects and infer a causal relationship between cigarette smoking and ad- 20–22 verse health outcomes. Quitting smoking improves the progno- number of deaths caused by smoking. Under this new land- sis of cancer patients. scape, novel and reemergent noncigarette tobacco products such 2. In cancer patients and survivors, the evidence is sufficient to as cigars, cigarillos, snuff, chewing tobacco, water pipes (hookahs), infer a causal relationship between cigarette smoking and in- and other forms of tobacco consumption have been growing in demand as a consequence of aggressive and sophisticated market- creased all-cause mortality and cancer-specific mortality. 23 3. In cancer patients and survivors, the evidence is sufficient to ing by the tobacco industry. Consumption patterns have also infer a causal relationship between cigarette smoking and in- changed due to efforts by the tobacco industry to make cigarettes appear safer, such as low tar or filtered cigarettes, and the inclusion creased risk for second primary cancers known to be caused by 24 cigarette smoking, such as lung cancer. of flavoring (menthol, vanilla, fruits, etc.). Although these efforts 4. In cancer patients and survivors, the evidence is suggestive but may have changed consumption patterns, they have not reduced not sufficient to infer a causal relationship between cigarette cancer risk. Large patient cohorts demonstrate that the introduc- smoking and the risk of recurrence, poorer response to treat- tion of low tar and filtered cigarettes actually increased risk by pro- ment, and increased treatment-related toxicity. moting deeper inhalation and higher rates of addiction with no reductions in cancer risk,24,25 resulting in subsequent changes in The overall objective of this chapter is to discuss tobacco use by can- lung cancer from centrally located squamous cell cancers to pe- cer patients, the clinical effects of smoking in cancer patients, methods ripherally located nonsquamous cell cancers. to address tobacco use by cancer patients, and areas of needed research. The relatively recent introduction of electronic cigarettes (i.e., e-cigarettes, e-cigs, nicotine vaporizers, or electronic nicotine de- livery systems [ENDS]) is noteworthy. These electronic or battery- NEUROBIOLOGY OF TOBACCO powered devices activate a heating element that vaporizes a liquid DEPENDENCE solution contained in a cartridge, and then the user inhales this vapor. Levels of nicotine as well as other chemical additives and Nicotine is the primary addictive component of tobacco that flavors in the cartridge are uncertain and vary according to the increases extracellular concentrations of dopamine in the nu- brand.26 Although there are no research studies that have evalu- cleus accumbens and stimulates the mesolimbic dopaminergic ated the potential harmful effects of the use of e-cigarettes for

322

tahir99 - UnitedVRG Chapter 31 Tobacco Use and the Cancer Patient 323 cancer patients,27 organizations such as the World Health Orga- before diagnosis, within the 12 months prior to diagnosis, after di- nization have already expressed concerns about the safety of these agnosis, within the past 10 years), former (e.g., recent, intermedi- increasingly popular products.28,29 To date, e-cigarettes have not ate-, or long-term quit for 1 month, 3 month, 6 month, 12 month, been approved by the U.S. Food and Drug Administration (FDA) 2 years, 5 years, 10 years), never, quitting after diagnosis, and ac- as therapeutic devices to aid in quitting smoking.26 Readers are cording to exposure (e.g., multiple pack year cutoffs, Brinkman referred to a recent editorial on the use of e-cigarettes by cancer index, years of smoking, years of smoking within a predefined pe- patients27; however, it will likely be several years before evidence- riod of time such as 5 years prior to diagnosis). Though the non- based health information is available. standard method of addressing tobacco use in cancer patients has been observed in several reports,54–57 there are no current standard recommendations for the definition of tobacco use by any national TOBACCO USE BY THE CANCER PATIENT organization. There are four primary categories for smoking status: 1. Never smoking is typically defined as having smoked less than The prevalence of current smoking among long-term adult can- 100 cigarettes in a person’s lifetime and no current cigarette cer survivors appears to have declined in the past decade,30 but use. These patients are generally considered as a reference data suggest higher rates of smoking among cancer survivors than group in many studies. Categories 2 through 4 require that a in the general population.30–32 These data are often biased by the person has smoked at least 100 cigarettes in their lifetime. fact that assessments in cancer patients may not include cancer 2. Former smoking is typically defined as no current cigarette patients who were current smokers at the time of death. As a result, use, usually within the past year. estimates of smoking rates in cancer survivors may be mislead- 3. Recent smoking (or recent quit) is generally defined as having ing and may underestimate true tobacco use patterns for cancer stopped smoking within the recent past, typically for a period of patients. Furthermore, alternative tobacco products are often not 1 week to 1 year. assessed in cancer patients. Data from the Childhood Cancer Sur- 4. Current smoking is typically defined as smoking one or more vivor Study and the 2009 Behavioral Risk Factor Surveillance Sys- cigarettes per day every day or some days. tem indicate that approximately 3% to 8% of cancer survivors use smokeless tobacco products.33,34 Patients may be attracted to these Ever smoking is a combination of categories 2 through 4 (i.e., alternative products due to less social stigma and the nonevidence- former, recent, and current smokers) that has been used to report based perception that these products are healthier alternatives negative associations between smoking and cancer outcomes in a compared to cigarette smoking. number of studies.58–70 Defining smoking according toever smok- Continued tobacco use by cancer patients often represents ing status limits the ability to interpret the effects of current smok- a combined failure by the patient to recognize the need to stop ing on a clinical outcome, and nothing can be done to address a smoking even after a cancer diagnosis and the effort by health-care prior tobacco use history. However, defining exposure according to providers to address tobacco use with evidence-based assessments current smoking status allows for the analysis of potentially revers- and tobacco cessation support. Approximately 30% of all cancer ible effects as well as for the potential implementation of smoking patients use tobacco at the time of cancer diagnosis with higher cessation to prevent the adverse outcomes of smoking on cancer rates in traditionally tobacco-related disease sites, such as head and patients. The primary focus for the remainder of this chapter will AND SCREENING CANCER PREVENTION neck or lung cancers, and lower rates in traditionally nontobacco- be on current smoking and will include a discussion of methods related disease sites, such as breast or prostate cancers.35–44 How- to address tobacco use with the cancer patient through accurate ever, findings from several studies indicate that cancer patients are assessments and structured tobacco cessation support. receptive to smoking cessation interventions even as they continue to smoke.35,38,45–50 A cancer diagnosis can be used as a window of opportunity, THE CLINICAL EFFECTS OF SMOKING ON or teachable moment, to intervene and provide assistance in the THE CANCER PATIENT quitting process.51 A recent study in 12,000 cancer patients, in- cluding 2,700 patients who smoked, capitalized on the teachable Cancer treatment is generally defined according to disease site, moment and demonstrated that less than 3% of patients who were stage, treatment type (e.g., surgery, chemotherapy [CT], radio- contacted by the cessation program rejected tobacco cessation as- therapy [RT], or biologic therapy), and primary treatment ob- sistance.45 However, only 1.2% of patients who received a mailed jective, such as cure or palliation. A comprehensive discussion invitation participated in the program. This highlights the idea that of the effects of smoking on cancer patients is beyond the scope patients may be interested in quitting, but methods such as mailed of a single chapter, but the 2014 SGR provides an excellent evi- tobacco cessation information may not yield effective participation dence base, concluding that “the evidence is sufficient to infer a by cancer patients. Once enrolled, patients and clinicians must causal relationship between cigarette smoking and adverse health realize that although relapses in the general population usually outcomes.”7 Overall, approximately 75% to 80% of studies in the occur within 1 week of cessation, relapses in cancer patients may SGR demonstrated a negative association between smoking and be delayed due to cancer treatment–related variables such as surgi- outcome, with approximately 65% to 70% of studies demonstrat- cal or other posttreatment healing.52 Consequently, it is important ing statistically significant negative associations. This chapter will to continue offering tobacco assessments and cessation support for provide an illustrative review of studies that demonstrate the ad- cancer survivorship efforts. verse effects of tobacco across disease sites and treatment modali- ties (e.g., surgery, CT, RT), and effects will be discussed across the Defining Tobacco Use by the Cancer Patient categories of mortality, recurrence and cancer-related mortality, tox- icity, and risk of a second primary cancer. Evidence for the benefits In dealing with tobacco use by cancer patients, it is important of smoking cessation will also be presented within each section. to note that virtually all of the evidence associating tobacco with cancer treatment outcomes deals with smoking. Few studies report The Effect of Smoking on Overall Mortality associations between other forms of tobacco use (e.g., smokeless, cigars, cigarillos) and outcomes in cancer patients. Furthermore, Substantial evidence demonstrates that current smoking by can- the definition of smoking across published studies varies substan- cer patients increases the risk of overall mortality across virtually tially.53 In studies of cancer patients, smoking has been defined as all cancer disease sites and for all treatment modalities. Currently current (e.g., smoking after diagnosis, at diagnosis, in the weeks smoking significantly increased the risk of overall mortality by 324 Cancer Prevention and Screening

between 17% to 38% as compared with never, former, and re- gastrointestinal cancers,82,120,121 prostate cancer,41,84,122 gyneco- cent quit smokers in a large cohort of patients across 13 disease logic cancers,89,90,106,123–125 and lung cancer.126 Cancer recurrence, sites.71 Similar but larger observations were noted in elderly cur- whether local or metastatic, is a key driver behind cancer-related rent smokers from a separate cohort (hazard ratio [HR], 1.72, 95% mortality. Several studies demonstrate that current smoking in- confidence interval [CI], 1.23 to 2.42).72 A large analysis of over creases the risk of recurrence and decreases response across multi- 20,000 patients treated with surgery demonstrated that current ple disease sites.76,84,107,127,128 The effects of smoking on increasing smoking increased mortality by 62% in gastrointestinal cancer recurrence or cancer-related mortality have also been reported in patients and by 50% in thoracic cancer patients with a nonsig- several relatively rare cancers.120,129 In a remarkable report of pa- nificant trend in urologic cancer patients.73 Several larger studies tients with recurrent head and neck cancers treated with salvage with at least 500 patients demonstrated that current smoking in- surgery, continued smoking after salvage treatment continued to creases mortality in head and neck cancer,74–77 breast cancer,78–81 increase the risk of yet another recurrence by 42%.130 The striking gastrointestinal cancers,82,83 prostate cancer,84–87 renal cancer,88 nature of this last study highlights the continued risks even in re- gynecologic cancers,89,90 and lung cancer.91–102 Smaller studies current cancer patients and the resilience with which some cancer demonstrate similar effects for hematolymphoid cancers such as patients will continue to smoke. leukemia and lymphoma.103,104 Studies suggest that the effects The effects of smoking are also noted in premalignant lesions. of current smoking on mortality may be dose and time depend- In patients with high-grade vulvar intraepithelial neoplasia, cur- ent, with higher risks in heavier smokers105,106 and lesser risks in rent smoking increased the risk of persistent disease after therapy patients whose time since quitting was longer.105 by 30-fold.131 In a prospective trial of progesterone to treat cervi- Whereas many reports rely on retrospective chart reviews, sev- cal intraepithelial neoplasia (CIN), current smoking increased the eral prospective studies demonstrate that current smoking increases risk of progression as compared with former and never smokers mortality.71 Browman et al.107 was one of the first prospective stud- combined.132 A prospective trial of 516 low-grade cervical intraepi- ies to demonstrate that current smoking increased mortality by 2.3- thelial neoplasia patients demonstrated that current smoking de- fold in patients who continued to smoke during RT as compared creased response by 36%, although a similar effect was also noted with nonsmokers. Results from Radiation Therapy Oncology in former smokers.133 Group (RTOG) 9003 and 0129 cooperative group trials demon- As noted with overall mortality, several studies demonstrated strated that current smoking increased mortality in advanced head that the effects of current smoking are worse than the effects of and neck cancer patients treated with RT or concurrent chemo- former smoking76,86,89,109,127,134–136,137 and that the effects of smok- radiotherapy (CRT),108 with a similar effect noted in 165 cervical ing may be acutely reversible. Several studies also demonstrate that cancer patients treated with CRT.109 In the randomized current smoking increases recurrence or cancer mortality, whereas chemoprevention trial of 1,190 early stage head and neck cancer former smoking has no significant effect.41,78,82,84,85,119,122–124,138 patients, current smoking increased mortality by 2.5-fold.110 The acutely reversible effects of smoking were shown by Browman Numerous studies have demonstrated that current smoking et al.139 who demonstrated that continued smoking increased the increases overall mortality as compared with former and never risk of cancer-related mortality by 23% as compared with patients smokers combined.72,75,76,101,102,107,108 The adverse effects of smok- who quit within 12 weeks of starting RT. In 284 colorectal cancer ing compared with former and never smokers not only reflect the patients, smoking at the first postoperative visit increased the risk negative effects of smoking on mortality as a whole, but also dem- of cancer mortality by 2.5-fold as compared with all other patients onstrate that the effects of smoking are reversible. Current smoking suggesting that smoking after treatment significantly predict for increased mortality risk as compared with patients who quit within adverse outcome.121 In a notable study of over 1,400 prostate can- the year71 or 1 to 3 months prior to diagnosis.111,112 Furthermore, in cer patients treated with surgery, continued smoking 1 year after 284 limited-stage small-cell lung cancer patients, patients who quit treatment increased the risk of recurrence 2.3-fold, but quitting smoking at or following a cancer diagnosis had a 45% reduction in smoking 1 year after treatment did not confer an increased risk of mortality as compared with current smokers.113 These studies sug- recurrence.128 Chen et al.138 demonstrate that patients who con- gest that the effects of smoking on mortality are reversible. tinue to smoke before and following a bladder cancer diagnosis Collectively, these studies provide significant data associat- have an increased risk of recurrence as compared with patients ing current smoking with increased overall mortality across most who quit in the year prior to diagnosis or within the first 3 months disease sites, tumor stages, treatment modalities, and in both tra- after diagnosis. The reversible effects of smoking on recurrence ditionally tobacco-related as well as nontobacco-related cancers. and mortality are consistent with observations on overall mortality The potential significance of smoking is perhaps best exemplified and continue to emphasize the benefit of tobacco cessation for by Bittner et al.,114 who analyzed causes of death in prostate can- cancer patients who smoke at diagnosis. cer patients and demonstrated that more than 90% died of causes other than prostate cancer, but that current smoking increased the The Effect of Smoking on Cancer Treatment risks of non–prostate cancer deaths between 3- and 5.5-fold. As a result, tobacco use and cessation may be of paramount importance Toxicity to cancers with high cure rates, such as prostate cancer or breast cancer, simply because patients may be at the most risk of death Discussion of the effects of smoking on cancer treatment toxicity from noncancer-related causes such as heart disease, pulmonary is highly dependent upon disease site, treatment modality (e.g., disease, or other diseases related to smoking and tobacco use. surgery, CT, RT), and timing of toxicity. Across disease sites and treatments, current smoking has been shown to increase complica- tions from surgery,140–149 pulmonary complications,150,151 toxicity The Effect of Smoking on Cancer Recurrence from RT,117,152–156 mucositis,157 hospitalization,158 and vasomo- and Cancer-Related Mortality tor symptoms.159 One of the largest recent studies in over 20,000 gastrointestinal, pulmonary, and urologic patients demonstrates The primary objective of cancer therapy is to cure cancer and that former or current smoking increased the risk of surgical site prevent recurrence. However, smoking has been shown to in- infection, pulmonary complications, or 30-day mortality in a crease cancer recurrence and cancer-related mortality. Across site-specific manner.73 The effects of current smoking were most a broad spectrum of cancer patients, current smoking increased significant for pulmonary complications where former smoking cancer mortality as compared with former and never smokers.71 had a lesser or nonsignificant effect. In 13,469 lung cancer patients Current smoking has been shown to increase cancer mortality in treated with surgery, current smoking increased the risk of postop- patients with head and neck cancer,108,115–118 breast cancer,78,119 erative death with no increased risk in former smokers.160 Current

tahir99 - UnitedVRG Chapter 31 Tobacco Use and the Cancer Patient 325 smoking increased the risk of complications, morbidity, or reopera- current smokers had a 1.8-fold increased risk of developing a sec- tion following esophagectomy, pancreatectomy, or colorectal sur- ond primary as compared with all other survivors.175 gery.161–163 A study of 836 prostate cancer patients treated with RT There are some studies suggesting that smoking, combined demonstrated that current smoking increased abdominal cramps, with cytotoxic therapy, may have an additive or synergistic ef- rectal urgency, diarrhea, incomplete emptying, and sudden emp- fect on the risk of developing a second primary cancer. In 9,780 tying between two- and nine-fold,164 with similar effects noted in prostate cancer patients from the Cancer of the Prostate Strategic 3,489 cervical cancer patients who smoked more than 1 pack per Urologic Research Endeavor (CaPSURE) study, RT increased day (PPD).156 the risk of bladder cancer by 1.6-fold, smoking increased the risk Several studies have demonstrated that the effects of smoking by 2.1-fold, and smoking combined with RT increased risk by 3.7- on cancer treatment toxicity are reversible. Stopping smoking fold.176 In ER-positive breast cancer patients, treatment with RT within 3 weeks of surgery reduced wound healing complications had no significant effect on the risk of developing a contralateral in esophageal cancer patients treated with surgery and reconstruc- breast cancer, but RT combined with current smoking increased tion.165 In 393 T1 laryngeal cancer patients treated with RT, quit- the risk of contralateral cancer by ninefold.173 In a detailed analy- ting smoking after diagnosis reduced laryngeal complications as sis of Hodgkin lymphoma patients, nonheavy smokers (defined as compared with continued smoking.152 In a large study of 7,990 never, former, and less than one PPD) had a second primary rela- lung cancer patients from the Society of Thoracic Surgeons Data- tive risk of between fourfold and sevenfold when treated with CT base, current smoking increased the risk of pulmonary complica- or RT as compared with patients who received no RT or CT.177 tions by 80% and hospital mortality 3.5-fold.151 However, smoking However, heavy smokers had a sixfold increased risk in the absence cessation for 2 weeks eliminated the risks for pulmonary compli- of RT and CT and a 17- to 49-fold increased risk when combined cations, and cessation for 1 month eliminated risks for hospital with RT and/or CT. These observations suggest that smoking com- mortality. Vaporciyan et al.166 also showed that current smoking in- bined with cytotoxic cancer therapy may complement the risk of creased the risk of pulmonary complications 2.7-fold as compared developing a second primary cancer perhaps through the promo- with smoking cessation for at least 1 month prior to surgery. In tion of mutations induced by CT and/or RT in the presence of a striking example of the potentially reversible effects of smoking tobacco smoke. The potential mechanisms of this effect have not in 205 head and neck cancer patients treated with RT,167 43% of been tested or defined at this time, but the mechanism of tobacco- smoking patients treated in the morning experienced Grade 3+ induced carcinogenesis in prior reports3 supports these observations. mucositis compared with 72% of smokers treated in the afternoon (p = 0.04). These data suggest that reducing smoking overnight may yield a clinical benefit in reduced toxicity. Whereas all tox- Human Papilloma Virus, Epidermal Growth icity may not be acutely reversed, these encouraging data show Factor Receptor, Anaplastic Lymphoma that patients can make clinically meaningful improvements in Kinase, Programmed Cell Death Protein 1, their health and or cancer treatment within a short time frame by quitting smoking. and Smoking

Data over the past decade has shown that head and neck cancers that AND SCREENING CANCER PREVENTION The Effect of Smoking on Risk of Second are human papilloma virus (HPV) positive are known to have an im- Primary Cancer proved prognosis as compared with HPV-negative tumors.178 Patients who have HPV-positive tumors typically have increased p16 expres- Several studies have reported the effects of smoking on the risk of sion and often respond better to conventional cancer therapy, includ- developing a second primary cancer. Park et al.168 reported on over ing RT and CT. Many HPV-positive patients are never smokers or 14,000 male cancer patients and demonstrated that current smok- have a lighter smoking history. However, smoking was an independ- ing increased the risk of developing a second tobacco-related pri- ent adverse risk factor for both overall and cancer-related mortality mary cancer twofold, with no increased risk in former smokers. A with a 1% increase in risk per pack-year smoked.178 Current smok- higher risk was observed in in head and neck cancer patients who ing increased cancer mortality approximately fivefold even in p16- smoked more than 10 cigarettes per day, with no increased risk positive patients treated with surgery.115 Smoking also increased the in lighter smokers.169 Kinoshita et al.170 showed an 82% increased risk of developing second primary cancer in both HPV-positive and risk of developing a second primary in gastric cancer patients who HPV-negative patients.179 As a consequence, the presence of HPV are current smokers with no increased risk in former smokers. In does not appear to negate the adverse effects of smoking. the phase III randomized trial of isotretinoin for the prevention of A similar effect is noted in lung cancer patients with epidermal a second primary tumor in 1,190 head and neck cancer patients, growth factor receptor (EGFR)-mutated or anaplastic lymphoma current smoking increased the risk of a second primary by 2.2-fold kinase (ALK)-mutated tumors. As with HPV-positive head and with a nonsignificant trend of 1.6-fold in former smokers.110 Nota- neck cancer patients, lung cancer patients who are light or never bly, 39% of patients who reported quitting within the previous year smokers have a higher rate of EGFR-positive tumors that may re- were biochemically confirmed smokers.171 As a result, these data spond to biologic therapy using EGFR tyrosine–kinase inhibitors. collectively suggest that some of the increased risk may be biased At this time, most information regarding EGFR-based therapy for by continued smoking in patients who deny smoking by self-report. lung cancer reports on the effects of ever smoking demonstrating The effects of smoking on the risk of a second primary cancer that ever smokers have a decreased response to EGFR therapy. are also noted in nontobacco-related cancers and in long-term sur- Early, large, randomized trials demonstrate that Tarceva (erlotinib) vivors. In 835 breast cancer patients, smoking increased the risk for and Iressa (gefitinib) provide survival and tumor control benefits the development of lung metastases after breast cancer by more specifically in never smokers.180,181 A very similar pattern is noted than threefold.172 Ford et al.173 demonstrated that breast cancer for ALK-positive patients with a much higher incidence in never patients who were former smokers had a threefold increased risk smokers and high response rate to the ALK kinase inhibitor crizo- of developing lung cancer, but that current smokers had a 13-fold tinib.182 Paik et al.183 have described the importance of driver muta- increased risk. In nearly 1,100 estrogen receptor (ER)-positive tions in EGFR, ALK, and KRAS demonstrating that smokers have breast cancer patients, current smokers had a 1.8-fold increased a higher preponderance for K-ras drivers, whereas nonsmokers risk of developing a second contralateral breast cancer, and cur- tend to have EGFR or ALK driver mutations. In general, patients rent smokers at most recent follow-up had a 2.2-fold increased risk, who are smokers may be best served with conventional cancer but former smoking at the diagnosis or most recent follow-up had treatments rather than these biologic therapies, but randomized no increased risk.174 In 2,700 5-year survivors of testicular cancer, controlled trials confirming this suggestion are lacking at this time. 326 Cancer Prevention and Screening

Although there are essentially no biologic therapies that have Smoking Cessation Guidelines shown to have a better response in smokers, there are exciting data presented at the 2013 European CanCer Organization (ECCO) Overall, the approach to tobacco cessation for the cancer patient annual conference, suggesting that anti–programmed cell death is very similar to the approach for the general population. How- protein 1 (PD-1)–based therapies may have a better response rate ever, there are a few specific details that are important to consider in smokers.184 These very preliminary data have yet to be replicated when approaching the cancer patient who smokes. It is important or expanded into randomized trials, but if expanded trials prove ef- to recognize that virtually all newly diagnosed cancer patients are fective, they may represent one of the only cancer treatments that faced with a life-changing diagnosis that will require intensive may specifically benefit smokers. treatment approaches. Treatments, toxicity, and outcomes differ according to disease site and treatment modality. Whereas some Summarizing the Clinical Effects of Smoking cancer patients may have a curable cancer, others may have incur- able cancer. Smoking in cancer patients is also often associated on the Cancer Patient with comorbid psychiatric diseases, such as depression, that may affect dependence.193 The urgency of cessation is also important Smoking by cancer patients increases mortality, toxicity, recur- to consider. If smoking decreases the efficacy of cancer treatment, rence, and the risk of a second primary cancer. There are four then every effort should be made to stop tobacco use as soon as pos- important conclusions, and a fifth implied conclusion, to the evi- sible rather than choosing a quit date several weeks or months after dence previously presented: a cancer diagnosis. Patients may also be burdened with a “stigma” 193–197 1. One or more adverse effects of smoking affect all cancer disease associated with certain tobacco-related cancers, where they sites. may be viewed by others, or themselves, as causing their cancer 2. One or more adverse effects of smoking affect all treatment due to tobacco use. As a result, the rationale and motivation for modalities. quitting tobacco use likely differs among cancer patients, but there 3. The effects of current smoking are distinct from an ever or is a consistent theme that exists. (1) All patients should be asked f ormer smoking history. about tobacco use with structured assessments; (2) all patients who 4. Several lines of evidence demonstrate that many of the effects use tobacco or are at risk for relapse should be offered evidence- of smoking are reversible. based cessation support; and 3) tobacco assessment and cessation support should occur at the time of diagnosis, during treatment, Although substantial data demonstrate that smoking by cancer and during follow-up for all cancer patients. patients increases the risk for one or more outcomes, the largest Empiric treatment of tobacco use by cancer patients is funda- limitations are the lack of standard tobacco use definitions, the mentally supported by Public Health Service (PHS) Guidelines lack of assessing tobacco use in cancer patients at follow-up, and that are based on evidence from tobacco cessation efforts in non- the lack of structured tobacco cessation for cancer patients. Impor- cancer patients. Originally issued in 1996 and renewed in 2008, tantly, patients may further misrepresent tobacco use. Several stud- The Clinical Practice Guideline: Treating Tobacco Use and Depen- ies suggest that approximately 30% of cancer patients who smoke 171,185,186 187 dence is a PHS-sponsored, evidence-based guideline designed to deny tobacco use. Marin et al. exemplify the impor- assist health-care providers in delivering and supporting effective tance of an accurate assessment, demonstrating that patients who smoking cessation treatment.198,199 The basic recommendation self-reported smoking had no significant risk associated with surgi- states that clinicians should consistently identify, document, and cal complications; however, biochemical confirmation of smoking treat every tobacco user seen in a health-care setting. Details of significantly increased the risk of surgical wound complications. cessation support range from brief to intensive intervention, but This highlights the potential discrepancy between the effects of emphasize that consistent repeated cessation support and even smoking based on subjective versus biochemically confirmed as- brief counseling are effective methods to assist patients with stop- sessments. Due to this discrepancy, the fifth implied conclusion is ping tobacco use. It is important to note that physician-delivered that the adverse effects of smoking and the benefits of cessation interventions significantly increase long-term abstinence rates.199 may be more pronounced than currently reported in the literature. Included are newer effective medication options and strong sup- port for counseling and the use of quit lines as effective interven- tion strategies. As described in the PHS Guidelines, the principal ADDRESSING TOBACCO USE BY THE steps in conducting effective smoking cessation interventions are CANCER PATIENT referred to as The 5 A’s: 1. Ask about tobacco use for every patient. National Oncology Association Statements 2. Advise every tobacco user to quit. and Clinical Practice Guidelines 3. Assess the willingness of patients to quit. 4. Assist patients with quitting through counseling and pharmaco- Professional societies are taking leadership roles in recognizing the therapy. need to assess patients’ tobacco use and to examine the effects of 5. Arrange follow-up cessation support, preferably within the first tobacco use in medical treatment, including the important role week after the quit date. of tobacco cessation. The American Medical Association (AMA) There is a strong evidence base for these interventions as docu- passed a resolution supporting documentation of smoking be- mented in the clinical practice guideline.199 havior in clinical trials, from trial registration through treatment, follow-up, and to end of the study or death.188 The Oncology Nurs- ing Society (ONS) has also advocated for assessment and cessa- Implementing Smoking Cessation Into tion.189,190 Both the AACR5,191 and ASCO6,192 have issued policy Clinical Practice statements specifically addressing tobacco use in cancer patients, detailing that clinicians have a responsibility to address tobacco An algorithm is provided to guide clinicians in implementing the use, that all patients should be screened, that all patients who use five A’s into clinical cancer care (Fig. 31.1).5,45,194,199 Included in the tobacco should receive evidence-based tobacco cessation support, algorithm are suggested questions that are useful to accurately as- and that tobacco use should be included in clinical practice and sess tobacco use by cancer patients where patients can generally be research. These provide strong counsel to address tobacco use in divided into current, former, or never smokers. The first step (ASK) the general population as well as in cancer patients. is to inquire about and document tobacco use behaviors for every

tahir99 - UnitedVRG Chapter 31 Tobacco Use and the Cancer Patient 327

ASK: For CURRENT smokers: 4. On average, how many cigarettes per day Baseline tobacco assessment questions to did you smoke in the past 7 days? determine smoking/tobacco use status: 1. Have you smoked at least 100 cigarettes in 5. How soon after waking do you smoke your you life? first cigarette? A. Yes A. <30 min (higher nicotine dependence) B. No B. >30 min (lower nicotine dependence) 2. Do you now smoke every day, some days, or not at all? For CURRENT and FORMER smokers: A. Every day 6. At what age did you start smoking regularly? B. Some days 7. At what age did you stop smoking regularly? C. Not at all 8. When you smoke regularly, how many 3. Do you use other forms of tobacco every day, cigarettes per day did you smoke on average? some days, or not at all? A. Every day 9. How long has it been since you smoked even B. Some days a single puff? C. Not at all A. <1 day B. 1–7 days C. 8–30 days CURRENT = Answers 2A, 2B, 3A or 3B E. 4–6 months FORMER = Answers 1A and 2C and 3C F. 6–12 months NEVER = Answers 1B and 2C and 3C G. More than 1 year

All CURRENT smokers and patients Maintain All FORMER who smoked within the past 30 days abstinence, smokers prevent relapse

ADVISE patients to stop smoking with clear strong personalized advice CANCER PREVENTION AND SCREENING CANCER PREVENTION

All CURRENT smokers and patients who smoked within the past 30 days ASSESS willingness to quit immediately or set quit date as soon as possible

ASSIST patients with behavioral REASSESS current tobacco counseling and pharmacotherapy use at follow-up

ARRANGE follow-up, ideally within 1–2 Standard cancer care weeks or in preparation for cancer treatment and follow-up

Figure 31.1 This 5 A’s screening and smoking cessation treatment schema for cancer patients may be integrated into clinical oncology practice.

patient at every visit including follow-up visits. Whereas a more Less than 1% of referrals were delayed when assessments were re- comprehensive evaluation is necessary at the first consult, only peated on a monthly basis rather than at every clinic visit.45 These updates to current tobacco use are needed at follow-up. Including findings reduce the clinical burden and patient fatigue associated smoking status assessments as a “vital sign” for all patients signifi- with repeated assessments as frequently as every day such as in pa- cantly increases the identification and treatment for patients.200 To- tients who are treated with daily RT or CT. bacco-use status stickers on paper charts or an automated reminder At the time of this chapter release, there were no national system for electronic records can increase compliance with tobacco guidelines for implementation of specific questions to assess to- assessments.45 With the recent Meaningful Use standards that were bacco use in cancer patients. However, Figure 31.1 provides ef- implemented in 2011, hospitals using an electronic medical record fective questions for assessing tobacco use in cancer patients based (EMR) are essentially required to document tobacco use.201 A re- on advice from published reports.5,45,194,199 Current, former, and cent report utilized the EMR to implement mandatory tobacco as- never smokers are identified in a structured manner. Patients who sessments in cancer patients demonstrating that just a few questions use tobacco within the past 30 days should have structured sup- at the initial evaluation and at follow-up could yield high referral. port to quit tobacco use, maintain abstinence, and prevent relapse. 328 Cancer Prevention and Screening

Although not explicitly stated by any specific guidelines, asking TABLE 31.1 about tobacco use in family members of cancer patients may be important because family members often support cancer patients Additional Tobacco Cessation Resources for Patients during and following treatment, but continued smoking by family and Cliniciansa members can make quitting much more difficult.202–204 Advising is the second step in promoting effective tobacco ces- American Association for Cancer Research sation that involves giving clear, strong, and personalized advice Information about the adverse effects of tobacco and advocacy to stop tobacco use. This advice should include the importance of for tobacco control quitting smoking, such as explicit information on the risks of con- (http://www.aacr.org/home/public--media/science-policy-- tinued smoking and the benefits of cessation for cancer treatment government-affairs/science-policy--government-affairs- outcomes and overall health regardless of cancer diagnosis. This committee/tobacco-and-cancer.aspx) includes a discussion of how it is not “too late” to quit and that quit- American Cancer Society ting will in fact benefit their cancer treatment efficacy and cancer A national cancer organization providing brochures and fact 5 outcome. Patients can also consider the cost savings of stopping sheets on the health effects of tobacco and resources for a smoking habit. Clinicians must be particularly sensitive to avoid smoking cessation 195–197,205 contributing to any perceived blame for the patient’s illness. (http://www.cancer.org/cancer/cancercauses/tobaccocancer/) Clinicians must remember that most patients started smoking in adolescence and did not completely understand the risks associated American Legacy Foundation with tobacco use. At the same time, the severe addiction associated A national independent public health foundation offering with chronic tobacco use makes it difficult to stop. programs to help people quit and resources about the health The next step is assessing dependence and willingness to quit. effects of tobacco use Asking “How soon after waking do you smoke your first cigarette?” (http://www.legacyforhealth.org/) assesses nicotine dependence, with high dependence associated American Society of Clinical Oncology with a shorter interval between waking and the first cigarette.206 Resources for tobacco cessation with an emphasis on cancer Nicotine dependence is predictive of smoking cessation outcomes patients and can be used as a good indicator of the intensity of cessation (http://www.asco.org/practice-research/tobacco-cessation-and- treatment needed, such as the need for pharmacotherapy.207,208 control-resources) Determining the patient’s motivation and interest in quitting are Centers for Disease Control and Prevention critical parameters that influence the types of intervention strate- A collection of online resources, information, and materials about gies to be employed. Different strategies for quitting are based on quitting tobacco use the transtheoretical model of change and motivational interview- (http://www.cdc.gov/TOBACCO/) ing stance, which recognizes that unique intervention messages and strategies are needed to optimally promote smoking cessation North American Quitline Consortium based on a patient’s readiness to quit smoking.209,210 In the gen- Information on local and national cessation quitlines, eral population, recommendations encourage that clinicians set 1-800-QUIT-NOW a target quit date within 30 days. However, for cancer patients, (http://www.naquitline.org/) the reader is encouraged to consider an urgent need to stop smok- Smoking Cessation Leadership Center ing immediately. If patients are unable to quit immediately, then A collaborative dedicated to disseminating knowledge about the patients should be encouraged to immediately reduce tobacco use health effects of tobacco and assistance in cessation and to set a quit date as soon as possible based on the typical need (http://smokingcessationleadership.ucsf.edu/) to start cancer treatment in the immediate future. Smokefree.gov Assisting patients with smoking cessation involves clinicians Tips and resources for people trying to stop smoking helping the patient design and implement a specific quit plan or (http://smokefree.gov/) broadly enhancing the motivation to quit tobacco. Promoting an effective quit strategy for cancer patients should consist of (1) set- Tobacco Free Nurses ting a quit date (immediately or as soon as possible), (2) removing An organization aimed at engaging nurses in tobacco cessation all tobacco-related products from the environment (e.g., cigarettes, efforts ashtrays, lighters), (3) requesting support from family and friends, (http://www.tobaccofreenurses.org/) (4) discussing challenges to quitting, and (5) discussing or prescrib- U.S. Department of Health & Human Services, ing pharmacotherapy where appropriate. Patients should also be Surgeon General.gov provided information on cessation support services (Table 31.1). Tobacco use and cessation information from the Surgeon General In the cancer setting, patients can also be informed that smoking (http://www.surgeongeneral.gov/initiatives/tobacco/index.html) cessation is a critical component of cancer care over which they have complete control, thereby conferring some personal control World Health Organization over their cancer care. An international organization tasked with implementing and Patients who are unwilling to quit should continue to receive monitoring public health with a focus on tobacco control repeated assessments and counseling to help motivate patients to (http://www.who.int/topics/tobacco/en/) quit smoking. These patients should be encouraged to make im- a Current as of December 2013. mediate reductions in tobacco use and work toward abstinence as soon as possible. Clinician education, reassurance, and gentle encouragement can help them to consider changing their smok- reflective listening, summarizing).198,210,212,213 Table 31.2 provides ing behaviors. Specific strategies include discussing the personal suggested methods to help clinicians promote tobacco cessation. relevance of smoking and benefits to cessation, providing support The final step in a clinician-delivered smoking cessation inter- and acknowledging the difficulty of quitting, educating patients vention involves arranging a follow-up contact with the patient. about the positive consequences of quitting smoking, and discuss- Ideally, cancer patients will follow an immediate quit strategy and ing available pharmacologic methods to assist with quitting.211 follow-up should occur preferably within 1 to 2 weeks. However, a The emphasis should be placed on patient autonomy to quit. short-term follow-up may also benefit patients who are reluctant to Motivational strategies for patients unwilling to quit can be em- quit smoking. The clinician must remember that a new cancer di- ployed (e.g., asking open-ended questions, providing affirmations, agnosis is stressful and patients may rely on continued smoking to

tahir99 - UnitedVRG Chapter 31 Tobacco Use and the Cancer Patient 329

TABLE 31.2 form of patches, lozenges, inhalers, sprays, and gum, varenicline (Chantix), and bupropion (Zyban) are the three principal first- Select Treatment Strategies Used for Tobacco line pharmacotherapies recommended for use either alone or in Cessation Treatments combination according to PHS Guidelines.199 Table 31.3 presents information on these first-line agents. Nicotine is the primary ad- Provide and monitor the use of nicotine replacement or other dictive substance in tobacco and NRT facilitates smoking cessation pharmacotherapy. by reducing craving and withdrawal that smokers experience dur- Provide education regarding the health effects of tobacco use ing abstinence. NRT also weans smokers off nicotine by providing and its addictive and relapsing nature. a lower level and, in some cases, slower infusion of nicotine than smoking.214 Strong evidence from over 100 randomized clinical Identify and change environmental and psychological cues for trials support the use of NRT to increase the odds of quitting ap- tobacco use. proximately twofold as compared with placebo.215 Pooled analyses Generate alternative behaviors for tobacco use. demonstrate that 17% of smokers receiving NRT were able to quit Assist in optimization of social support for cessation efforts and versus 10% with placebo after at least 6 months. Recent evidence address tobacco use in family members. further shows that combination therapy, or dual NRT (such as a nicotine patch and lozenge), is a very effective smoking cessa- Prevent relapse including the identification of future high-risk tion therapy that produces high quit rates.216,217 Data suggest that situations and plans for specific behaviors in those situations. activation of the nicotinic acetylcholine receptor (nAChR) may Provide motivational interventions as needed throughout promote tumor development,218 but evidence suggests that the treatment. negative aspects of smoking outweigh these concerns.219,220 Fur- thermore, there are no clinical trials reporting negative outcomes Identify relaxation techniques such as guided imagery and for NRT in cancer patients as related to mortality or recurrence. progressive muscle relaxation. Studies also demonstrate that NRT is not associated with an in- Provide behavioral strategies to address depressed mood (e.g., creased risk of carcinogenesis in the general population.221,222 As a increasing pleasurable activities). result, NRT should be used as a clinically proven method to help Provide crisis intervention including appropriate referrals and cancer patients stop smoking. emergency intervention if indicated. Antidepressants have been studied as non-nicotine–based pharmacotherapy in part due to depression and psychiatric disease Recognize and congratulate patients on success with reducing being comorbid conditions in smokers.223 Bupropion (Zyban) is and/or quitting smoking. currently the only FDA-approved antidepressant for the treat- ment of tobacco dependence.199 Bupropion inhibits the reup- take of both dopamine and norepinephrine, thereby increasing dopamine and norepinephrine concentrations in the mesolimbic 12,224 relieve stress, but after absorbing the psychological effects of a new systems. Bupropion also antagonizes the nAChR, thereby 225 AND SCREENING CANCER PREVENTION cancer diagnosis, patients may be more receptive to smoking cessa- lowering the rewarding effects of nicotine. Should an abstinent tion. During follow-up, clinicians should congratulate patients on smoker relapse, bupropion may function to reduce the pleasure 226 successful cessation efforts, discuss accomplishments and setbacks, of cigarette smoking experienced by the smoker and help to and assess pharmacotherapy use and problems. Patients should not prevent further relapse. A meta-analysis found that smokers who be criticized for returning to smoking; rather, it is critical to create received bupropion were twice as likely as those who received a supportive environment for patients to communicate progress, placebo to have achieved long-term abstinence at either a 6- or 227 failure, and personal needs. Framing relapses as a learning experi- 12-month follow-up. ence can be helpful, and patients should be encouraged to set an- Varenicline (Chantix) is a α4β2 nAChR partial agonist that other quit date. Referrals to a psychologist or professionally trained produces sustained dopamine release in the mesolimbic system smoking cessation counselor should be considered for patients that received FDA approval for treating tobacco dependence in with numerous unsuccessful quit attempts, comorbid depression, 2006. Sustained dopamine release maintains a normal systemic anxiety, additional substance abuse disorders, or inadequate social level of the neurotransmitter, which helps to reduce craving and 228 support. withdrawal during abstinence. Varenicline also antagonizes the Clinicians who are not well versed in tobacco cessation should rewarding effects of nicotine. Because varenicline attenuates the realize that smoking is an extremely difficult addiction to overcome pleasure smokers experience from smoking, it may decrease moti- and should recognize the clinical pattern associated with cessa- vation to smoke and protect them from relapse. One of the initially tion. As patients stop smoking, many will experience symptoms of reported randomized clinical trials that compared varenicline withdrawal, including dry or sore throat, constipation, cravings to (2 mg), bupropion (300 mg), and placebo showed that varenicline smoke, irritability, anxiety, trouble concentrating, restlessness, in- was superior to bupropion and placebo, with overall continuous 229 creased appetite, depression, and insomnia. In the first few weeks, abstinence rates between 10% to 23%. A meta-analysis demon- patients may also report an increase in mucous secretions from the strated that the 1-mg daily dose approximately doubled, whereas airways, a cough, and other upper respiratory tract symptoms. Pa- the 2-mg daily dose approximately tripled the likelihood of long- 199 tients and clinicians should realize that tobacco cessation requires term abstinence at 6 months as compared to placebo. As a result, a concerted effort, may require repeated attempts, and symptoms the 1-mg daily dose can be considered as an alternative should will not resolve immediately. Clinicians should counsel patients the patient experience significant dose-related side effects. Several on a repeated basis, recognize success, and provide repeated as- meta-analyses have shown that varenicline is superior to bupro- 230–233 sistance if patients relapse. pion and placebo in the general population. In July 2009, the FDA issued a warning after reports that some patients attempting to quit smoking while using varenicline or Pharmacologic Treatment for Smoking bupropion experienced unusual changes in behavior, depressed Cessation mood, worsening of depression, or had thoughts of suicide. This has prompted recommendations that health-care providers elicit The principles of pharmacotherapy to help patients quit smoking information about a patient’s psychiatric history prior to pre- are fundamentally based on reducing the craving associated with scribing varenicline or bupropion to closely monitor changes in nicotine withdrawal. Nicotine replacement therapy (NRT), in the mood and behavior during the course of treatment. However, 330 Cancer Prevention and Screening

TABLE 31.3 First-Line Pharmacotherapy Agents for the Treatment of Nicotine Depedence

Agent Dose Mechanism Use Nicotine Replacement Transdermal (patches) 16 h or 24 h Steady state NRT to reduce 6–10 CPD: 14 mg daily × 8 wks then 7 mg 7, 14, or 21 mg craving and withdrawal daily × 2 wks 1 patch/d >10 CPD: 21 mg daily × 6 wks, then 14 mg × 2 wks, then 7 mg × 2 wks Gum 2 or 4 mg Short-term NRT to reduce craving First cigarette >30 min after waking: 2 mg Max: 24 pieces/d and withdrawal PO q1–2 hr First cigarette <30 min after waking: 4 mg PO q1–2 hr Lozenge 2 or 4 mg Short-term NRT to reduce craving 1st cigarette >30 min after waking: 2 mg PO Max: 20 lozenges/d and withdrawal q1–2 hr 1st cigarette <30 min after waking: 4 mg PO q1–2 hr Nasal spray 0.5 mg/spray Short-term NRT to reduce craving 1 spray/nostril q1–5 hr Max:10 sprays/hr or and withdrawal 80 sprays/d Inhaler 4 mg/cartridge Short-term NRT to reduce craving 1 cartridge inhaled over 20 min q1.5–6 hr Max: 16 cartridges/d and withdrawal Bupropion (Zyban) 150 mg Block nicotinic receptors and 1 tablet daily × 3 d, then 1 tablet twice daily reduces reward for 7–12 wks Varenicline (Chantix) 0.5 or 1 mg Dopaminergic reward and partial 0.5 mg daily × 3 d, then 0.5 mg twice daily nicotinic receptor antagonist × 3 d, then 1 mg twice daily

CPD, cigarettes per day; PO, by mouth; NRT, nicotine replacement therapy.

updated recent safety studies examining very large databases intervention comparison in 186 newly diagnosed head and neck (one database of N = 119,546, one database of N = 35,800) re- cancer patients. Patients were treated with either minimal ad- garding safety have shown no difference in neuropsychiatric side vice or an enhanced intervention with trained clinicians consist- effects between varenicline or bupropion as compared to NRT ing of strong personalized advice to stop smoking, a contracted and no increased risk of depression.234,235 Another prospective quit date, tailored written materials, and booster advice sessions. study showed no adverse events when treating participants with No significant differences were found between treatments, but a current or past major depression and also showed higher absti- 70.2% continuous abstinence rate was found at 12-month follow- nence rates for the varenicline group as compared to placebo at up regardless of treatment condition, suggesting that many cancer weeks 9 to 52 (20.3% versus 10.4%, p <0.001).236 Varenicline patients can benefit from brief physician-delivered advice. A later should be considered a viable cessation pharmacotherapy for study by Schnoll et al.,239 comparing cognitive behavioral treat- cancer patients. ment with standardized health education advice, also failed to find The clinical practice guideline also identifies two non-nicotine– significant differences in quit rates. All patients received NRT, and based medications—clonidine and nortriptyline—as second-line quit rates in both groups approached 50% at 1-month follow-up pharmacotherapies for tobacco dependence. A second-line agent and 40% at 3-month follow-up. is used when a smoker cannot use first-line medications due to Additional studies, ranging from 15 to 80 patients, examined either contraindications or lack of effectiveness. Both clonidine, an nurse-delivered cessation interventions for a variety of cancer pa- antihypertensive, and nortriptyline, a tricyclic antidepressant, have tients. The lowest cessation rates were found with a single session been shown to effectively assist smokers achieve abstinence.227,237 intervention: a 21% cessation rate in the intervention group versus Unfortunately, many patients who quit will eventually relapse, 14% in the usual care group 6 weeks’ postintervention.240 Higher and rates of long-term abstinence remain low. Because smoking cessation rates were associated with a more intensive intervention poses enormous health risks to individuals and their families, even consisting of three inpatient visits, supplementary materials, and a modest reduction in smoking may translate into a significant im- five postdischarge follow-up contacts. Additional studies dem- pact on public health. Clinicians should continue to encourage onstrate higher cessation rates with more intensive intervention recalcitrant smokers to stop tobacco use and use pharmacotherapy (40% to 75%) as compared with usual care (43% to 50%), sug- where appropriate with repeated quit attempts. gesting more intensive interventions may yield higher cessation rates.241–243 In general, more intense interventions appear to be more efficacious, but even brief advice is important to achieve to- Empirically Tested Cessation Interventions bacco cessation. with Cancer Patients In a randomized trial of 432 cancer patients coordinated by the Eastern Cooperative Oncology Group (ECOG) with a The overwhelming majority of cessation research has been per- physician-delivered intervention (comprised of cessation advice, formed in the general population, but there are several studies that optional NRT, and written materials) or usual care (unstruc- have been performed in cancer patients. Gritz et al.238 conducted tured advice from physicians), there were no significant inter- the first physician- or dentist-delivered randomized cessation vention effects and generally low abstinence rates (12% to 15%

tahir99 - UnitedVRG Chapter 31 Tobacco Use and the Cancer Patient 331 at 6 to 12 months).244 However, patients with head and neck or Current Tobacco Assessment and Cessation lung cancer were significantly more likely to have quit smok- Support by Oncologists ing compared to patients with tumors that were not smoking related. Analyses of outcomes from the Mayo Clinic Nicotine Access to cessation support is critical to address tobacco use by Dependence Center found that although lung cancer patients cancer patients. A recent survey of 58 NCI-designated cancer were more likely to achieve 6-month tobacco abstinence than centers indicated that about 80% reported a tobacco use program controls (22% versus 14%), no significant differences were ob- 245 246 available to their patients and about 60% routinely offered edu- served after adjusting for covariates. Garces et al. also found cational materials, but less than 50% had a designated individual no significant differences in abstinence rates between head and who provided services.254 A recent survey of over 1,500 members neck cancer patients and controls (33% versus 26%). However, of the International Association for the Study of Lung Cancer higher abstinence rates were found for both head and neck and (IASLC)255 and a parallel study of 1,197 ASCO members256 ob- lung cancer patients treated within 3 months of diagnosis com- served that approximately 90% of physicians believe that tobacco pared to those treated for more than 3 months after the diagnosis, affects outcomes, tobacco cessation should be a standard part of emphasizing the potential importance of the teachable moment cancer care, and approximately 80% regularly advise patients to at the time of the cancer diagnosis. stop using tobacco, but only approximately 40% discuss medica- The potential importance of addressing smoking combined tions or assist with quitting. Dominant perceived barriers to cessa- with considering comorbid disease has been noted in a few studies. tion support were patient resistance to treatment, an inability to get In a randomized head and neck cancer patients of usual care ver- patients to quit, a lack of cessation resources, and a lack of clinician sus 9 to 11 sessions of a nurse-administered intervention consisting education. These data showed that even motivated clinicians are of cognitive-behavioral therapy and medications, targeting comor- not regularly providing tobacco cessation support. A recent survey bid smoking, drinking, and depression significantly increased quit of 155 actively accruing cooperative group clinical trials further rates at 6-month follow-up for the intervention group compared to < 247 demonstrated that only 29% of active trials collected any tobacco the usual control group (47% versus 31%, p 0.05). In a ran- use information, 4.5% collected any tobacco use information at domized trial of 246 cancer patients treated with 9 weeks of NRT follow-up, and none addressed tobacco cessation.55 Few oncology with or without bupropion, there was no significant difference with meetings offer educational workshops or talks, and they are often the addition of bupropion to NRT, but in patients with depressive poorly attended when they are offered.257 Collectively, these data symptoms, bupropion increased abstinence rates, lowered with- 248 demonstrate that oncologists are not regularly providing cessation drawal, and improved quality of life. Patients without depression support and that we are not capturing tobacco use information that symptoms did equally well when treated with bupropion versus may be critical to understanding the effects of tobacco on cancer transdermal nicotine and counseling alone. treatment outcomes. Patient recruitment has been a problem noted by some studies, More in-person talks as well as written and Web-based training including 5.5 years to accrue 246 patients with telephone screen- 249 should be made available, as well as new approaches that move ing of over 7,500 potential patients. A pilot trial of varenicline from the traditional 5 A’s model delivered by a single professional in thoracic oncology patients required screening 1,130 patients to

to referral systems that efficiently connect tobacco users to mul- AND SCREENING CANCER PREVENTION accrue 49 participants randomized to a 12-week course of either tiple resources for tobacco cessation.258–260 The ASCO Prevention varenicline or placebo paired with a behavioral counseling plat- 250 Curriculum has a chapter devoted to educating oncology health- form of seven sessions. A randomized trial of 185 smoking can- care professionals on the evaluation and treatment of tobacco cer patients comparing the efficacy of a hospital-based standard use.213 Innovative curricula, such as the Texas Tobacco Outreach care smoking cessation model versus standard care augmented by Education Program (TOEP), are available and can facilitate pro- a behavioral tapering regimen via a handheld device before inpa- gram development in other states.261 However, specialty programs tient hospitalization for cancer surgery demonstrated no difference 251 in tobacco cessation treatment that are based in cancer centers in quit rates (both 32%,). However, over 29,000 patients were and other medical centers are valuable resources that need to be screened to conduct a randomized clinical trial with a smoking further developed. cancer patient population. These studies highlight the potential Addressing tobacco use in cancer patients may be approached difficulty recruiting participants who smoke, including consid- in a systematic and efficient manner. A recent report highlighted erations for the importance of medical comorbidity in guiding the potential utility of automated tobacco assessment and smoking smoking cessation treatment, patient mix (multiple tumor sites), cessation using structured assessments in the EMR where all pa- treatment status (awaiting treatment to completed treatment), tients were automatically referred to a dedicated cessation program variation in stage of disease, and considering how psychiatric consisting of phone-based cessation support.45 In 2,700 patients re- conditions such as depression reflect the difficulty of conducting ferred for cessation support, half received only a mailing and only research in the oncology setting and the importance of these vari- 1% contacted the cessation program. However, in the arm with at ables in future studies. least five phone call attempts made by the cessation service, 81% of Although accruing patients to intervention trials may seem patients were successfully contacted and only 3% refused cessation discouraging, several studies demonstrate the benefit of counsel- 252 support. Furthermore, assessments implemented every 4 weeks, ing over self-help. Emmons et al. conducted a randomized rather than more frequent assessments every 2 weeks, resulted in controlled trial in 796 young adult survivors of pediatric cancer delayed cessation referrals in less than 1% of smokers. This is the that included six calls, tailored and targeted written materials, and first report to try and identify clinically efficient mechanisms of optional NRT as compared with self-help. Significantly higher quit addressing tobacco use that may be useful in clinical practice or rates were found in the counseling group compared to the self-help research that may be an effective method of increasing patient par- group at all reported follow-up time points, including 12 months < ticipation in cessation support, but substantial work is needed to (15% versus 9%; p 0.01). A randomized trial of a motivational assess who may benefit from low versus high intensity support in interviewing-based smoking cessation intervention in a south Aus- such a program. tralian hospital was delivered over a 3-month period, consisted of multiple contacts with a trained counselor, and provided supple- mentary material tailored to cancer patients with NRT.253 The Examples of Model Tobacco Treatment control group received brief advice to quit and generic supplemen- Programs tary material. Quit rates did not differ by treatment group (5% to 6% at 3-month follow-up), but the intervention group was signifi- Several dedicated tobacco treatment programs at cancer centers cantly more likely to report attempts to quit smoking. have been developed. Table 31.4 contrasts the core elements of 332 Cancer Prevention and Screening

TABLE 31.4 Attributes of Prototypical Tobacco Treatment Programs

Attribute MDACC RPCI Yale MSKCC Tobacco assessment of all patients (EMR) Yes Yes Yes Yes Automatic referral of all patients Yes Yes No Yes In-person counseling Yes No Yes Yes Telephone counseling Yes Yes No Yes Medications prescribed Yes No Yes Yes Biochemical confirmation (CO) testing Yes No Yes No Free to patient Yes Yes No Yes Third-party payment No No Yes Yes Research studies of new treatments Yes Yes Yes Yes

MDACC, M.D. Anderson Cancer Center, Houston, TX; RPCI, Roswell Park Cancer Institute, Buffalo, NY; Yale, Yale University Hospital, Smilow Cancer Center, New Haven, CT; MSKCC, Memorial Sloan Kettering Cancer Center, New York, NY.

four active model programs at the end of 2013 (University of Texas compliance, second primary, and noncancer-related comor- M.D. Anderson Cancer Center, Roswell Park Cancer Institute, bidity is needed. All cancer disease sites and stages are im- Yale Cancer Center, and Memorial Sloan Kettering Cancer Cen- portant to consider. ter), each of which employ different methods to help cancer pa- 2. Understanding the effects of tobacco and cessation on cancer tients quit smoking. All programs follow the evidence-based 5 A’s biology. Although not a primary focus of this chapter, tobacco model described previously from PHS Guidelines.199 All programs and tobacco-related products increase tumor growth, angiogen- were made available to patients at their respective medical cen- esis, migration, invasion and metastasis and decrease response ters and are now designed to evaluate and treat all patients who to conventional cancer treatments such as CT and RT. These self-report current tobacco use. Importantly, not all cancer centers and other areas are important to consider, including the po- can treat smoking cessation in the same manner. Financing of a tential effects on immune-related therapy and vaccine devel- cessation program is critical and may include institutional funds, opment. In vivo models of exposure and cancer response are state funds, research funds, and third-party billing. Notably, given not well developed, yet are critical to this research area. Work the broad spectrum of adverse health effects associated with smok- is also needed to assess the effect of emerging tobacco-related ing, cancer centers should carefully consider the potential health products such as e-cigarettes. benefits and cost savings associated with tobacco cessation due to 3. Advance understanding of models to increase access to cessa- reductions in treatment complications and recurrence associated tion support and increase efficacy of tobacco cessation meth- with smoking by cancer patients. There is no one “correct” way to ods for cancer patients. This diverse area includes assessing create and sustain a tobacco treatment program at a cancer center, the timing of intervention, intensity, duration, follow-up, and but at the very least and consistent with evidence, rigorous behav- the potential effects of harm-reduction strategies. Cessation ioral counseling should be provided and, if possible, medication pharmacology requires additional consideration in combina- management as well. tion with unique approaches to motivational and behavioral counseling in cancer patients. Significant work is needed to FUTURE CONSIDERATIONS disseminate evidence-based cessation support and to assess the cost-effectiveness of different cessation strategies, particularly with regard to improving the cost of cancer care as a whole. Research Considerations Preventing relapse and evaluating the safety of transition to al- ternative products such as e-cigarettes is equally important and The past several years have shown a surge in activities identifying increasingly complex with the addition of new tobacco-related the effects of tobacco in cancer patients and increasing awareness products. Identifying and addressing barriers to effective ces- is being developed for cessation support at cancer centers as well as sation support is also needed. As related to the cancer patient, through several national organizations. There are three fundamen- clinicians and cessation specialists should consider how their tal areas of research that need to be expanded: research relates to cancer care. Taking advantage of new in- tegrated medical management systems presents a significant 1. Evaluating the effects of tobacco use and cessation on clini- opportunity to improve cessation support access as well as to cal cancer outcomes. The 2014 SGR concluded that smok- develop a more effective tracking of patient outcomes. ing caused adverse outcomes in cancer patients,7 but several limitations remain. Tobacco-use definitions should be stan- dardized and implemented at diagnosis, during treatment, Policy Implications and Systematic Issues and follow-up. Biochemical confirmation with cotinine or exhaled carbon monoxide may improve the accuracy of to- Several national and international organizations have emphasized bacco assessment in at-risk groups such as current smokers the importance of tobacco assessments and cessation for the general who are trying to quit or patients who reported quitting in population and for cancer patients that include tools to evaluate the past year.171,185,186,262 Although smoking is the predomi- tobacco use at diagnosis, during treatment, and follow-up appoint- nant form of tobacco consumption, all tobacco products ments, as well as routine support for smoking cessation.5,6,188–191 should be considered. A further understanding of the ef- In 2012, ASCO, with the contribution of the American Legacy fects of tobacco on the efficacy and toxicity of cancer treat- Foundation, published a Tobacco Cessation Toolkit for the on- ment, tumor response, quality of life, survival, recurrence, cology setting.263 This evidence-based guideline intends to help

tahir99 - UnitedVRG Chapter 31 Tobacco Use and the Cancer Patient 333 oncology providers integrate tobacco cessation strategies into their tremendous public health burden from tobacco-related disability patient care. Utilization of the EMR and standardized, automated and death has not been countered by a proportional level of fund- systems for more efficacious and efficient access to tobacco cessa- ing in tobacco control, cancer treatment research, or public advo- tion support has also been suggested,45 but requires participation cacy. Researchers, clinicians, and advocates must come together to by clinicians, institutions, insurers, and health departments. Not persuade policy makers to increase funding in tobacco-related re- only should providers be aware of the need for tobacco cessation search, treatment, and policy initiatives on behalf of healthy indi- and available interventions, but health-care institutions must also viduals and patients. A united front is critically needed in support build such treatment into their overall system of care. Thus, the of a common agenda that includes both increased tobacco-control identification of patients who smoke or use any alternative tobacco efforts and additional funding for disease-related research and treat- product, referral or direct treatment by providers, billing and reim- ment. With clinical rationale, guidelines, and advocacy in place, bursement for treatment provided, and consistent efforts from pro- the final steps in effective tobacco control and improving health fessional oncology organizations are critically important.257 The outcomes are to implement these recommendations into practice.

SELECTED REFERENCES

The full reference list can be accessed at lwwhealthlibrary.com/oncology. 110. Khuri FR, Lee JJ, Lippman SM, et al. Randomized phase III trial of low-dose isotretinoin for prevention of second primary tumors in stage I and II head 1. U.S. Department of Health, Education, and Welfare. Smoking and Health: and neck cancer patients. J Natl Cancer Inst 2006;98:441–450. Report of the Advisory Committee to the Surgeon General of the Public Health 114. Bittner N, Merrick GS, Galbreath RW, et al. Primary causes of death after per- Service. PHS Publication No. 1103. Washington, D.C.: U.S. Department of manent prostate brachytherapy. Int J Radiat Oncol Biol Phys 2008;72:433–440. Health, Education, and Welfare, Public Health Service, Center for Disease 166. Vaporciyan AA, Merriman KW, Ece F, et al. Incidence of major pulmonary Control; 1964. morbidity after pneumonectomy: association with timing of smoking cessa- 2. Office of the Surgeon General, Office on Smoking and Health. The Health tion. Ann Thorac Surg 2002;73:420–425. Consequences of Smoking. A Report of the Surgeon General. Atlanta: Centers 167. Bjarnason GA, Mackenzie RG, Nabid A, et al. Comparison of toxicity as- for Disease Control and Prevention; 2004. sociated with early morning versus late afternoon radiotherapy in patients 3. Centers for Disease Control and Prevention, National Center for Chronic with head-and-neck cancer: a prospective randomized trial of the National Disease Prevention and Health Promotion, Office on Smoking and Health. Cancer Institute of Canada Clinical Trials Group (HN3). Int J Radiat Oncol How Tobacco Smoke Causes Disease: The Biology and Behavioral Basis for Biol Phys 2009;73:166–172. Smoking-Attributable Disease: A Report of the Surgeon General. Atlanta: 177. Travis LB, Gospodarowicz M, Curtis RE, et al. Lung cancer following Centers for Disease Control and Prevention; 2010. chemotherapy and radiotherapy for Hodgkin’s disease. J Natl Cancer Inst 5. Toll BA, Brandon TH, Gritz ER, et al. Assessing tobacco use by cancer 2002;94:182–192. patients and facilitating cessation: an American Association for Cancer Re- 185. Morales N, Romano M, Cummings KM, et al. Accuracy of self-reported search policy statement. Clin Cancer Res 2013;19:1941–1948. tobacco use in newly diagnosed cancer patients. Cancer Causes Control 6. Hanna N, Mulshine J, Wollins DS, et al. Tobacco cessation and control a 2013;24:1223–1230. decade later: American society of clinical oncology policy statement update. 186. Warren GW, Arnold SM, Valentino JP, et al. Accuracy of self-reported to- J Clin Oncol 2013;31:3147–3157. bacco assessments in a head and neck cancer treatment population. Radio-

7. U.S. Department of Health and Human Services. The Health Consequences ther Oncol 2012;103:45–48. AND SCREENING CANCER PREVENTION of Smoking—50 Years of Progress: A Report of the Surgeon General. Atlanta: 187. Marin VP, Pytynia KB, Langstein HN, et al. Serum cotinine concentration U.S. Department of Health and Human Services, Centers for Disease Con- and wound complications in head and neck reconstruction. Plast Reconstr trol and Prevention, National Center for Chronic Disease Prevention and Surg 2008;121:451–457. Health Promotion, Office on Smoking and Health; 2014. 192. American Society of Clinical Oncology. American Society of Clinical On- 9. Benowitz NL. Nicotine addiction. N Engl J Med. 2010;362:2295–2303. cology policy statement update: tobacco control—Reducing cancer inci- 24. Warren GW, Cummings KM. Tobacco and lung cancer: risks, trends, and dence and saving lives. 2003. J Clin Oncol 2003;21:2777–2786. outcomes in patients with cancer. In: American Society of Clinical Oncol- 193. Gritz ER, Fingeret MC, Vidrine DJ, et al. Successes and failures of the teach- ogy 2013 Educational Book. 201;359–364. ASCO University Web site. http:// able moment: smoking cessation in cancer patients. Cancer 2006;106:17–27. meetinglibrary.asco.org//content/200-132. Accessed November 25, 2013. 194. Gritz ER, Dresler C, Sarna L. Smoking, the missing drug interaction in clini- 25. Thun MJ, Carter BD, Feskanich D, et al. 50-year trends in smoking-related cal trials: ignoring the obvious. Cancer Epidemiol Biomarkers Prev 2005;14: mortality in the United States. N Engl J Med. 2013;368:351–364. 2287–2293. 27. Cummings KM, Dresler CM, Field JK, et al. E-cigarettes and cancer pa- 199. Fiore MC, Jaén CR, Baker TB, et al. Treating Tobacco Use and Dependence: 2008 tients. J Thorac Oncol 2014;9:438–441. Update. Rockville, MD: U.S. Department of Health and Human Services; 2008. 36. Gritz ER. Smoking and smoking cessation in cancer patients. Br J Addict http://www.ncbi.nlm.nih.gov/books/NBK63952/. Accessed November 25, 2013. 1991;86:549–554. 211. Toll BA, Rojewski AM, Duncan L, et al. “Quitting smoking will benefit your 45. Warren GW, Marshall JR, Cummings KM, et al. Automated tobacco assess- health”: the evolution of clinician messaging to encourage tobacco cessation. ment and cessation support for cancer patients. Cancer 2014;120:562–569. Clin Cancer Res 2014;20:301–309. 48. Gritz ER, Nisenbaum R, Elashoff RE, et al. Smoking behavior following 213. Gritz ER, Fingeret MC, Vidrine DJ. Tobacco control in the oncology set- diagnosis in patients with stage I non-small cell lung cancer. Cancer Causes ting. In: Brawley OW, Khuri FR, Rock CL, eds. ASCO Cancer Prevention Control 1991;2:105–112. Curriculum. Alexandria, VA: American Society of Clinical Oncology; 2007. 55. Peters EN, Torres E, Toll BA, et al. Tobacco assessment in actively accru- 218. Warren GW, Singh AK. Nicotine and lung cancer. J Carcinog 2013;12:1–8. ing National Cancer Institute Cooperative Group Program Clinical Trials. J 221. Jorgensen ED, Zhao H, Traganos F, et al. DNA damage response induced by Clin Oncol 2012;30:2869–2875. exposure of human lung adenocarcinoma cells to smoke from tobacco- and 56. Land SR. Methodologic barriers to addressing critical questions about to- nicotine-free cigarettes. Cell Cycle 2010;9:2170–2176. bacco and cancer prognosis. J Clin Oncol 2012;30:2030–2032. 222. Murray RP, Connett JE, Zapawa LM. Does nicotine replacement therapy cause 57. Gritz ER, Dresler C, Sarna L. Smoking, the missing drug interaction in cancer? Evidence from the Lung Health Study. Nicotine Tob Res 2009;11: clinical trials: ignoring the obvious. Cancer Epidemiol Biomarkers Prev 1076–1082. 2005;14:2287–2293. 227. Hughes JR, Stead LF, Lancaster T. Antidepressants for smoking cessation. 71. Warren GW, Kasza K, Reid M, et al. Smoking at diagnosis and survival in Cochrane Database Syst Rev 2003;(2):CD000031. cancer patients. Int J Cancer 2013;132:401–410. 229. Jorenby DE, Hays JT, Rigotti NA, et al. Efficacy of varenicline, an alpha4be- 73. Gajdos C, Hawn MT, Campagna EJ, et al. Adverse effects of smoking on post- ta2 nicotinic acetylcholine receptor partial agonist, vs placebo or sustained- operative outcomes in cancer patients. Ann Surg Oncol 2012;19:1430–1438. release bupropion for smoking cessation: a randomized controlled trial. 80. Holmes MD, Murin S, Chen WY, et al. Smoking and survival after breast JAMA 2006;296:56–63. cancer diagnosis. Int J Cancer 2007;120:2672–2677. 235. Thomas KH, Martin RM, Davies NM, et al. Smoking cessation treatment 84. Kenfield SA, Stampfer MJ, Chan JM, et al. Smoking and prostate cancer and risk of depression, suicide, and self harm in the Clinical Practice Re- survival and recurrence. JAMA 2011;305:2548–2555. search Datalink: prospective cohort study. BMJ 2013;347:f5704. 107. Browman GP, Wong G, Hodson I, et al. Influence of cigarette smoking on 236. Anthenelli RM, Morris C, Ramey TS, et al. Effects of varenicline on smok- the efficacy of radiation therapy in head and neck cancer. N Engl J Med ing cessation in adults with stably treated current or past major depression: 1993;328:159–163. a randomized trial. Ann Intern Med 2013;159:390–400. 108. Gillison ML, Zhang Q, Jordan R, et al. Tobacco smoking and increased risk 238. Gritz ER, Carr CR, Rapkin D, et al. Predictors of long-term smoking cessa- of death and progression for patients with p16-positive and p16-negative oro- tion in head and neck cancer patients. Cancer Epidemiol Biomarkers Prev pharyngeal cancer. J Clin Oncol 2012;30:2102–2111. 1993;2:261–270. 334 Cancer Prevention and Screening

244. Schnoll RA, Zhang B, Rue M, et al. Brief physician-initiated quit-smoking 256. Warren GW, Marshall JR, Cummings KM, et al. Addressing tobacco use in strategies for clinical oncology settings: a trial coordinated by the Eastern cancer patients: a survey of American Society of Clinical Oncology (ASCO) Cooperative Oncology Group. J Clin Oncol 2003;21:355–365. members. J Oncol Pract 2013;9:258–262. 248. Schnoll RA, Martinez E, Tatum KL, et al. A bupropion smoking cessation 257. Gritz ER, Sarna L, Dresler C, et al. Building a united front: aligning the clinical trial for cancer patients. Cancer Causes Control 2010;21:811–820. agendas for tobacco control, lung cancer research, and policy. Cancer Epide- 251. Ostroff JS, Burkhalter JE, Cinciripini PM, et al. Randomized trial of a presur- miol Biomarkers Prev 2007;16:859–863. gical scheduled reduced smoking intervention for patients newly diagnosed 262. Society for Research on Nicotine and Tobacco Committee on Biochemical with cancer. Health Psychol 2013 [Epub ahead of print]. Verification. Biochemical verification of tobacco use and cessation. Nicotine 254. Goldstein AO, Ripley-Moffitt CE, Pathman DE, et al. Tobacco use treatment Tob Res 2002;4:149–159. at the U.S. National Cancer Institute’s designated Cancer Centers. Nicotine 263. American Society of Clinical Oncology. Tobacco Cessation Guide: for On- Tob Res 2013;15:52–58. cology Providers. ASCO Web site. http://www.asco.org/sites/default/files/to- 255. Warren GW, Marshall JR, Cummings KM, et al. Practice patterns and per- bacco_cessation_guide.pdf. Accessed November 25, 2013. ceptions of thoracic oncology providers on tobacco use and cessation in can- cer patients. J Thorac Oncol 2013;8:543–548.

tahir99 - UnitedVRG Role of Surgery in Cancer 32 Prevention

José G. Guillem, Andrew Berchuck, Jeffrey F. Moley, Jeffrey A. Norton, Sheryl G. A. Gabram-Mendola, and Vanessa W. Hui

INTRODUCTION Following referral for genetic assessment, three groups of pa- tients emerge.8 The first consists of those women who have under- Since the heritable component of some cancer predispositions has gone genetic testing and have been found to harbor a mutated gene been linked to mutations in specific genes, clinical interventions associated with high penetrance for breast cancer. Given that the have been formulated for mutation carriers within affected fami- possibility of developing breast cancer in this group may be as high lies. The primary interventions for mutation carriers for highly pen- as 90%, there is a role for enhanced surveillance or risk-reduction etrant syndromes, such as multiple endocrine neoplasia (MEN), surgery. The American Cancer Society has published guidelines familial adenomatous polyposis (FAP), hereditary nonpolyposis for magnetic resonance imaging (MRI) screening as a method colorectal cancer (CRC), and hereditary breast and ovarian cancer for enhanced surveillance.9 Women in this first group qualify for syndromes, are primarily surgical. This chapter is divided into five such screening, which can be offered annually but scheduled at sections addressing breast (S.G.A.G.), gastric (J.N.), ovarian and 6-month intervals with screening mammography to increase the endometrial (A.B.), and MENs (J.F.M.) and colorectal (J.G.G., rate of identifying interval cancers. Alternatively, simultaneous V.W.H.). For each, the clinical and genetic indications and timing screening with MRI and mammography to compare one modal- of prophylactic surgery and its efficacy, when known, are provided. ity with the other on an annual basis may also be offered. Another Prophylactic surgery in hereditary cancer is a complex process, choice for this group of women is to pursue bilateral risk-reduction requiring a clear understanding of the natural history of the disease mastectomy with an option for immediate reconstruction. Bilat- and variance of penetrance, a realistic appreciation of the potential eral salpingo-oophorectomy for BRCA1 and BRCA2 mutation car- benefit and consequence of a risk-reducing procedure in an other- riers may also be considered, as this procedure has been shown wise potentially healthy individual, and the long-term sequelae of to reduce breast cancer risk by almost 50%.8,10 This is especially such surgical intervention, as well as the individual patient’s and true for BRCA2 mutation carriers, who tend to develop hormone receptor–positive breast cancers. family’s perception of surgical risk and anticipated benefit. AND SCREENING CANCER PREVENTION The second group consists of women with strong family his- tories suggestive of hereditary breast cancer who test negative for PATIENTS AT HIGH RISK FOR BREAST both the BRCA1 and BRCA2 mutations as well as the other de- CANCER scribed syndromes. In this group, there may not have been a family member with cancer who was tested for the mutation. Therefore, a negative test does not necessarily indicate that a woman’s risk is Identification of Patients at Risk equivalent to that of the general population.7 There may also be an undetected mutation in such a family, indicating the possibility of A detailed family history is the most important tool for identifying higher-than-average risk for that particular woman. These women individuals at increased risk for hereditary cancers. The US Pre- may or may not qualify for enhanced surveillance with MRI ventive Services Task Force updated their recommendation for risk screening,9 and accurate assessment of their risk may require the assessment, genetic counseling, and genetic testing for asympto- use of other risk prediction tools,3 in addition to evaluating for the matic women who have not been diagnosed with a BRCA-related presence of lobular carcinoma in situ, atypical lobular hyperplasia, cancer. In this update, the use of a risk screening tool is highly or atypical ductal hyperplasia, and determining if a more intensive recommended to identify appropriate patients for referral for ge- surveillance regimen is necessary based on heterogeneously or ex- netic counseling.1 The American Society of Clinical Oncology tremely dense breast tissue on mammography. has also updated the policy on genetic and genomic testing for The third group consists of women with a strong family his- cancer susceptibility, and this update includes information on tory of breast cancer, who for various reasons, have chosen not to genetic tests of uncertain clinical utility and direct-to-consumer pursue genetic testing. These individuals may have other health- marketing, both of which impact the practice of oncology and related problems, psychological concerns, cost issues, or they may preventive medicine.2 Historically, genetic counseling and testing fear perceived medical insurance discrimination. Women in all were offered by health-care providers. However, with the advent groups can be educated that with passage of the Genetic Infor- of direct-to-consumer marketing, individuals may obtain tests and mation Nondiscrimination Act in 2008, significant advances have receive results directly from a company. The American Society of occurred that protect patients from discrimination by employers Clinical Oncology still endorses pre- and posttest counseling for and health insurers.11 thorough disclosure of the impact of testing. Before any woman Women in the second and third groups may still qualify for considers risk-reduction surgery such as bilateral mastectomy or bilateral risk-reduction mastectomy and immediate reconstruc- salpingo-oophorectomy, referral to a high-risk or genetic screen- tion. Often, women who elect this path are influenced by their ing program is desirable, as women often overestimate their actual family history or by witnessing breast and/or ovarian cancer deaths breast cancer risk.3 in close family members, giving them a significant fear of a breast The most common cancer syndromes that place women at risk or ovarian cancer diagnosis. For women in all three groups, the for breast cancer are BRCA14 and BRCA25 gene mutations. Other decision of whether to pursue risk-reducing surgery is difficult. less common syndromes are listed in Table 32.1.6,7 Often, the expertise of a cancer clinical psychologist or psychiatrist

335 336 Cancer Prevention and Screening

TABLE 32.1 Hereditary Carcinoma Syndromes Including Breast Cancer

Breast Cancer Syndrome Chromosome/Gene Primary Carcinoma Secondary Carcinoma Penetrance Familial breast cancer/ 17g21; BRCA1 Breast cancer, Colon, prostate 60%–80% ovarian cancer syndrome Autosomal dominant ovarian cancer Familial breast cancer/ 13q12; BRCA2 Breast cancer, Male breast cancer, 60%–80% ovarian cancer syndrome Autosomal dominant ovarian cancer endometrial, prostate, oropharyngeal, pancreatic Li-Fraumeni syndrome 17p13.1 and 22q12.1; TP53 Soft tissue cancers Soft tissue sarcoma, 50%–85% (for all and CHEK2 (including breast) leukemia, osteosarcoma, types of cancers Autosomal dominant melanoma, colon, pancreas, in this syndrome) adrenal syndrome, cortex, and brain tumors PTEN hamartoma syndrome 10q23.31; PTEN mutation Breast cancer Thyroid (follicular) and 25%–50% (Cowden’s) Autosomal dominant endometrial carcinoma Peutz-Jeghers syndrome 19p13.3; STK11 Gastrointestinal Esophagus, stomach, small 29% Autosomal dominant cancers intestine, large bowel, pancreas, lung, ovary, endometrial Diffuse gastric cancer 16q22.1; CDH1 Diffuse gastric cancer Colorectal, lobular breast 39% (lobular Autosomal dominant cancer breast cancer) Louis-Bar syndrome 11q22.3; ATM Leukemia and Ovarian, breast, gastric, 38% (for all types Autosomal recessive lymphoma melanoma, leiomyomas, of cancers in the sarcomas syndrome)

PTEN, phosphatase and tensin homolog. Data from Lux MP, Fasching PA, Beckmann MW. Hereditary breast and ovarian cancer: review and future perspectives. J Mol Med 2006;84:16–28; and Shannon KM, Chittenden A. Genetic testing by cancer site: breast. Cancer J 2012;18:310–319.

is enlisted, as risk-reduction mastectomy involves an irreversible time and thus may have contributed to the number of increased procedure with body image and sexual implications.8 recurrences. Updated in 2007, the Society of Surgical Oncology published Another surgical option for high-risk women is bilateral salpingo- a position statement on the role of prophylactic mastectomy for oophorectomy. Among a cohort of women with BRCA1 and BRCA2 patients at high risk for breast cancer, as well as those patients mutations, this procedure has been associated with a lower risk of recently diagnosed with breast cancer who are considering con- mortality from both breast and ovarian cancer.10 As an additional tralateral prophylactic breast surgery.12 For women at high risk, benefit, this procedure also decreases the risk of breast cancer in indications fall into three broad categories: presence of a mutation this patient population, likely through the mechanism of decreasing in BRCA or other susceptible genes, strong family history with no hormonal exposure at a younger age. demonstrable mutation, and histologic risk factors (biopsy-proven Contemporary surgical procedures for risk-reducing bilateral atypical ductal hyperplasia, atypical lobular hyperplasia, or lobular mastectomy include total mastectomy, skin-sparing mastectomy carcinoma in situ especially in patients with a strong family his- (preservation of the skin envelope by removal of the entire breast tory of breast cancer). Recommendations for patients with recently through a circumareolar incision around the nipple–areola com- diagnosed breast cancer are similar in that they include the indica- plex), subcutaneous mastectomy, areola-sparing mastectomy (re- tions for high-risk individuals previously noted, as well as future moval of the nipple while sparing the areola), and nipple-sparing surveillance challenges for the opposite breast (clinically and mastectomy (removal of entire breast and nipple core tissue but mammographically dense breast tissue or diffuse, indeterminate preservation of nipple–areolar skin).14 Given advances in recon- microcalcifications in the contralateral breast). Another impor- structive nipple–areolar techniques, it appears that total mastec- tant consideration is the need for symmetry in patients with large, tomy with or without skin-sparing methods reduces the risk of ptotic, or disproportionately sized contralateral breasts. breast cancer to the greatest extent with reasonable cosmesis. More limited and long-term follow-up data are available on areola- and nipple-sparing techniques. The potential limitations of these pro- Surgical Issues and Technique cedures are distortion of the nipple–areola complex and lack of sensitivity after breast tissue has been completely removed.8 In a single institution’s 33-year experience,13 the risk for breast can- Immediate reconstruction is offered to patients and performed cer in both moderate- and high-risk groups of women based on in the vast majority undergoing bilateral risk-reduction mastec- family history was reduced by at least 89% for women who under- tomy. Choices of reconstruction include a bilateral pedicled or went bilateral prophylactic mastectomy. From a technical perspec- free tissue transverse rectus abdominis muscle flap, a free bilateral tive, in this study, women either had a subcutaneous mastectomy deep inferior epigastric perforator flap or superficial inferior epi- (removal of the majority of breast tissue with sparing of the nipple– gastric artery flap, bilateral latissimus flaps with or without implant areola complex) or total mastectomy (removal of the entire breast or expanders, or bilateral implant or expander placement alone.14 through the nipple–areola complex). Most of the recurrences Although tissue flap transfer gives a more natural appearance and occurred in women undergoing a subcutaneous mastectomy. texture to the reconstructed site, individual body contour drives However, this was the most frequent procedure performed at that the ultimate plan for reconstruction. The decision about the type

tahir99 - UnitedVRG Chapter 32 Role of Surgery in Cancer Prevention 337 of reconstruction should be made by the plastic surgeon with input a decrease in intestinal-type gastric cancers, the overall incidence from the surgical oncologist, especially for the group of women of gastric cancer has declined significantly in the past 50 years. with breast cancer desiring bilateral mastectomies who may re- However, the incidence of diffuse gastric cancer (DGC), which is quire adjuvant radiation for treatment. also called signet ring cell or linitis plastica, has remained stable Although the risk reduction is dramatic for bilateral mastec- and, by some reports, may be increasing. tomy, residual breast tissue may be left behind, especially with skin- Hereditary DGC (HDGC) is a genetic cancer susceptibility sparing procedures. Patients should be educated that careful chest syndrome defined by one of the following: (1) two or more docu- wall surveillance is recommended after such a procedure. Local mented cases of DGC in first- or second-degree relatives, with at recurrences after bilateral implant reconstruction are reliably de- least one diagnosed before the age of 50; or (2) three or more cases tected by clinical examination. Recurrences after reconstruction of documented DGC in first- or second-degree relatives, indepen- with autologous tissue present most commonly on the skin 50% dent of age of onset. The average age of onset of HDGC is 38, and to 72% of the time and are detectable by physician examination.15 the pattern of inheritance is autosomal dominant.20 Figure 32.1 Nonpalpable deeper recurrences in this setting are less common, shows a pedigree with HDGC. and use of mammography image surveillance may be indicated, In 1998, inactivating germline mutations in the E-cadherin especially if significant breast tissue was left behind unintention- gene CDH1 were identified in three Maori families, each with ally during the bilateral mastectomy procedure. At times, an ini- multiple cases of poorly differentiated DGC.21 The CDH1 muta- tial “screening” mammogram may be performed, if significant tions in these families were inherited in an autosomal dominant residual breast tissue is suspected; this should occur well after all pattern, with incomplete but high penetrance. Onset of clinically healing has taken place to delineate the amount of visible breast apparent cancer was early, with the youngest affected individual tissue on imaging. This drives future decisions of whether to follow dying of DGC at the age of 14.21 Since then, germline muta- a patient with imaging. Finally, all patients should be instructed tions of CDH1 have been identified in 30% to 50% of all patients to return for clinical breast examination with the health provider with HDGC.19,22 More than 50 mutations have been recognized if any change is noted on the reconstructed breasts, regardless of across diverse ethnic backgrounds, including European, African imaging plan. American, Pakistani, Japanese, Korean, and others.19 In addition Although risk-reduction bilateral mastectomy may be exceed- to gastric cancers, germline CDH1 mutations are associated with ingly beneficial for high-risk women, especially for those testing increased risk of lobular carcinoma of the breast, and this was the positive for BRCA1, BRCA2, or other deleterious mutations, or first manifestation of aCDH1 mutation in one series.23 CDH1 is, belonging to a family afflicted with a cancer syndrome, they are to date, the only gene implicated in HDGC. Penetrance of DGC never emergent procedures. Along with risk-reduction bilateral in patients carrying a CDH1 mutation is estimated at 70% to 80%, salpingo-oophorectomy, risk-reduction bilateral mastectomy re- but may be higher. The need for a systematic study of specimens sides at the far end of the spectrum of an individual’s choices.16 is supported by recent work by Gaya et al.24 in which initial total These procedures should be offered only after appropriate genetic gastrectomy specimens were reported as negative, but detailed sec- counseling and accurate assessment of a woman’s actual risk for tioning and analysis showed invasive carcinoma. breast and ovarian cancer. An in-depth consultation with the pa- CDH1 is localized on chromosome 16q22.1 and encodes the tient and her family members is necessary prior to proceeding with calcium-dependent cell adhesion glycoprotein E-cadherin. Func- AND SCREENING CANCER PREVENTION an operative plan. tionally, E-cadherin impacts maintenance of normal tissue mor- phology and cellular differentiation. It is hypothesized that CDH1 acts as a tumor suppressor gene in HDGC, with loss of function HEREDITARY DIFFUSE GASTRIC CANCER leading to loss of cell adhesion and subsequently to proliferation, invasion, and metastases. Figure 32.2 shows the CDH1 mutation Gastric cancer is the fourth most common cause of cancer for the pedigree depicted in Figure 32.1. worldwide and is the second leading cause of cancer mortality.17 The germline CDH1 mutation is most frequently a truncat- Although environmental agents, including Helicobacter pylori and ing mutation. Germline missense mutations are causative in a few diet, are the primary risk factors for this disease, approximately 10% HDGC kindreds, but are more often clinically insignificant. In of gastric cancers are a result of familial clustering.18,19 Histologi- vitro assays for cellular invasion and aggregation may predict the cally, gastric cancers may be classified as either intestinal or diffuse functional impact of missense mutations to aid in this distinction.22 types. The intestinal type histopathology is linked to environmen- Within the gastric mucosa, the “second hit” leading to complete tal factors and advanced age. The diffuse type occurs in younger loss of E-cadherin function results from CDH1 promoter methyla- patients and is associated with a familial predisposition. Because of tion, as has been described in sporadic gastric cancer.25

63 80 GC GC BC

55 40 45 56 45 40 25 69 GC GC GC GC GC GC Figure 32.1 A family pedigree showing autosomal dominant inheritance of gastric cancer. Individual mutation testing results for the codon 1003 CDH1 mutation are indicated 51 52 55 4951 56 50 56 57 42 42 5351 50 by + or −. Individuals affected with gastric cancer are shaded. (From Norton JA, Ham CM, GC Van Dam J, et al. CDH1 truncating mutations in the E-cadherin gene: an indication for total 22 23 25 26 23 gastrectomy to treat hereditary diffuse gastric cancer. Ann Surg 2007;245:873.) 338 Cancer Prevention and Screening

aa 1 335 882

N SPRE EXTRACELLULAR TM CP C Exon # 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Nucleotide 1003 Arginine (CGA) Stop (TGA)

Figure 32.2 The mutation in this kindred is located in the central region of the E-cadherin gene that codes for the extracellular cadherin domains of the protein containing calcium-binding motifs important in the adhesion process. The C → T transition in exon 7 of nucleotide 1003 results in a premature stop codon (R335X), producing truncated peptides that lack the transmembrane and cytoplasmic β-catenin–binding domains essential for tight cell-cell adhesion. Black area indicates truncated portion of peptide. N, N-terminus; C, C-terminus; S, signal peptide; PRE, precursor sequence; TM, transmembrane domain; CP, cytoplasmic domain. (From Norton JA, Ham CM, Van Dam J, et al. CDH1 truncating mutations in the E-cadherin gene: an indication for total gastrectomy to treat hereditary diffuse gastric cancer. Ann Surg 2007;245:873.)

It remains unclear whether specific CDH1 mutations are as- clinically apparent DGC is only 10%, with the majority dying sociated with distinctive phenotypic characteristics or rates of before age 40. penetrance, although this may become apparent as more recur- Because of high cancer penetrance, poor outcome, and inad- rent mutations are recognized. To date, most mutations identified equacy of clinical screening tools for HDGC, prophylactic total have been novel and distributed throughout CDH1. Recognition gastrectomy is recommended as a management option for asymp- of recurrent mutations has usually resulted from independent tomatic carriers of CDH1 mutations.18 Although total gastrectomy events; however, there is evidence for the role of founder effects is performed with prophylactic intent in these cases, most speci- in certain kindreds.22 At present, it is also unclear whether patients mens have been found to contain foci of diffuse signet ring cell can- with HDGC without detectable CDH1 mutations have mutation cer.19,28,29 Foci of DGC have been identified even in patients who of a different gene or merely a CDH1 mutation that has gone have undergone extensive negative screening, including high-reso- unrecognized. lution computed tomography, positron emission tomography scan, New recommended screening criteria for CDH1 mutations are chromoendoscopy-guided biopsies, and endoscopic ultrasonogra- as follows: phy.19 However, HGDC in asymptomatic CDH1 carriers is usu- ally completely resected by prophylactic gastrectomy, as pathologic 1. Families with one or more cases of DGC analyses of resected specimens have shown only T1N0 disease. 2. Individuals with DGC before the age of 40 years without a Because these signet ring cell cancers are multifocal and dis- family history tributed throughout the entire stomach, especially in the cardia,30 3. Families or individuals with cases of DGC (one case below the prophylactic gastrectomy should include the entire stomach, and age of 50 years) and lobular breast cancer the surgeon must transect the esophagus and not the proximal 4. Cases where pathologists detect in situ signet ring cells or pa- stomach. Furthermore, it should be performed by a surgeon ex- getoid spread of signet ring cells adjacent to diffuse type gastric perienced in the technical aspects of the procedure and familiar cancer18,26 with HDGC. In asymptomatic patients, lymph node metastases As in other familial cancer syndromes, genetic counseling have not been observed; therefore, D2 lymph node resection is not should take place prior to genetic testing so that the family un- necessary. The optimal timing of prophylactic gastrectomy in indi- derstands the potential impact of the results. After obtaining viduals with CDH1 mutations is unknown, but recent consensus informed consent, a team comprising a geneticist, gastroenterolo- recommendations indicate that age 20 is reasonable.18 gist, surgeon, and oncologist should discuss the possible outcomes Although it is a potentially lifesaving procedure, prophylactic of testing and the management options associated with each. gastrectomy for CDH1 mutation carries significant risks that must Genetic testing should first be performed on a family member with be considered. Overall mortality for total gastrectomy is estimated HDGC or on a tissue sample if no affected relative is living. In to be as high as 2% to 4%, although it is estimated to be 1% when addition to direct sequencing, multiplex ligation-dependent probe performed prophylactically. Patients must also be aware that there amplification is recommended to test for large genomic rearrange- is a nearly 100% risk of long-term morbidity associated with this ments. If a CDH1 mutation is identified, asymptomatic family procedure, including diarrhea, dumping, weight loss, and difficulty members may proceed with genetic testing, preferably by the age eating.19 A recent study of the effects of prophylactic gastrectomy of 20.19 If no mutation is identified in the family member with for CDH1 mutation demonstrated that physical and mental func- DGC, the value of testing asymptomatic relatives is low. tion were normal at 12 months, but specific digestive issues were Among individuals found to carry a germline CDH1 muta- recognized. Overall, 70% had diarrhea, 63% fatigue, 81% eating dis- tion, clinical screening is problematic. Histologically, DGC is comfort, 63% reflux, 45% eating restrictions, and 44% had altered characterized by multiple infiltrates of malignant signet ring cells, body image, suggesting that this operation impacted negatively on which may underlie normal mucosa.27 Because these malignant quality of life.31 Because of these complications and the fact that foci are small in size and widely distributed, they are difficult to lymph node spread has not been observed, some recommend vagus- identify via random endoscopic biopsy. Chromoendoscopy and preserving gastrectomy done either open or laparoscopically. In ad- positron emission tomography have reportedly been used, but the dition, because the penetrance of CDH1 mutations is incomplete, clinical utility of these tools in early detection remains unproven. some patients who undergo prophylactic gastrectomy would never Lack of a sensitive screening test for HDGC makes early diag- have gone on to develop clinically significant gastric cancer. Pro- nosis extremely challenging. By the time patients are symptom- phylactic gastrectomy has, in fact, been performed on several pa- atic and present for treatment, many have diffuse involvement tients reported to show no evidence of gastric cancer on pathology.29 of the stomach or linitis plastica, and rates of mortality are high. Some individuals with CDH1 mutations choose not to pur- Published case reports describe patients who have presented with sue prophylactic gastrectomy. These individuals should undergo extensive DGC despite recent normal endoscopy and negative careful surveillance, including biannual chromoendoscopy with biopsies.28 The 5-year survival rate for individuals who develop biopsies, beginning when they are at least 10 years younger than

tahir99 - UnitedVRG Chapter 32 Role of Surgery in Cancer Prevention 339 the youngest family member with DGC was at time of diagnosis. in other genes that appear to affect the risk of ovarian cancer in It is recommended that any endoscopically visible lesion is tar- BRCA1/2 carriers.38 Based on the known ovarian cancer risk– geted and that six random biopsies are taken from the following re- modifying loci, it has been reported that the 5% of BRCA1 carriers gions: antrum, transitional zone, body, fundus, and cardia. Careful at lowest risk have a lifetime risk of ≤28% of developing ovarian white-light examination with targeted and random biopsies com- cancer, whereas the 5% at highest risk have a ≥63% lifetime risk. bined with detailed histopathology can identify early lesions and In the future, when modifier loci are more completely catalogued, help to inform decision making with regard to gastrectomy.32 Addi- more precise estimates of cancer risk may be provided to individ- tionally, because women with CDH1 mutations have a nearly 40% ual patients who are considering RRSO. lifetime risk of developing lobular breast carcinoma, they should As about 20% of women with high-grade serous ovarian cancers be carefully screened with annual mammography and breast MRI have BRCA1/2 mutations, it has been suggested that all of these starting at age 35.23 They should also do monthly self-examinations women undergo genetic testing regardless of family history.39 Mu- and have a breast examination by a physician every 6 months. The tational analysis in women with these cancers may increasingly same surveillance recommendations are probably appropriate for become standard practice as the cost of genetic testing declines. HDGC families without identifiable CDH1 mutations, although Testing may also be driven by the availability of poly(ADP-ribose) no current guidelines for this exist. polymerase inhibitor therapy for women whose cancers have germ- The emergence of gene-directed gastrectomy as a treatment line or sporadic mutations in genes such as BRCA1/2 and others strategy for patients with HDGC represents the culmination of a that are involved in homologous recombination DNA repair. successful collaboration between molecular biologists, geneticists, RRSO is strongly recommended in women who carry BRCA1/2 oncologists, gastroenterologists, and surgeons. It is anticipated that mutations because of the high mortality rate of ovarian cancer the recognition of similar molecular markers in other familial can- and the lack of effective screening and prevention approaches. cer syndromes will transform the approach to the early diagnosis Although screening with pelvic ultrasound and serum CA125 is and treatment of a variety of tumors. generally recommended for BRCA1/2 carriers during their 20s and 30s, it is not proven to reduce ovarian cancer mortality be- cause even early stage high-grade cancers have a very high mortal- SURGICAL PROPHYLAXIS OF HEREDITARY ity. Oral contraceptives reduce the risk of ovarian cancer in the OVARIAN AND ENDOMETRIAL CANCER general population and appear to have a similar effect in BRCA1/2 carriers, but this must be balanced against concerns regarding increased breast cancer risks. Hereditary Ovarian Cancer (BRCA1, BRCA2) The past practice of performing RRSO based solely on family history has been replaced by reliance on genetic testing. Clinical Inherited mutations in BRCA1 and BRCA2 strongly predispose management of women with a strong family history in whom a del- women to breast cancer and to high-grade serous cancers of the eterious germline mutation is not found, or those with variants of ovary, fallopian tube, and peritoneum.33,34 About two-thirds are uncertain significance, should be resolved on a case-by-case basis. due to BRCA1 mutations and one-third BRCA2 mutations, and RRSO may be deemed appropriate in some cases, despite the ab- these account for about 15% to 20% of high-grade serous cases. sence of a clearly deleterious mutation. Fortunately, the risk of AND SCREENING CANCER PREVENTION The lifetime risk of these gynecologic cancers increases from a hereditary ovarian cancer does not rise dramatically until the mid- baseline of 1.5% to about 15% to 25% in BRCA2 carriers and 30% 30s in women with BRCA1 mutations and the 40s for women with to 60% in BRCA1 carriers.33,34 BRCA1/2 mutations are rare in BRCA2 mutations.36 As a result, most women are able to complete most populations (<1 in 500 individuals); one notable exception childbearing prior to undergoing RRSO. It is advisable for BRCA1 is the Ashkenazi Jewish population, in which the carrier frequency carriers to undergo RRSO around age 35, as there is a 4% risk of is 1 in 40.35 BRCA1-associated cases peak in the 50s and BRCA2- ovarian cancer being discovered clinically or at the time of RRSO associated cancers in the 60s.36 In addition to BRCA1/2 mutations, by age 40.10 BRCA2 carriers may choose to delay surgery into their germline mutations in a number of other genes in the homolo- 40s due to their lower risk of ovarian cancer, but this could dimin- gous recombination DNA repair pathway confer high penetrance ish the protection against breast cancer that is afforded by RRSO. susceptibility to ovarian cancer (e.g., RAD51C, RAD51D, BRIP1, If a mutation carrier, particularly a BRCA1 carrier, chooses to pur- PALB2).37 This has led to the development of more comprehen- sue fertility into her 40s, then she should be counseled that she is sive cancer genetic testing panels that are increasingly being used at considerable risk of developing a life-threatening cancer that is to identify women who are candidates for risk-reducing salpingo- largely preventable. oophorectomy (RRSO). Several studies have provided evidence of the efficacy of RRSO. Genetic testing for inherited high-penetrance mutations in In one early study of BRCA1/2 carriers, RRSO reduced the rate of BRCA1/2 and other genes should be discussed with women who breast and ovarian cancer by 75% over several years of follow-up.40 A have a significant family history of early onset breast cancer and/or separate study in 2002 examined outcome in 551 BRCA1/2 carriers cancers of the ovary, fallopian tube, or peritoneum. Involvement from various registries.41 Among 259 women who had undergone of a genetic counselor prior to testing is helpful, as they have ex- RRSO, 6 (2.3%) were found to have stage I ovarian cancer at the pertise in managing the inherent clinical and social issues. Most time of the procedure and 2 (0.8%) subsequently developed serous BRCA1/2 mutations involve base deletions or insertions in the peritoneal carcinoma. Among the controls, 58 (20%) women devel- coding sequence or splice sites that encode truncated protein prod- oped ovarian cancer after a mean follow-up of 8.8 years. With the ex- ucts that are clearly dysfunctional. Less frequently, disease-causing clusion of the six women whose cancers were diagnosed at surgery, mutations may occur that alter a single amino acid, though most RRSO reduced ovarian cancer risk by 96%. More recently, in 2014, of these missense variants represent innocent polymorphisms. an international registry study of over 5,783 subjects with median The clinical significance of missense mutations can sometimes be follow-up of 5.6 years found that RRSO reduced ovarian, tubal, and elucidated by determining whether they segregate with cancer in peritoneal cancer risk by 80%.36 There was an estimated lifetime risk other family members. In addition, genomic rearrangements may of primary peritoneal cancer after RRSO of about 4% for BRCA1 occur that inactivate BRCA1 or BRCA2, and identification of such carriers and 2% for BRCA2 carriers.36 The risk of death from all alterations requires molecular testing beyond sequencing. causes was reduced by 77%. A prospective cohort study noted that Penetrance of ovarian cancer is not 100% in those with clearly RRSO was associated with reduction in breast cancer–specific (haz- deleterious BRCA1/2 mutations, but presently it is not possible ard ratio [HR] = 0.44; 95% confidence interval [CI]= 0.26 to 0.76), to provide more precise personalized risk estimates to guide the ovarian cancer–specific (HR= 0.21; 95% CI = 0.06 to 0.80), and use of RRSO. However, common variants have been discovered all-cause mortality (HR = 0.40; 95% CI = 0.26 to 0.61).10 340 Cancer Prevention and Screening

A B Figure 32.3 Hematoxylin and eosin (A) and immunohistochemical staining (B) demonstrating overexpression of mutant TP53 in serous carcinoma in situ of the fallopian tube from a BRCA1 mutation carrier who underwent risk-reducing bilateral salpingo-oophorectomy.

Removal of the ovaries, as internal organs, usually has little ef- appreciated, and more recent studies support the contention that fect on body image and self-esteem, and most BRCA1/2 mutation RRSO reduces the risk of breast cancer by about half in BRCA1/2 carriers elect to undergo RRSO. Insurance payers will almost al- carriers.42 However, a meta-analysis showed that while RRSO was ways pay for RRSO in proven mutation carriers. strongly protective against estrogen receptor–positive breast cancer RRSO can be performed laparoscopically in most women, with (HR = 0.22), there was no protection against estrogen receptor– discharge to home the same day. If a laparoscopic approach is prob- negative breast cancer.42 Many carriers are identified after develop- lematic due to obesity or adhesions, the surgery can be performed ing early onset breast cancer, and this group represents the most through a small lower abdominal incision. Morbidity including difficult in which to balance the potential risks and benefits of es- bleeding, infection, and damage to the urinary or gastrointestinal trogen replacement therapy. tracts can occur, but the incidence of serious complications is very Early stage high-grade serous cancers and in situ lesions with low. As the fallopian tubes and ovaries are small discrete organs, TP53 mutations have been identified in the fallopian tubes of they are relatively easy to remove completely. Attention should some RRSO specimens (Fig. 32.3). This has led to a paradigm be paid to transecting the ovarian artery and vein proximal to the shift in which it is now thought that most high-grade serous can- ovary and tube so that remnants are not left behind. This involves cers found in the ovary, fallopian tube, and peritoneum are derived opening the pelvic sidewall peritoneum, visualizing the ureter, from cells that originate in the tubal fimbria.43 The frequency of and then isolating the ovarian blood supply. If there are adhesions occult malignancies has varied between reports, but appears to be between the adnexa and adjacent structures, careful dissection about 3%.43 In view of this, the pelvis and peritoneal cavity should should be performed to ensure complete removal of the ovaries be examined carefully. Malignant cells also have been found in and fallopian tubes. If the uterus is not removed, care should be peritoneal cytologic specimens, and washings of the pelvis should taken to remove the entire fallopian tube. A small portion of the be obtained when performing RRSO. The pathologist should be tube inevitably will be left in the cornu of the uterus, but the risk informed of the indication for surgery and serial sections of the of fallopian tube cancer developing in such remnants appears to fallopian tubes should be performed to look for the presence of be negligible. early lesions. Patients found to have occult invasive high-grade se- Though there is not strong evidence that BRCA1/2 mutations rous cancers should be treated with chemotherapy after surgery. increase uterine cancer risk, many women elect to have the uterus Those with in situ lesions appear to have a good outcome without removed as part of the surgical procedure because they have com- chemotherapy.44 pleted their family or have other gynecologic indications. Although Cases of peritoneal serous carcinoma indistinguishable from the addition of a hysterectomy may increase operative time, blood ovarian cancer have been observed years after RRSO, but the ori- loss, surgical complications, and hospital stay, it usually can be gin of these cancers is unclear. Some may represent recurrences performed laparoscopically and serious adverse outcomes are in- of occult ovarian or tubal cancers. In this regard, retrospective ex- frequent. Furthermore, the likelihood of future exposure to tamox- amination of the ovaries and fallopian tubes sometimes has revealed ifen in the context of breast cancer prevention or treatment, which primary cancers that were not originally recognized. In contrast, increases endometrial cancer risk two- to three-fold, also argues some of these cancers likely arise directly from fallopian tube cells for concomitant hysterectomy. Women who receive hormone re- that have implanted in the peritoneum and subsequently become placement therapy after surgery will require a progestin along with malignant. Patients who undergo RRSO should be made aware of estrogen to protect against the development of endometrial cancer their residual risk of peritoneal cancer, but there is no evidence that if the uterus is not removed. continued surveillance using CA125 and/or ultrasound is beneficial. In younger women, surgical menopause after RRSO is associ- ated with vasomotor symptoms, vaginal atrophy, decreased libido, and an accelerated onset and incidence of osteoporosis and car- HEREDITARY ENDOMETRIAL CANCER diovascular disease. In premenopausal women who do not have (LYNCH SYNDROME) a personal history of breast cancer, estrogen replacement can be administered to ameliorate many of the deleterious effects of pre- Although Lynch syndrome (LS, also known as hereditary non- mature menopause. Systemic estrogen levels are lower in oopho- polyposis CRC syndrome) typically manifests as familial cluster- rectomized premenopausal women taking hormone replacement ing of early onset CRC, there is also an increased incidence of than if the ovaries had been left in place. The therapeutic ben- several other types of cancers—most notably endometrial cancer efit of oophorectomy in women with breast cancer has long been in women.45 About 3% of endometrial cancers are attributable to

tahir99 - UnitedVRG Chapter 32 Role of Surgery in Cancer Prevention 341 inherited mutations in the DNA mismatch repair (MMR) genes Many women with LS elect to undergo risk-reducing colec- that cause LS. Most often, MSH2 and MLH1 are implicated, but tomy, which provides an opportunity to perform concomitant mutations in MSH6 and PMS2 also occur.45 The risk of ovarian hysterectomy. Hysterectomy in concert with colectomy, either cancer is also significantly increased in LS, but to a lesser degree via laparoscopy or laparotomy, does not greatly increase operative than in BRCA1/2 mutation carriers, and accounts for only about time or surgical complications. If an endometrial biopsy has not 1% of all ovarian cancers. been performed preoperatively, an intraoperative inspection of the Cells in which one of the LS genes have been inactivated ex- uterine cavity and possibly frozen section should be performed to hibit a phenomenon called microsatellite instability (MSI).46 This exclude the presence of cancer. If cancer is found in the uterus, occurs as DNA mismatches cause shortening or lengthening of surgical staging—including sampling of the regional lymph repetitive DNA sequences and these mismatches go unrepaired. nodes—should be considered in addition to hysterectomy.53 It is This results in generation of alleles in the cancer that contain a also appropriate to discuss risk-reducing hysterectomy with LS car- greater or lesser number of repeats than are present in normal cells riers who do not elect to undergo prophylactic colectomy. The from that individual. MSI occurs in most LS-associated colorectal operative approach (vaginal versus laparotomy versus laparoscopy) and endometrial cancers.46 However, MSI is found in about 20% can be determined based on the presence or absence of uterine of sporadic cancers that arise in these organs, and in most cases pathology (e.g., myomas), whether the patient has had prior ab- is caused by silencing of the MLH1 gene due to promoter hyper- dominal surgery, and whether the ovaries are also to be removed. methylation. Screening strategies for identification of MMR gene alterations in families with LS-associated cancers include analysis of tumor tissue for MSI and/or loss of DNA MMR gene expression GYNECOLOGIC CANCER RISK IN VERY using immunohistochemistry (IHC).46 In cancers with MSI or loss RARE HEREDITARY CANCER SYNDROMES of expression of one of the MMR genes, or in families with pedi- grees suggestive of LS, these genes can be sequenced to identify Several very rare hereditary cancer syndromes also increase the disease-causing mutations, most of which cause truncated protein 47 risk of gynecologic cancers, and some of these women could po- products. Although it has been suggested that it may be cost- tentially benefit from risk-reducing surgery to remove the ovaries effective to do these tests on all endometrial cancers, this approach 47 and/or uterus. Peutz-Jeghers syndrome is characterized by intesti- has not been widely adopted. nal polyps and an increased risk of colorectal and breast cancers. The risk of a woman who carries a LS mutation developing 45,48 This rare syndrome is due to inherited mutations in the STK11 endometrial cancer ranges from 20% to 60% in various reports. gene. Affected women also have an increased risk of ovarian sex The risk of ovarian cancer is increased to about 5% to 12%. cord–stromal tumors with annular tubules and adenoma malig- Whereas the mean age of women with sporadic endometrial can- num of the cervix. Li-Fraumeni syndrome is caused by inherited cers is in the early 60s, cancers that arise in association with LS are mutations in the TP53 gene, and carriers are predisposed to a num- often diagnosed before menopause, with the average age in the ber of types of cancers including sarcomas and breast cancer. The 40s. The clinical features of these endometrial cancers are similar risk of ovarian cancer is increased as well, but is not a major cause to those of most sporadic cases (well-differentiated, endometrioid of cancer in these families. Cowden syndrome is due to germline CANCER PREVENTION AND SCREENING CANCER PREVENTION histology, early stage), and survival is about 90%. The mean age PTEN mutations and increases the risk of several malignancies in- of onset of ovarian cancer in LS is in the early 40s, and the clini- cluding breast, thyroid, mucocutaneous, and endometrial cancers. cal features of these cancers are generally more favorable than in Finally, small cell carcinoma of the ovary, hypercalcemic type, is sporadic cases. They usually are identified at an early stage, are due to mutations in the SMARCA4 gene. These highly lethal ovar- well- or moderately differentiated, have favorable survival, and ian cancers occur at a very young age (median 24 years) and pre- some occur in the setting of a synchronous endometrial cancer. sent difficult challenges related to timing of RRSO. There are no Recommendations for screening and risk-reducing surgery in well-accepted evidence-based guidelines for early detection and LS are better established for CRC than for extracolonic malignan- 49 prevention of gynecologic cancers in these very rare hereditary cies. Transvaginal ultrasound has been proposed as a screening cancer syndromes. An awareness of the risk and natural history of test for endometrial cancer (and ovarian cancer), but its efficacy 50 gynecologic cancers in these families provides a basis for counsel- is unproven. Endometrial biopsy is the most sensitive means of ing individual patients. diagnosing endometrial cancer, and it has been suggested that this should be employed periodically beginning around age 30 to 35. However, there are no published studies demonstrating that this MULTIPLE ENDOCRINE NEOPLASIA TYPE 2 approach prevents endometrial cancer deaths compared to simply performing a biopsy if abnormal uterine bleeding occurs. Most experts believe that risk-reducing hysterectomy has a role Gene Carriers in the management of some women with LS because of the high incidence of endometrial cancer. The risk of endometrial cancer The MEN type 2 syndromes include MEN 2A, MEN 2B, and is low during the prime reproductive years, and the uterus does familial (non-MEN) medullary thyroid carcinoma (FMTC).54–56 not serve a vital function once childbearing has been completed. These are autosomal dominant inherited syndromes caused by In view of the increased risk of ovarian cancer in LS, concomi- germline mutations in the RET proto-oncogene. Their hallmark tant bilateral salpingo-oophorectomy should also be considered. is the development of multifocal bilateral medullary thyroid carci- One study demonstrated that there were no cases of endometrial noma (MTC) associated with C-cell hyperplasia. MTCs arise from or ovarian cancer in 61 LS carriers who underwent risk-reducing the thyroid C-cells, also called parafollicular cells. C-cells secrete hysterectomy and bilateral salpingo-oophorectomy, while endo- the hormone calcitonin, a specific tumor marker for MTC. A slow- metrial cancer occurred in 33% and ovarian cancer in 5% who growing tumor in most cases, MTC causes significant morbidity retained their uterus and ovaries.51 Despite the low risk of death and death in patients with uncontrolled local or metastatic spread. from gynecologic cancers in LS, cost-effectiveness analyses of Large tumor burden is associated with diarrhea and flushing. In various approaches suggest that risk-reducing hysterectomy and the MEN 2 syndromes, there is almost complete penetrance of salpingo-oophorectomy leads to both the lowest cost and the great- MTC. Other features are variably expressed, with incomplete pen- est increase in quality-adjusted life-years.52 Estrogen replacement etrance (summarized in Table 32.2). after removal of the ovaries in premenopausal women with LS In MEN 2A, all patients develop MTC. Approximately 42% of is not contraindicated, as there is no evidence that this adversely affected patients also develop pheochromocytomas, associated with affects the incidence of other cancers. adrenal medullary hyperplasia. Hyperparathyroidism develops in 342 Cancer Prevention and Screening

TABLE 32.2 Clinical Features of Sporadic Medullay Thyroid Carcinoma, Multiple Endocrine Neoplasia 2A, Multiple Endocrine Neoplasia 2B, and Familial Medullay Thyroid Carcinoma

Clinical Setting Features of MTC Inheritance Pattern Associated Abnormalities Genetic Defect Sporadic MTC Unifocal None None Somatic RET mutations in >20% of tumors MEN 2A Multifocal, bilateral Autosomal dominant Pheochromocytomas, Germline missense hyperparathyroidism, cutaneous mutations in extracellular lichen amyloidosis, Hirshprung’s cysteine codons of RET disease MEN 2B Multifocal, bilateral Autosomal dominant Pheochromocytomas, mucosal Germline missense neuromas, megacolon, skeletal mutation in tyrosine abnormalities kinase domain of RET FMTC Multifocal, bilateral Autosomal dominant None Germline missense mutations in extracellular or intracellular cysteine codons of RET

MTC, medullary thyroid carcinoma; MEN, multiple endocrine neoplasia; FMTC, familial medullary thyroid carcinoma.

10% to 35%. Cutaneous lichen amyloidosis and Hirschsprung’s mutations at other codons are also observed (see Fig. 32.4). Some disease are infrequently associated with MEN 2A.57– 60 patients do extremely well for many years, even with distant me- MEN 2B appears to be the most aggressive form of hereditary tastases, while others develop inanition, symptomatic liver, lung MTC. In MEN 2B, MTC develops in all patients at a very young or skeletal metastases, as well as disabling diarrhea. Recurrence in age (infancy). All affected individuals develop neural gangliomas, the central neck, with invasion of the airway or great vessels, may particularly in the mucosa of the digestive tract, conjunctiva, lips, cause death. and tongue; 40% to 50% develop pheochromocytomas. Patients In patients with FMTC, MTC is usually indolent. These in- with MEN 2B may also have megacolon, skeletal abnormalities, dividuals most commonly have mutations of codons 609, 611, and markedly enlarged peripheral nerves. They do not develop 618, 620, 768, 804, or 891, although mutations of other codons hyperparathyroidism. have been identified (see Fig. 32.4). Many patients with FMTC FMTC is characterized by development of MTC in the ab- are cured by thyroidectomy alone, and even those with persistent sence of any other endocrinopathies. MTC in these patients has a elevation of calcitonin levels do well for many years. Occasion- more indolent clinical course. Some individuals with FMTC may ally, patients with FMTC survive into the seventh or eighth decade never manifest clinical evidence (i.e., symptoms or a lump in the without clinical signs of disease, although pathologic examination neck), although biochemical testing and histologic evaluation of of the thyroid will reveal MTC or C-cell hyperplasia.66 the thyroid demonstrates MTC.55,56 Risk-Reducing Thyroidectomy in RET RET Genotype-Phenotype Correlations Mutation Carriers

Mutations in the RET proto-oncogene are responsible for MEN Genetic counseling and informed consent should be obtained 2A, MEN 2B, and FMTC.61–64 This gene encodes a transmem- prior to genetic testing. Specific issues that should be covered in ge- brane tyrosine kinase protein.57,65 The mutations that cause the netic counseling sessions include explaining the patterns of herita- MEN 2 syndromes are activating gain-of-function mutations bility, likelihood of expression of different tumors, their prevention affecting constitutive activation of the protein. This is unusual and treatment, insurability, nonpaternity, survivor guilt, and others. among hereditary cancer syndromes, which are usually caused by It has been shown that RET mutation carriers may harbor foci loss-of-function mutations in the predisposition gene (e.g., famil- of MTC in the thyroid gland, even when calcitonin levels are nor- ial polyposis, BRCA1 and 2, von Hippel-Lindau, and MEN 1). mal.67 While the age of onset and rate of disease progression may More than 30 missense mutations have been described in patients differ, the lifetime penetrance of MTC is near 100% in carriers of affected by the MEN 2 syndromes (Fig. 32.4). RET mutations associated with MEN 2 syndromes. At-risk indi- There is a relationship between the type of inherited RET mu- viduals who are found to have inherited a RET gene mutation are tation and presentation of MTC. The most virulent form is seen therefore candidates for thyroidectomy, regardless of their plasma in patients with MEN 2B. These patients most commonly have a calcitonin levels. germline mutation in codon 918 of RET (ATG->ACG), although The best option for prevention of MTC in RET mutation other mutations have been described (codon 883 and 922). As carriers is complete surgical resection prior to malignant trans- noted previously, MTC in MEN 2B has an extremely early age formation. Prophylactic thyroidectomy prior to the development of onset (infancy). Despite its distinctive clinical appearance and of MTC is the goal in these patients. A number of studies have associated gastrointestinal difficulties, the disease is often not de- demonstrated improved biochemical cure rates and/or decreased tected until the patient develops a neck mass. Metastatic spread recurrence rates from early thyroidectomy, performed after posi- is usually present at the time of initial treatment, and calcitonin tive screening by calcitonin testing or RET mutation testing.68–70 levels often remain elevated postoperatively. MEN 2B mutations are the highest risk level, designated level MTC has a variable course in patients with MEN 2A, similar III (see Fig. 32.4).55,71 Patients with MEN 2B have the most ag- to that of sporadic MTC. Codon 634 and 618 mutations are the gressive form of MTC, with invasive disease reported in patients most common RET mutations associated with MEN 2A, although <1 year of age. These patients should have preventative surgery

tahir99 - UnitedVRG Chapter 32 Role of Surgery in Cancer Prevention 343

MEN 2A Risk MEN Codon FMTC HSCR Level 2B MTC Pheo HPT 533 I × × × 9-bp ins I* × 606 I* × 609 II* × × × × × Exons 8–11 611 II × × × × × Cysteine-rich × × × × × 618 II domain 620 II × × × × 630 II* × × × 631 I* × × × 634 II × × × ×

768 I × × × 777 I* × 790 I × × × Exons 13,14 791 I × × × × First tyrosine kinase domain 804 I × × × ×

804 III* × +806

883 III × Exons 15,16 891 I × × × Second tyrosine × 912 I* kinase domain 918 III ×

Figure 32.4 RET mutation sites associated with multiple endocrine neoplasia (MEN) 2 syndromes. Codons previously reported in AND SCREENING CANCER PREVENTION association with MEN-2 syndromes are listed by structural domain within the RET protein. Risk level is based on consensus guidelines or more recent clinical reports. Previously reported phenotypes for each codon are shown. MTC, medullary thyroid carcinoma; Pheo, pheochromocytoma; HPT, hyperparathyroidism; FMTC, familial medullary thyroid carcinoma; HSCR, Hirschsprung’s disease. Asterisk indicates risk level based on recent clinical reports not available at publication of the consensus guidelines. (From Traugott AL, Moley JF. The RET protooncogene. Cancer Treat Res 2010;153:303–319.)

early in the first year of life, if possible. Identification and pres- be enough published data to direct timing of interventions based ervation of parathyroid glands can be extremely difficult in these upon this information. As with the level II mutations, the need for infants, due to their small size, translucent appearance, and the central lymph node dissection should be guided by calcitonin levels presence of exuberant thymic and perithyroidal nodal tissue. and clinical features of the patient and kindred. These procedures should be performed by surgeons experienced Until recently, some groups recommended total thyroidectomy in parathyroid and/or pediatric thyroid operations. with central neck lymph node dissection and total parathyroi- Patients with MEN 2A with mutations in codons 634, 620, 618, dectomy with autotransplantation for all RET mutation carriers. and 611 are also considered high risk (level II).55,71 Patients with Recent studies and personal experience, however, have demon- level II mutations should undergo a total thyroidectomy at 5 to strated an extremely low likelihood of nodal metastases in patients 6 years of age. There is evidence that the risk of lymph node me- with MEN 2A or FMTC younger than 8 years of age, and in pa- tastasis is very low in patients with MEN 2A under the age of 8, tients with a normal calcitonin level.70 Our current strategy is to with normal calcitonin levels. Central lymph node dissection is leave the parathyroid in situ in these patients, if possible.74 Often, associated with higher risk of hypoparathyroidism, and recurrent however, the desired complete removal of thyroid tissue results laryngeal nerve injury and should be reserved for patients with in compromise of parathyroid blood supply. In these situations, elevated calcitonin levels. autotransplantation of devascularized parathyroid is required. A larger subset of RET mutations, associated with MEN 2A We routinely remove and autotransplant the parathyroid if a cen- and/or FMTC, is considered the lowest risk (level I).55,71 These in- tral node dissection is done. In parathyroid autotransplantation, clude mutations at codons 768, 790, 791, 804, and 891. For patients parathyroid glands are sliced into 1 mm × 3 mm fragments and with low-risk level I mutations, total thyroidectomy is recommended autotransplanted into individual muscle pockets in the muscle before age 5 to 10 years. This decision, however, regarding ideal of the nondominant forearm in patients with MEN 2A, or in the age at preventative thyroidectomy in low-risk mutation carriers, is sternocleidomastoid muscle in patients with FMTC or MEN 2B. currently being reviewed, and may be driven by additional clini- Patients are maintained on calcium and vitamin D supplementa- cal data such as the basal or stimulated serum calcitonin level.72,73 tion for 4 to 8 weeks postoperatively. There are no guidelines at present that address the issue of timing of In a recent series of thyroidectomies performed in 50 individuals surgery based on calcitonin level, and at present, pentagastrin (the with MEN 2A (identified by genetic screening), total thyroidec- primary calcitonin secretagogue used in testing) is not available in tomy and central node dissection with parathyroidectomy and para- the United States. It is anticipated that within a decade, there will thyroid autografting were performed in all patients (Fig. 32.5).70 344 Cancer Prevention and Screening

A B Figure 32.5 Total thyroidectomy specimen with attached central nodes from a patient with germline RET mutation and elevated calcitonin levels. Note small visible foci of medullary thyroid carcinoma (arrows).

All autografts functioned, but three patients required supplemental MTC. The importance of regular monitoring of patients’ compli- calcium. The percentage of individuals requiring calcium supple- ance with thyroid medication following thyroidectomy should not mentation following parathyroidectomy with parathyroid autograft- be underestimated. Children and teenagers are frequently non- ing reportedly ranges from 0% to 18%. Parathyroidectomy should compliant, and this can be determined by routine measurement be performed in all patients showing gross parathyroid enlarge- of thyroid-stimulating hormone levels. Continued noncompliance ment or biochemical evidence of parathyroid disease at time of sur- can result in growth problems. Occasionally, local human services gery. The operating surgeon should have expertise in preservation agencies may need to be involved in particularly difficult cases. of parathyroid function. It is important that the surgeon performing The term “biochemical cure” is used to refer to patients with an operative procedure for MTC be familiar with the techniques normal calcitonin levels after surgery for MTC. Complete post- described here. If not, the patient should be referred to a center operative normalization of calcitonin has been associated with de- where these procedures are routinely performed. creased long-term risk of MTC recurrence, though the evidence is Some patients with MEN 2 will be found to have elevated cal- less clear for a survival benefit. A persistent or recurrent elevation citonin levels prior to thyroidectomy. This is usually associated with in calcitonin indicates residual or recurrent MTC and warrants ad- medullary thyroid carcinoma or C-cell hyperplasia in the gland, ditional investigation by imaging. However, as most MTC has a and may be associated with lymph node metastases. Much has fairly indolent course, patients with biochemical evidence of recur- been written about the correlation between preoperative calcitonin rent disease may not have corollary imaging findings for some time. levels and extent of nodal involvement. It has been suggested that preoperative calcitonin level may guide the extent of node dissec- Conclusions tion. In a study of 300 European patients with MTC, node metasta- ses were not identified when the preoperative basal calcitonin level Identification ofRET gene mutations in individuals at risk for < 75 was 20 pg/ml. Involvement of nodal groups was correlated with developing hereditary forms of MTC has simplified manage- basal calcitonin level as follows: ipsilateral central and lateral neck ment, expanding the scope of indications for surgical interven- nodes (basal calcitonin >20 pg/ml), contralateral central nodes > tion. Patients who carry this mutation can be offered operative (basal calcitonin 50 pg/ml), contralateral lateral neck nodes treatment at a very young age, hopefully before the cancer has (basal calcitonin >200 pg/ml), and mediastinal nodes (basal calci- > developed or spread, and those identified as not having the muta- tonin 500 pg/ml). Based upon these findings, this group (who also tion are spared further genetic and biochemical screening. This wrote the European guidelines) recommends thyroidectomy only < achievement marks a new paradigm in surgery: the indication that if basal calcitonin is 20 pg/ml, ipsilateral central and lateral neck an operation be performed based on the results of a genetic test. dissection if the calcitonin is 20 to 50 pg/ml, and contralateral cen- As in the decision to perform any surgical procedure, meticulous tral neck dissection if the basal calcitonin is 50 to 200 pg/ml, with preparation and detailed discussion with patient and family must the addition of contralateral lateral neck dissection if the calcitonin precede the final recommendation. It is also important that the is 200 to 500 pg/ml. Most experts agree that sternotomy with medi- patient and family be involved in preoperative discussions with astinal neck dissection should be reserved for patients with image genetic counselors. Postoperative follow-up for compliance with evidence of mediastinal disease. In contrast, most North American thyroid medication is important, especially in children and teenag- surgeons rely heavily upon preoperative ultrasound imaging to map ers who are still growing and developing into adults. the extent of nodal involvement and determine extent of surgery based upon calcitonin and imaging results.55,74,76 FAMILIAL ADENOMATOUS POLYPOSIS, Follow-up MYH-ASSOCIATED POLUPOSIS, AND LYNCH SYNDROME Following thyroidectomy, thyroid hormone replacement is re- quired for life. Patients may need several weeks of oral calcium and Inherited CRC syndromes with multiple adenomatous polyps vitamin D until parathyroid function recovers. Intermittent calci- include FAP, MYH-associated polyposis (MAP), and LS. In some tonin testing may be done to monitor for persistent or recurrent cases, the diagnosis is suspected because of a striking family history

tahir99 - UnitedVRG Chapter 32 Role of Surgery in Cancer Prevention 345

Patient with multiple adenomatous polyps

Extensive No family Amsterdam polyposis history criteria

>10–20 <10 polyps polyps

– – APCMYH MMR/MSI Figure 32.6 Schematic demonstrating the potential genetic ++– + – workup for a patient with multiple adenomatous colorectal polyps and suspected of having an inherited colorectal cancer syndrome. APC, adenomatous polyposis coli; MMR, mismatch FAP Unknown MAP FCC X or Lynch repair; MSI, microsatellite instability; FAP, familial adenomatous unknown syndrome polyposis; MAP, MYH-associated polyposis; FCC X, familial colorectal cancer syndrome X. of CRC, while in others, suspicion arises from a very young onset At-risk individuals who belong to families with an AFAP phenotype of CRC or florid polyposis. should undergo colonoscopic screening every 2 to 3 years starting Although adenomatous polyp burden and family history may in their late teens. Informative genetic testing is possible in families suggest one syndrome over another, an initial negative genetic test with a demonstrated APC mutation, and mutations are detected result should be followed by further evaluation for other syndromes. in most pedigrees. However, approximately 25% of patients with For example, in clinical practice, a negative adenomatous polyposis FAP will have a de novo APC mutation.87 Severity of polyposis coli (APC) gene test in a patient with a suspected CRC syndrome is should be established during colonoscopy, as the timing of surgery followed by reflex testing for MAP and LS, as shown in Figure 32.6. and the risk of developing colorectal is dependent on the extent of FAP is an autosomal dominant syndrome that accounts for polyp burden. Patients with mild polyposis and a correspondingly <1% of the annual CRC burden, is caused by mutations in the lower CRC risk can undergo surgery in their late teens. Patients tumor-suppressor APC gene. It is characterized by the presence of with severe polyposis, a high degree of dysplasia, multiple adeno- AND SCREENING CANCER PREVENTION ≥100 adenomatous polyps in the colorectum, nearly 100% pen- mas >5 mm in size, and symptoms (bleeding, persistent diarrhea, etrance, and an inevitable risk of CRC if prophylactic colectomy is anemia, failure to thrive, psychosocial stress, etc.) should undergo not performed.8,77 Patients with a less severe form known as attenu- risk-reducing colorectal surgery as soon as is practical after diagno- ated FAP (AFAP) usually present with <100 colorectal adenomas sis.90,91 However, in carefully selected, fully asymptomatic patients that tend to be proximally located. MAP is an autosomal recessive who have small adenomas but a strong family history of aggressive syndrome that often presents phenotypically as attenuated polypo- abdominal desmoid disease, consideration can be given to delay- sis. While an estimated 2% of the general population are mono- ing prophylactic colectomy, as the risk of desmoid-related compli- allelic carriers of a mutated base-excision-repair MUTYH (MYH) cation may be greater than the risk of CRC development. gene, biallelic germline mutations may account for 9% to 18% of The three current surgical options for patients with FAP are patients with FAP or AFAP phenotypes who have no demonstrable total proctocolectomy (TPC) with permanent ileostomy, total co- APC mutation.78–80 lectomy with ileorectal anastomosis (IRA), and proctocolectomy LS accounts for 1% to 4% of all newly diagnosed CRC and is at- with ileal pouch-anal anastomosis (IPAA). IPAA can be a double- tributable to a germline mutation in one of the DNA MMR genes stapled, end-of-pouch-to-anus anastomosis, which may leave be- (MLH1, MSH2, MSH6, and PMS2).81–83 Epigenetic silencing of hind approximately 1 cm of anal transition zone. An alternative the MSH2 gene via a 3′-end deletion in EPCAM (TACSTD1), approach, which is preferred when there is carpeting of the anal a neighbor of MSH2 that plays a role in cell adhesion, also ac- transition zone with adenomas, is to perform a mucosal stripping counts for 20% to 25% of all suspected MSH2 cases and 1% to 6% of the anal transition zone down to the dentate line followed by of LS cases overall.84–86 LS is characterized by early age-of-onset a hand-sewn per anal anastomosis of pouch to the dentate line. CRC, predominance of lesions proximal to the splenic flexure, an Selection of the optimal procedure for an individual patient is increased rate of metachronous CRC, and a unique spectrum of based on several factors, including characteristics of the FAP syn- benign and malignant extracolonic tumors. Lifetime risk of CRC drome within the patient and family, differences in likely postop- in patients with LS may be as high as 80%.83,87 MSI reflects a defi- erative functional outcome, preoperative anal sphincter status, and ciency in DNA repair secondary to MMR gene mutation and is a patient preference.8 hallmark feature of LS-associated tumors. TPC with permanent ileostomy, although rarely chosen as a Variability in penetrance, phenotypic expression, and certainty primary procedure, is used in patients with invasive cancer involv- of disease development mandate distinctly different surgical ap- ing the sphincters or levator complex, or patients for whom an proaches in these three syndromes, including the type and timing IPAA is not technically feasible (secondary to desmoid disease and of risk-reducing colon and rectal surgery.88 foreshortening of the small bowel mesentery, making it surgically impossible to bring the ileal pouch to the anus) nor likely to lead to good function such as massive obesity or weak anal sphincters. Familial Adenomatous Polyposis However, TPC is occasionally chosen as a primary procedure by patients who perceive that their lifestyle would be compromised Surveillance of at-risk family members should begin around age 10 by the frequent bowel movements (five to six per day) sometimes to 15 years with an annual colonoscopy or flexible sigmoidoscopy.89 associated with the IPAA procedure. 346 Cancer Prevention and Screening

In addition to these issues, the key in deciding between an IPAA fulguration and polypectomy over many years can lead to difficulty and an IRA is based primarily on the risk of rectal cancer develop- with subsequent polypectomy, reduced rectal compliance, and ment if the rectum is left in situ. The risk of rectal cancer following difficulty identifying flat cancers in the background of scar tissue. IRA may range from 3% to 10% at 10 years, while the risk for a sec- The development of severe dysplasia and/or villous adenomas not ondary proctectomy for uncontrolled rectal polyposis ranges from amenable to endoscopic removal is indication for proctectomy. 10% to 61% at 20 years following initial colectomy with IRA.92–94 The magnitude of risk in an individual patient is, however, related Long-Term Considerations from Extracolonic to the overall extent of colorectal polyposis. IRA may be considered Manifestations for patients with <1,000 colorectal polyps (including those with attenuated FAP) and <20 rectal adenomas, as these individuals Despite the reduced risk of CRC-related death following prophy- have a relatively low risk of developing rectal cancer.88,93 Patients lactic colectomy, patients with FAP are still at increased risk of with severe rectal (>20 adenomas) or colonic (>1,000 adenomas) mortality from both rectal cancer and other causes relative to the polyposis, an adenoma >3 cm, or an adenoma with severe dys- general population. The three main causes of death following IRA plasia should ideally undergo a risk-reducing procedure that will are progression of desmoid disease, stomach and duodenal cancer, include a proctectomy.90,91,93 and perioperative mortality. Additional FAP-related extraintestinal The risk of secondary rectal excision, due to uncontrollable rec- manifestations include epidermoid cysts, supernumerary teeth, tal polyposis or rectal cancer, may be estimated by identifying the osteomas of the jaw and/or skull, congenital hypertrophy of the specific location of the causativeAPC mutation. Patients with mu- retinal pigment epithelium, cancers of the hepatopancreatobiliary 102–104 tations located between codons 1250 and 1464 have been shown tract and genitourinary tract, and thyroid cancer. to have a six-fold increased risk of developing rectal cancer, com- pared to those with mutations prior to codon 1250 or after codon Desmoids 8,92 1464 (mean number of rectal polyps 42 versus 22, respectively). Desmoids may occur in 10% to 25% of patients with FAP.105,106 Although the use of the genotype-phenotype relationship to guide 92 Unlike those found in the general population, FAP-associated des- patient management may be appealing, it is important to recog- moids tend to be intra-abdominal and arise following abdominal nize the variability of phenotypic expression that exists even among surgery.106,107 Although conflicting reports exist, it appears that fe- members of the same family. This suggests that at the current time, male patients, those with extracolonic manifestations of FAP, a pos- the choice between an IRA and an IPAA should be based primarily ′ 90 itive family history of desmoids, and APC mutations located at 3 of on clinical (rather than genetic) grounds. codon 1440 are at increased risk of developing desmoids.106,108,109 The risk of polyp and cancer development following primary These tumors often involve the small bowel mesentery as well as surgery is not limited to patients undergoing IRA. In patients under- the retroperitoneum and are often life-threatening due to invasion going IPAA, neoplasia may occur at the site of ileal pouch anasto- or compression of adjacent viscera. Further, recurrence and mor- mosis; the frequency appears to be greater after stapled anastomosis bidity rates are high following attempted resection, with recurrent (28% to 31%) than after mucosectomy and hand-sewn anastomo- 95 disease often more aggressive than the initial desmoid. Estimated sis (10% to 14%). In the case of neoplasia developing at the anal 5-year overall survival for patients with intra-abdominal desmoids transition zone after a stapled anastomosis, transanal mucosectomy causing severe symptoms such as significant pain and septic fis- may be performed, followed by advancement of the pouch to the tula/abscess, diameter >20 cm or rapidly growing, and/or need for dentate line. Of additional concern is the development of adenom- parenteral nutrition is only 53%.107 Therefore, desmoid resection atous polyps in the ileal pouch, which occurs in approximately 45% 96 is evaluated on an individualized case-by-case basis with surgery of patients by 10-year follow-up. Consequently, depending on reserved for highly select cases. polyp burden, lifetime endoscopic surveillance of the rectal rem- Desmoids that involve the small bowel mesentery may preclude nant (after IRA) every 6 to 12 months or the ileal pouch (after IPAA) 89 the formation of an IPAA secondary to foreshortening of the small every 1 to 3 years is required following either procedure. bowel mesentery, especially in patients undergoing proctectomy Another important consideration in choosing between IPAA after an initial IRA.110 Surgery for intra-abdominal and abdomi- and IRA is postoperative bowel function and quality of life. Some nal wall desmoids should be reserved for limited disease where the studies have associated IPAA with higher frequency of both daytime likelihood of clear margins is high. and nocturnal bowel movements, higher incidence of passive in- In symptomatic cases where resection of an intra-abdominal continence and incidental soiling, and greater postoperative mor- 97 desmoid may not be feasible, intestinal bypass or ureteral stenting bidity. However, long-term follow-up demonstrates a comparable may be necessary to alleviate bowel or urinary obstruction second- quality of life following IPAA for FAP relative to the patient’s pre- 98 ary to mass effect. In addition to surgical intervention, several medi- operative baseline. Therefore, although the choice of procedure cal options with variable efficacy are available for the management must be carefully individualized, because of the risk of rectal cancer of desmoid disease and include nonsteroidal anti-inflammatory associated with IRA, the authors favor IPAA for most patients with drugs (e.g., sulindac), selective estrogen receptor modulators (e.g., FAP whenever feasible. However, an IRA should be considered in tamoxifen), immunomodulators (e.g., imatinib, sorafenib, inter- specific circumstances, such as when there is mild rectal polyposis feron), doxorubin-based cytotoxic chemotherapy, and radiation. (as in AFAP), or a young patient with rectal sparing who is not in- terested in undergoing the multiple procedures that accompany an IPAA and diverting loop ileostomy, or a young woman interested MYH-Associated Polyposis in having children and trying to avoid the decreased fecundity as- sociated with an IPAA procedure.99 The use of minimally invasive MAP should be suspected in patients with >10 colorectal adeno- techniques such as laparoscopy may reduce the risk of infertility mas, a weak history of CRC, and no family history of FAP. The associated with IPAA.100,101 Though a diverting loop ileostomy diagnosis is confirmed by MUTYH (MYH) gene testing.80,88 should be performed in all IPAA procedures, it is not always feasible Depending on the polyp burden, the management of the colon due to a number of anatomic factors such as body habitus. and rectum of a patient with a biallelic MYH mutation can be Endoscopic surveillance of the rectal segment at 6- to 12-month endoscopic or surgical. If the polyp burden is limited and an en- intervals after the index surgery is recommended, with subsequent doscopic approach is pursued, colonoscopy should be performed surveillance frequencies dependent on the number and size of every 1 to 3 years.87,89 If the polyp burden is not amenable to an adenomas observed.89 Although small (<5 mm) scattered ad- endoscopic approach at the time of diagnosis, then a resection is enomas can be safely observed or removed with biopsy forceps, indicated. In most cases in which surgery is deemed necessary, an polyps >5 mm should be removed by snare. However, repeated IRA is sufficient. However, if rectal polyposis is severe, an IPAA

tahir99 - UnitedVRG Chapter 32 Role of Surgery in Cancer Prevention 347 may be indicated. Indications for surgery following an endoscopic of only 42% of LS mutation carriers.115 Families meeting Amster- surveillance program include increasing polyp size or number, or dam criteria but lacking an MMR mutation are referred to as hav- worsening histology. ing “familial colorectal cancer type X” and appear to have a lower Extracolonic manifestations of MAP are similar to FAP and in- incidence of colorectal and extracolonic cancers than those with a clude osteomas, desmoids, congenital hypertrophy of the retinal LS germline MMR mutation (see Fig. 32.6). Of note, they have an pigment epithelium, as well as cancers of the thyroid, ovary, blad- increased incidence of left-sided and nonmucinous microsatellite der, sebaceous gland, and breast. In addition, patients with MAP are stable tumors.77,88 also at a 4% lifetime risk of developing duodenal cancer and require Patients with CRC who belong to pedigrees suspicious for LS upper endoscopies every 1 to 3 years beginning as early as ages 18 to should be offered screening by IHC for loss of MMR protein ex- 20 years and starting no later than ages 30 to 35 years.87,89,111 pression or by MSI analysis. As the sensitivity of IHC testing for loss of MMR protein expression is comparable to MSI testing, 112 Lynch Syndrome either approach can be pursued. However, IHC testing is less ex- pensive and can also identify a specific MMR protein loss, which can help target subsequent germline testing. Routine IHC testing Due to the discordance associated with the term hereditary non- for loss of MMR protein in individuals younger than 50 years at the polyposis colorectal cancer, the use of this term has largely been time of CRC diagnosis is feasible and has led to the identification abandoned with reversion back to the eponym LS, which refers to of patients with LS who might otherwise have been missed.116,117 individuals with a predisposition to CRC and other malignancies Patients with MSI-high tumors should undergo testing for germ- as a result of a germline MMR mutation.112 Overall, CRC occurs line MMR mutations in MSH2, MLH1, MSH6, and PMS2. Reflex in up to 80% of patients with LS by their mid-40s.8,82 Endometrial IHC and/or MSI testing on all newly diagnosed CRC has been cancer occurs in 40% to 60%, gastric cancer in 11% to 19%, uri- advocated by some expert groups and has been successfully imple- nary tract cancer in 5% to 18%, and ovarian cancer in 9% to 15% mented at some institutions.83,112,118 However, a majority of cancer of affected individuals.8,82,87 programs nationwide currently do not have a protocol for reflex The Amsterdam criteria and revised Bethesda guidelines113 testing for LS, citing lack of institutional protocols as well as fear (Table 32.3) are used in clinical practice to identify patients at risk of nonreimbursement.119 As such, a unified move toward universal for LS who require further genetic evaluation. The Amsterdam testing remains some time away. In families for which tumor tissue criteria, which led to the identification of the LS-causing MMR is not available, initial germline testing may be considered though gene mutations require that there be: the financial burden is not insignificant, with the cost of finding a ■ Three relatives (one a first-degree relative of the other two) with single LS carrier measuring approximately $58,000 (compared to colorectal, endometrial, stomach, ovary, small bowel, ureteral/ the $5,000 spent in finding a single LS carrier using IHC screen- renal pelvis, brain, hepatobiliary, and/or sebaceous cancer; ing).120 As in FAP, a mutation in an affected individual must be ■ In two or more successive generations; established for testing in at-risk individuals to be conclusive. ■ With at least one case of cancer diagnosed before the age of 50; In lieu of universal testing, several predictive models such as the 114 ■ And that FAP as a diagnosis is excluded. MMRpredict, MMRpro, and PREMM1,2,6 have been devised in order to assess an individual’s likelihood of harboring LS.115,121,122 AND SCREENING CANCER PREVENTION Though the Amsterdam criteria can be used clinically to identify These models quantify an individual’s risk for carrying an MLH1, potential patients with LS, using it alone will result in identification MSH2, or MSH6 germline mutation by using clinical charac- teristics such as age at onset of CRC and/or other LS-associated cancers, location of CRC, family history, history of synchronous TABLE 32.3 or metachronous CRC, among others. A study of these predic- The Revised Bethesda Guidelines for Testing tive models demonstrated that they all performed better than the Colorectal Tumors for Microsatellite Instability revised Bethesda guidelines in terms of identifying patients with germline mutations for LS.123 The MMRpredict model appeared Tumors from individuals should be tested for MSI in the following to have to be the best predictor, with a sensitivity and specificity for situations: LS of 94% and 91%, respectively. Other validation studies, how- 1. Colorectal cancer diagnosed in a patient who is <50 y of age. ever, have not demonstrated the superiority of MMRpredict com- 124,125 2. Presence of synchronous, metachronous colorectal, or other pared to the other aforementioned models. It appears that the HNPCC-associated tumors,a regardless of age. use of clinical characteristics in combination with MSI or MMR 3. Colorectal cancer with the MSI-Hb histologyc diagnosed in a protein expression status in predictive models may potentially im- patient who is <60 y of age.d prove our ability to establish LS diagnoses in patients with CRC. 4. Colorectal cancer diagnosed in one or more first-degree However, the practicality and applicability of these tools in a clini- relatives with an HNPCC-related tumor, with one of the cal setting requires further assessment. cancers being diagnosed under age 50 years. Although development of CRC in LS is not a certainty, the 5. Colorectal cancer diagnosed in two or more first- or second- 80% lifetime risk, the 16% to 30% risk of metachronous CRC, degree relatives with HNPCC-related tumors, regardless of age. and the possibly accelerated adenoma-to-carcinoma sequence mandate consideration of prophylactic surgical options.82,87,126–129 MSI, microsatellite instability; HNPCC, hereditary nonpolyposis colorectal Patients with LS who have a CRC or more than one advanced cancer; MSI-H, microsatellite instability–high. adenoma should be offered the options of prophylactic total colec- a HNPCC-related tumors include colorectal, endometrial, stomach, ovarian, pancreas, ureter and renal pelvis, biliary tract, and brain (usually glioblastoma tomy with IRA or segmental colectomy with annual postoperative as seen in Turcot syndrome) tumors, sebaceous gland adenomas and surveillance colonoscopy. Careful surveillance is also necessary keratoacanthomas in Muir -Torre syndrome, and carcinoma of the small bowel. after total colectomy and IRA, as the risk of high-risk adenomas b MSI-H in tumors refers to changes in two or more of the five National and cancer in the retained rectum at a median of 104 months Cancer Institute–recommended panels of microsatellite markers. 126 c Presence of tumor-infiltrating lymphocytes, Crohn’s-like lymphocytic are 11% and 8%, respectively. Although there has been no reaction, mucinous/signet ring differentiation, or medullary growth pattern. study demonstrating an improved survival for patients with LS d There was no consensus among the workshop participants on whether to undergoing total colectomy and IRA versus segmental colectomy, include the age criteria in guideline 3; participants voted to keep <60 years of mathematical models suggest a slight survival benefit for total age in the guidelines. colectomy and IRA, especially for individuals under the age of From Umar A, Boland CR, Terdiman JP, et al. Revised Bethesda Guidelines 130,131 for hereditary nonpolyposis colorectal cancer (Lynch syndrome) and 30. In addition, because of increased rates of metachronous microsatellite instability. J Natl Cancer Inst 2004;96:261–268. CRC development and the risk of multiple abdominal surgeries 348 Cancer Prevention and Screening in those undergoing a segmental resection, a total colectomy and TABLE 32.4 IRA has emerged as the procedure of choice for the index cancer, with consideration for TPC in cases where a high risk of metachro- Prophylactic Total Abdominal Colectomy and nous rectal cancer can be predicted.126–128 Targeted genetic testing Ileorectal Anastomosis for Lynch Syndrome approaches—such as the single amplicon MSH2 A636P mutation Patients without Cancer test in Ashkenazi Jewish patients with CRC—have demonstrated how a rapid and inexpensive preoperative genetic test can help Pros direct the extent of colon resection.132 ■ Elimination of colon cancer risk LS mutation carriers with a normal colon and without a his- ■ Elimination of need for surveillance colonoscopy tory of CRC may also be offered prophylactic colectomy in highly ■ Alleviating patient anxiety over the prospect of colon cancer select situations. One rationale for this approach is the similar- development ity of lifetime cancer risk between patients with APC and MMR Cons gene mutations, and the fact that total abdominal colectomy ■ Persistence of risk of rectal cancer development with IRA produces less functional disturbance than the prophy- ■ Rectum still requires flexible endoscopic surveillance lactic procedure recommended for FAP (TPC with IPAA). How- ■ Patient anxiety of prospect of rectal cancer persists ever, an alternate strategy for these individuals is surveillance by ■ Possible altered bowel function colonoscopy, which is cost-effective and greatly reduces the rate ■ Risk of surgery and possible associated complications of CRC development and overall mortality.133 There is a risk of CRC development in the interval between colonoscopies, though most interval cancers tend to be early stage.134,135 As such, given that metachronous CRC may develop in as short a duration as a median of 11.3 months,136 the recommended interval for surveil- family, early age-of-onset in affected family members, functional lance colonoscopies is now every 1 to 2 years.89 While prophylac- and quality-of-life considerations, and likelihood of compliance tic colectomy is not routinely recommended, it may be indicated with surveillance. Table 32.4 lists some of the pros and cons of in highly select patients for whom colonoscopic surveillance is a prophylactic colectomy for germline mutation carriers for LS not technically possible or in those who refuse to undergo regu- without a history of CRC. Patients with LS and an index rectal lar surveillance. A decision analysis model suggests that prophy- cancer should be offered the options of TPC with IPAA or ante- lactic subtotal colectomy at age 25 may offer a survival benefit rior proctosigmoidectomy with primary reconstruction.128,138 The of 1.8 years, compared with surveillance colonoscopy. The ben- rationale for TPC is the 10% to 15% associated risk of metachro- efit of prophylactic colectomy decreases when surgery is delayed nous colon cancer in the remaining colon following the index until later in life and is negligible when performed at the time of rectal cancer. Choosing between the two procedures depends, in cancer development. 137 Thus, the decision between prophylactic part, on the patient’s willingness to undergo intensive surveillance surgery and surveillance for a gene-positive unaffected individual of the retained proximal colon, as well as issues regarding quality is based on many factors including penetrance of disease in the of life and bowel function.

SELECTED REFERENCES

The full reference list can be accessed at lwwhealthlibrary.com/oncology. 18. Bardram L, Hansen TV, Gerdes AM, et al. Prophylactic total gastrectomy in hereditary diffuse gastric cancer: identification of two novel CDH1 gene 1. Moyer VA, US Preventive Services Task Force. Risk Assessment, Genetic mutations-a clinical observational study. Fam Cancer 2014;13:231–242. Counseling, and Genetic Testing for BRCA-Related Cancer in Women: U.S. 19. Norton JA, Ham CM, Van Dam J, et al. CDH1 truncating mutations in Preventive Services Task Force Recommendation Statement. Ann Intern the E-cadherin gene: an indication for total gastrectomy to treat hereditary Med 2014;160. diffuse gastric cancer. Ann Surg 2007;245:873–879. 2. Robson ME, Storm CD, Weitzel J, et al. American Society of Clinical 21. Guilford P, Hopkins J, Harraway J, et al. E-cadherin germline mutations in Oncology policy statement update: genetic and genomic testing for cancer familial gastric cancer. Nature 1998;392:402–405. susceptibility. J Clin Oncol 2010;28:893–901. 22. Kaurah P, MacMillan A, Boyd N, et al. Founder and recurrent CDH1 mutations 3. Amir E, Freedman OC, Seruga B, et al. Assessing women at high risk of in families with hereditary diffuse gastric cancer. JAMA 2007;297:2360–2372. breast cancer: a review of risk assessment models. J Natl Cancer Inst 2010; 23. Benusiglio PR, Malka D, Rouleau E, et al. CDH1 germline mutations and 102:680–691. the hereditary diffuse gastric and lobular breast cancer syndrome: a multicen- 8. Guillem JG, Wood WC, Moley JF, et al. ASCO/SSO review of current role tre study. J Med Genet 2013;50:486–489. of risk-reducing surgery in common hereditary cancer syndromes. J Clin 25. Lee KH, Hwang D, Kang KY, et al. Frequent promoter methylation of CDH1 Oncol 2006;24:4642–4660. in non-neoplastic mucosa of sporadic diffuse gastric cancer. Anticancer Res 9. Saslow D, Boetes C, Burke W, et al. American Cancer Society guidelines for 2013;33:3765–3774. breast screening with MRI as an adjunct to mammography. CA Cancer J Clin 28. Huntsman DG, Carneiro F, Lewis FR, et al. Early gastric cancer in young, 2007;57:75–89. asymptomatic carriers of germ-line E-cadherin mutations. N Engl J Med 10. Domchek SM, Friebel TM, Singer CF, et al. Association of risk-reducing 2001;344:1904–1909. surgery in BRCA1 or BRCA2 mutation carriers with cancer risk and mortal- 30. Rogers WM, Dobo E, Norton JA, et al. Risk-reducing total gastrectomy for ity. JAMA 2010;304:967–975. germline mutations in E-cadherin (CDH1): pathologic findings with clinical 11. US Equal Employment Opportunity Commission. Genetic information dis- implications. Am J Surg Pathol 2008;32:799–809. crimination. http://www.eeoc.gov/laws/types/genetic.cfm. Accessed January 2, 31. Worster E, Liu X, Richardson S, et al. The impact of prophylactic total gas- 2014. trectomy on health-related quality of life: a prospective cohort study. Ann 12. Society of Surgical Oncology. Position statement on prophylactic mastectomy. Surg 2014;260:87–93. http://www.surgonc.org/practice-policy/practice-management/consensus-state 32. Lim YC, di Pietro M, O’Donovan M, et al. Prospective cohort study assessing ments/position-statement-on-prophylactic-mastectomy. Accessed February 1, outcomes of patients from families fulfilling criteria for hereditary diffuse gastric 2014. cancer undergoing endoscopic surveillance. Gastrointest Endosc 2014;80:78–87. 14. Eldor L, Spiegel A. Breast reconstruction after bilateral prophylactic mastec- 33. Mavaddat N, Peock S, Frost D, et al. Cancer risks for BRCA1 and BRCA2 tomy in women at high risk for breast cancer. Breast J 2009;15:S81–S89. mutation carriers: results from prospective analysis of EMBRACE. J Natl 15. Zakhireh J, Fowble B, Esserman LJ. Application of screening principles to Cancer Inst 2013;105:812–822. the reconstructed breast. J Clin Oncol 2010;28:173–180. 34. Risch HA, McLaughlin JR, Cole DE, et al. Prevalence and penetrance of 16. Gabram SG, Dougherty T, Albain KS, et al. Assessing breast cancer risk and germline BRCA1 and BRCA2 mutations in a population series of 649 women providing treatment recommendations: immediate impact of an educational with ovarian cancer. Am J Hum Genet 2001;68:700–710. session. Breast J 2009;15:S39–S45. 35. Struewing JP, Hartge P, Wacholder S, et al. The risk of cancer associated with 17. Nadauld LD, Ford JM. Molecular profiling of gastric cancer: toward person- specific mutations of BRCA1 and BRCA2 among Ashkenazi Jews. N Engl J alized cancer medicine. J Clin Oncol 2013;31:838–839. Med 1997;336:1401–1408.

tahir99 - UnitedVRG Chapter 32 Role of Surgery in Cancer Prevention 349

36. Finch AP, Lubinski J, Moller P, et al. Impact of oophorectomy on cancer in- 94. Koskenvuo L, Renkonen-Sinisalo L, Jarvinen HJ, et al. Risk of cancer and cidence and mortality in women with a BRCA1 or BRCA2 mutation. J Clin secondary proctectomy after colectomy and ileorectal anastomosis in familial Oncol 2014;32:1547–1553. adenomatous polyposis. Int J Colorectal Dis 2014;29:225–230. 37. Walsh T, Casadei S, Lee MK, et al. Mutations in 12 genes for inherited ovar- 97. Aziz O, Athanasiou T, Fazio VW, et al. Meta-analysis of observational studies ian, fallopian tube, and peritoneal carcinoma identified by massively parallel of ileorectal versus ileal pouch-anal anastomosis for familial adenomatous sequencing. Proc Natl Acad Sci U S A 2011;108:18032–18037. polyposis. Br J Surg 2006;93:407–417. 38. Couch FJ, Wang X, McGuffog L, et al. Genome-wide association study in 101. Beyer-Berjot L, Maggiori L, Birnbaum D, et al. A total laparoscopic approach BRCA1 mutation carriers identifies novel loci associated with breast and reduces the infertility rate after ileal pouch-anal anastomosis: a 2-center ovarian cancer risk. PLoS Genet 2013;9:e1003212. study. Ann Surg 2013;258:275–282. 40. Kauff ND, Satagopan JM, Robson ME, et al. Risk-reducing salpingo- 102. Steinhagen E, Guillem JG, Chang G, et al. The prevalence of thyroid cancer oophorectomy in women with a BRCA1 or BRCA2 mutation. N Engl J Med and benign thyroid disease in patients with familial adenomatous polyposis 2002;346:1609–1615. may be higher than previously recognized. Clin Colorectal Cancer 2012; 41. Rebbeck TR, Lynch HT, Neuhausen SL, et al. Prophylactic oophorectomy in 11:304–308. carriers of BRCA1 or BRCA2 mutations. N Engl J Med 2002;346:1616–1622. 104. Jarrar AM, Milas M, Mitchell J, et al. Screening for thyroid cancer in patients 44. Wethington SL, Park KJ, Soslow RA, et al. Clinical outcome of isolated se- with familial adenomatous polyposis. Ann Surg 2011;253:515–521. rous tubal intraepithelial carcinomas (STIC). Int J Gynecol Cancer 2013;23: 106. Nieuwenhuis MH, Lefevre JH, Bulow S, et al. Family history, surgery, and 1603–1611. APC mutation are risk factors for desmoid tumors in familial adenoma- 45. Bonadona V, Bonaiti B, Olschwang S, et al. Cancer risks associated with tous polyposis: an international cohort study. Dis Colon Rectum 2011;54: germline mutations in MLH1, MSH2, and MSH6 genes in Lynch syndrome. 1229–1234. JAMA 2011;305:2304–2310. 107. Quintini C, Ward G, Shatnawei A, et al. Mortality of intra-abdominal des- 48. Watson P, Vasen HF, Mecklin JP, et al. The risk of endometrial cancer in moid tumors in patients with familial adenomatous polyposis: a single center hereditary nonpolyposis colorectal cancer. Am J Med 1994;96:516–520. review of 154 patients. Ann Surg 2012;255:511–516. 49. Koornstra JJ, Mourits MJ, Sijmons RH, et al. Management of extracolonic 111. Nieuwenhuis MH, Vogt S, Jones N, et al. Evidence for accelerated colorec- tumours in patients with Lynch syndrome. Lancet Oncol 2009;10:400–408. tal adenoma—carcinoma progression in MUTYH-associated polyposis? Gut 51. Schmeler KM, Lynch HT, Chen LM, et al. Prophylactic surgery to reduce 2012;61:734–738. the risk of gynecologic cancers in the Lynch syndrome. N Engl J Med 2006; 112. Palomaki GE, McClain MR, Melillo S, et al. EGAPP supplementary evi- 354:261–269. dence review: DNA testing strategies aimed at reducing morbidity and mor- 54. Traugott AL, Moley JF. Multiple endocrine neoplasia type 2: clinical mani- tality from Lynch syndrome. Genet Med 2009;11:42–65. festations and management. Cancer Treat Res 2009;153:321–337. 113. Umar A, Boland CR, Terdiman JP, et al. Revised Bethesda Guidelines for 55. American Thyroid Association Guidelines Task Force, Kloos RT, Eng C, hereditary nonpolyposis colorectal cancer (Lynch syndrome) and microsatel- et al. Medullary thyroid cancer: management guidelines of the American lite instability. J Natl Cancer Inst 2004;96:261–268. Thyroid Association. Thyroid 2009;19:565–612. 116. Lee-Kong SA, Markowitz AJ, Glogowski E, et al. Prospective immunohisto- 56. Wells SA Jr, Pacini F, Robinson BG, et al. Multiple endocrine neoplasia chemical analysis of primary colorectal cancers for loss of mismatch repair type 2 and familial medullary thyroid carcinoma: an update. J Clin Endocri- protein expression. Clin Colorectal Cancer 2010;9:255–259. nol Metab 2013;98:3149–3164. 117. Steinhagen E, Shia J, Markowitz AJ, et al. Systematic immunohistochemistry 57. Eng C, Clayton D, Schuffenecker I, et al. The relationship between spe- screening for Lynch syndrome in early age-of-onset colorectal cancer patients cific RET proto-oncogene mutations and disease phenotype in multiple en- undergoing surgical resection. J Am Coll Surg 2012;214:61–67. docrine neoplasia type 2. International RET mutation consortium analysis. 118. Heald B, Plesec T, Liu X, et al. Implementation of universal microsatel- JAMA 1996;276:1575–1579. lite instability and immunohistochemistry screening for diagnosing lynch 70. Skinner MA, Moley JA, Dilley WG, et al. Prophylactic thyroidectomy in syndrome in a large academic medical center. J Clin Oncol 2013;31: multiple endocrine neoplasia type 2A. N Engl J Med 2005;353:1105–1113. 1336–1340.

71. Brandi ML, Gagel RF, Angeli A, et al. Guidelines for diagnosis and therapy 119. Beamer LC, Grant ML, Espenschied CR, et al. Reflex immunohistochem- AND SCREENING CANCER PREVENTION of MEN type 1 and type 2. J Clin Endocrinol Metab 2001;86:5658–5671. istry and microsatellite instability testing of colorectal tumors for Lynch syn- 72. Elisei R, Romei C, Renzini G, et al. The timing of total thyroidectomy in drome among US cancer programs and follow-up of abnormal results. J Clin RET gene mutation carriers could be personalized and safely planned on Oncol 2012;30:1058–1063. the basis of serum calcitonin: 18 years experience at one single center. J Clin 126. Kalady MF, McGannon E, Vogel JD, et al. Risk of colorectal adenoma and Endocrinol Metab 2012;97:426–435. carcinoma after colectomy for colorectal cancer in patients meeting Amster- 74. Moley JF. Medullary thyroid carcinoma: management of lymph node metas- dam criteria. Ann Surg 2010;252:507–511, discussion 511–513. tases. J Natl Compr Canc Netw 2010;8:549–556. 127. Parry S, Win AK, Parry B, et al. Metachronous colorectal cancer risk for mis- 81. Hampel H, Frankel WL, Martin E, et al. Screening for the Lynch syndrome (he- match repair gene mutation carriers: the advantage of more extensive colon reditary nonpolyposis colorectal cancer). N Engl J Med 2005;352:1851–1860. surgery. Gut 2011;60:950–957. 83. Vasen HF, Blanco I, Aktan-Collan K, et al. Revised guidelines for the clinical 128. Kalady MF, Lipman J, McGannon E, et al. Risk of colonic neoplasia after management of Lynch syndrome (HNPCC): recommendations by a group proctectomy for rectal cancer in hereditary nonpolyposis colorectal cancer. of European experts. Gut 2013;62:812–823. Ann Surg 2012;255:1121–1125. 86. Rumilla K, Schowalter KV, Lindor NM, et al. Frequency of deletions of 129. Cirillo L, Urso ED, Parrinello G, et al. High risk of rectal cancer and of meta- EPCAM (TACSTD1) in MSH2-associated Lynch syndrome cases. J Mol chronous colorectal cancer in probands of families fulfilling the Amsterdam Diagn 2011;13:93–99. criteria. Ann Surg 2013;257:900–904. 87. Jasperson KW, Tuohy TM, Neklason DW, et al. Hereditary and familial 132. Guillem JG, Glogowski E, Moore HG, et al. Single-amplicon MSH2 A636P colon cancer. Gastroenterology 2010;138:2044–2058. mutation testing in Ashkenazi Jewish patients with colorectal cancer: role in 88. Steinhagen E, Markowitz AJ, Guillem JG. How to manage a patient with presurgical management. Ann Surg 2007;245:560–565. multiple adenomatous polyps. Surg Oncol Clin N Am 2010;19:711–723. 133. Barrow P, Khan M, Lalloo F, et al. Systematic review of the impact of reg- 89. National Comprehensive Cancer Network. NCCN Clinical Practice Guidelines istration and screening on colorectal cancer incidence and mortality in in Oncology: Colorectal Cancer Screening. http://www.nccn.org/professionals familial adenomatous polyposis and Lynch syndrome. Br J Surg 2013;100: /physician_gls/pdf/colorectal_screening.pdf. Accessed February 19, 2014. 1719–1731. 90. Vasen HF, Moslein G, Alonso A, et al. Guidelines for the clinical manage- 134. Vasen HF, Abdirahman M, Brohet R, et al. One to 2-year surveillance in- ment of familial adenomatous polyposis (FAP). Gut 2008;57:704–713. tervals reduce risk of colorectal cancer in families with Lynch syndrome. 91. Church J. Familial adenomatous polyposis. Surg Oncol Clin N Am 2009;18: Gastroenterology 2010;138:2300–2306. 585–598. 138. Giardiello FM, Allen JI, Axilbund JE, et al. Guidelines on genetic evalua- 93. Sinha A, Tekkis PP, Rashid S, et al. Risk factors for secondary proctec- tion and management of Lynch syndrome: a consensus statement by the US tomy in patients with familial adenomatous polyposis. Br J Surg 2010;97: Multi-Society Task Force on Colorectal Cancer. Dis Colon Rectum 2014; 1710–1715. 57(8):1025–1048. 33 Cancer Risk Reducing Agentss

Dean E. Brenner, Scott M. Lippman, and Susan T. Mayne

WHY CANCER PREVENTION AS A CLINICAL be detected at the gross (oral premalignant lesions, polyps), micro- ONCOLOGY DISCIPLINE scopic (metaplasia, dysplasia), and/or molecular (gene loss or am- plification) levels. Recent molecular studies detecting profound Until recently, clinical oncology has been defined as a medical genetic alterations in histologically normal tissue from high-risk specialty that attempts to intervene in order to slow or reverse the individuals have provided strong support for the field carcinogen- final stage of the cancer process—the clonally derived, genomi- esis concept. The implication of the field effect is that multifocal, genetically distinct, and clonally related premalignant lesions can cally damaged, invasive cell mass. Cancer is a long process, a step- 1 wise carcinogenic progression that encompasses critical molecular progress over a broad tissue region. The essence of cancer risk events that culminate in the loss of key cellular control homeo- reduction, then, is intervention within the multistep carcinogenic static functions (e.g., control of proliferation, apoptosis, invasion, process and throughout a wide field. angiogenesis).1 These events occur prior to and during the mor- phologic changes that have historically defined neoplasia. Mor- phologic changes, such as subtle increases in cellular proliferation IDENTIFYING POTENTIAL CANCER RISK– that progress to early and late precancerous lesions containing dys- REDUCING AGENTS plastic cells, characterize the carcinogenesis process (Fig. 33.1).1–3 Opportunities for intervention in this process can include diverse, Cancer risk–reducing agent identification results from the syn- nonpharmacologic approaches (e.g., obesity management via diet/ thesis of data from population, basic, translational, and clinical lifestyle interventions) or pharmacologic interventions (e.g., drugs sciences. Findings from all of these disciplines are combined to or nutrients/nonnutrient substances used as drugs) aimed at delay- contribute to the identification of agents with the potential to delay ing or reversing the carcinogenesis process prior to or following the or reverse the carcinogenesis process (see Fig. 33.1). appearance of early morphologic changes. Cancer screening and The Hanahan and Weinberg hallmarks of malignant transfor- early detection strategies (e.g., surveillance endoscopy, fecal occult mation—self-sufficiency in growth signals, insensitivity to growth- blood testing, mammography) identify not only those individuals inhibitory signals, evasion of apoptosis, limitless replication potential, with early stage, curable malignant transformations, but also those sustained angiogenesis, and tissue invasion and metastasis7—reflect individuals with noninvasive neoplasias who are at risk for progres- the loss of cellular signaling control. The molecular damage that re- sion to transformed invasive malignancies. sults in transformation is triggered by a large array of genetic and Recognizing that cancer is a continuum, oncologists are in- environmental stressors such as chronic inflammation, oxidation, in- creasingly expected to be knowledgeable about a diverse array of herited genetic mutations, and exogenous environmental exposures. cancer-related topics including lifestyle behaviors such as diet and Many such signaling intermediates have common functions in mul- exercise, risk assessment, screening, other preventive interventions, tiple organ sites (see Fig. 33.1). The complexity and overlap of signal in addition to current treatments for advanced malignancy. The transduction pathways suggests that single molecular therapeutic/ understanding, use, and management of interventions designed to preventive targets may have limited effectiveness. Interventions delay or reverse the carcinogenesis process have become integral aimed at preventing the occurrence of or overcoming the effects of components of the this role.4 molecular defects in multiple pathways or targets may be required to arrest or reverse carcinogenesis. Using the Hanahan and Weinberg hallmarks, examples of possible targets are shown in Table 33.1. DEFINING CANCER RISK–REDUCING AGENTS (CHEMOPREVENTION) PRECLINICAL DEVELOPMENT OF CANCER Cancer risk reduction, commonly referred to as chemopreven- RISK–REDUCING AGENTS tion, is the use of a range of interventions from drugs to isolated dietary components to whole-diet modulation to block, reverse, or Similar to the development of therapeutic interventions, the as- prevent the development of invasive cancer.5,6 Human cancer risk sessment of efficacy and toxicity of single chemically synthesized reduction asserts that one can intervene at many steps in the car- entities, agents designed in silico, botanicals, nutrients/nonnutri- cinogenic process, which occurs over many years. This prolonged ent substances used as drugs for cancer risk–reducing agent ef- latency provides opportunities to intervene at many time points ficacy proceeds through a translational paradigm that identifies and at multiple events in the carcinogenic process. Successful efficacy in cell culture models, in live animal models, and in hu- deployment of cancer risk–reducing agent interventions requires mans. Preclinical models that simulate the carcinogenesis process evidence of reduced cancer-associated incidence and/or mortality. in target epithelia identify molecular biomarkers for modulation The concept of field carcinogenesis was first described in the by interventions. These models can be used to identify potential early 1950s as field cancerization in squamous cell carcinomas of toxicity of interventions and to assess the effect of interventions the head and neck, and subsequently ascribed to many epithelial on the development and progression of preneoplasia/neoplasia.8 sites. The field carcinogenesis concept is that patients have a wide The U.S. National Cancer Institute’s (NCI) PREVENT Cancer surface area of precancerous or cancerous tissue change that can Preclinical Drug Development Program is a prime example of a 350

tahir99 - UnitedVRG Chapter 33 Cancer Risk Reducing Agents 351

Molecular Biomarkers of Carcinogenesis Dysplasia = Intraepithelial Neoplasia (IEN) Normal Initiated Mild Moderate Severe CIS Cancer

Prostate AR, SRD5A2, CYP, GSTP1 AR, GSTP1, TERT, NKX3.1, 8p, 13q, 10q, 167q, 7p p53, VEGF, FGF, Polymorphism 7q, Xq, DNA Ploidy, IGF, EGFR, HER-2, PCNA, Ki67 Cadherins, Genetic susceptibility to infection MMPs, PSA

Colon APC, BCL-2, RAS, SMAD 2, p53, p16, 8p, tPA, p15, Bub1, c-MYC COX-2 SMAD 4, DCC, 7q, VEGF, MMP, CEA, 22q, CD44 Hypomethylation STAT3 Cyclin D1 E-Cadherin

Breast E2 Metabolism DNA Adducts, p53, Cyclin D1, ERB-B2, EGFR, Angiogenesis, Cyt P450, ER, PR, Genomic instability, BRCA1, 2, IGF, VEGF, RXR, NM23 Collagenase, FGF DNA Repair Thrombospondin Aneuploidy

Lung 3p, 9p, 13q, 15p, P16 53, K-RAS, c-myc, 22q, 18q, β-Catenin

Head & Neck 3p, 9p, 53q, FHIT Cyclin D1, EGFR, COX-2 6p, 8p23, 4q26-q28 p16, p19 CANCER PREVENTION AND SCREENING CANCER PREVENTION

Esophagus p16, p53, DNA Content EGFR, VEGFR, Cyclin D1, APC, TGFα, VEGF, Cadherin

Liver HBV, HCV, TGF, IGF-2, TNF-2, IL6, Telomerase, c-MYC, p53, Rb, IGF2-R, PTEN, Carcinogen/ Genomic instability DLCI, p73, E-Cadherin, Cyclin D, Cyclin E, DNA Adducts p16, p21, Aberrant methylation

Figure 33.1 Genetic progression in major cancers. Carcinogenesis is driven by genetic progression. This progression is marked by the appearance of molecular biomarkers in distinctive patterns representing accumulating changes in gene expression and correlating with changes in histologic phenotype as cells move from normal through the early stages of clonal expansion to dysplasia and finally to early invasive, locally advanced, and metastatic cancer. The figure257 shows candidate molecular biomarkers of genetic progression in seven target organs: the prostate,300–302 the colon,2 the breast,303,304 the lung,305–307 the head and neck,308–311 the esophagus,312,313 and the liver.314 CIS, carcinoma in situ; AR, androgen receptor; CYP, cytochrome P - 450; GSTP1, glutathione S transferase P1; TERT, telomerase reverse transcriptase; NKX3.1, NK 3 transcription factor related, locus 1 (prostate specific, androgren regulated); IGF, insulin-like growth factor; EGFR, epidermal growth factor receptor; HER-2, human epidermal growth factor receptor-2; PCNA, proliferating cell nuclear antigen; VEGF, vascular endothelial growth factor; FGF, fibroblast growth factor; MMP, matrix metalloproteinase; PSA, prostate- specific antigen; APC, adenomatous polyposis coli; BCL - 2, B-cell lymphoma 2 gene, apoptosis control; c-MYC, v-myc avian myelocytomatosis viral oncogene homolog; COX-2, cyclooxygenase-2; SMAD, homolog of mothers against decapentaplegic + C. Elegans SMA protein; DCC, deleted in colon cancer gene; CEA, carcinoembryonic antigen; ER, estrogen receptor; PR, progesterone receptor; ERB-B2, Receptor tyrosine-protein kinase erbB-2, same as HER-2; RXR, retinoid X receptor; NM23, Nucleoside disphosphate kinase A; K-RAS, Kirsten rat sarcoma viral oncogene homolog; FHIT, fragile histidine triad protein; TGFα, tumor growth factor alpha; HBV, hepatitis B virus; HCV, hepatitis C virus; TNF-2, tumor necrosis factor 2; IL6, interleukin 6; PTEN, phosphatase and tensin homolog. (Figure and revised caption from Kelloff GJ, Lippman SM, Dannenberg AJ, et al. Progress in chemoprevention drug development: the promise of molecular biomarkers for prevention of intraepithelial neoplasia and cancer—a plan to move forward. Clin Cancer Res 2006;12:3661–3697, published with permission from the American Association for Cancer Research.)

rational strategy to select promising agents for clinical trials through These in vitro assays are rapidly completed for potential cancer risk– a stepwise approach of preclinical in vitro testing followed by in vivo reducing agents. Examples of such assays include carcinogen-DNA screening.8–10 This system involves several phases: biochemical pre- binding, prostaglandin synthesis inhibition, glutathione–S-transferase screening assays, in vitro efficacy models, in vivo short-term screen- inhibition, and ornithine decarboxylase inhibition. ing, animal efficacy testing, and preclinical toxicology testing. In Vitro Efficacy Models Biochemical Prescreening Assays In vitro assays test the cancer risk–reducing activity of a screened Prescreening assays are a series of short-term, mechanistic assays devel- compound in four epithelial cell systems (primary rat tracheal oped to evaluate the ability of a test compound to modulate biochem- epithelial cells, human lung tumor [A427] cells, mouse mammary ical events presumed to be mechanistically linked to carcinogenesis.8 organ cultures [MMOC], and human foreskin epithelial cells). 352 Cancer Prevention and Screening

TABLE 33.1 after carcinogen exposure measures activity during tumor promo- tion. Three of these assays (using rat tracheal epithelial, A427, and Molecular Mechanisms Common to Transforming MMOC cells) have shown predictive values of 76% to 83% for can- Cells and Potential Preventive Interventions cer risk–reducing agent efficacy in in vivo models.8 Characteristics of Neoplasia Possible Molecular Targets ■ Self-sufficiency in cell EGFR, platelet-derived growth Preclinical In Vivo Models for Cancer Risk– growth factor, MAPK, PI3K Reducing Agent Efficacy Testing ■ Insensitivity to antigrowth SMADs, pRb, cyclin-dependent signals kinases, MYC Animal models remain a crucial link in the efficacy assessment of cancer risk–reducing agents for epithelial cancer. Chemical ■ Limitless replicative hTERT, pRb, p53 carcinogenesis models provide the reproducible development of potential tumors in animals following the administration of a known chemi- ■ Evading apoptosis Bcl-2, BAX, caspases, Fas, tumor cal initiator or combination initiator/promoter and have been the necrosis factor receptor, insulin primary in vivo screening tool for cancer risk–reducing agents 12,13 growth factor/PI3K/Akt, mTOR, (Table 33.2A). Carcinogenesis models employing genetically p53, NF-κB, PTEN, RAS engineered mice permit the interrogation of targeted pathways and the corresponding efficacy of cancer risk–reducing agents. Al- ■ Sustained angiogenesis VEGF, basic fibroblast growth though useful for mechanistic studies, knockout or genetic muta- factor, integrins (αvβ3), α tional models create accelerated neoplastic progression that does thrombospondin-1, HIF-1 not accurately recapitulate the more complex, stepwise, human car- ■ Tissue invasion and MMPs, MAPK, E-cadherin cinogenesis process. Recombinant alleles can be driven by the ad- metastases dition of drug-sensitive regulatory elements, such as tetracycline or tamoxifen analogs. The drug-sensitive regulatory elements achieve EGFR, epidermal growth factor receptor; MAPK, mitogen-activated protein temporal control over a gene promoter through the administration kinase; PI3K, phosphoinositide 3-kinase; SMAD, drosophila protein, mothers against decapentaplegic gene and the Elegans protein SMA; pRb, of the drug that binds to the regulatory element. Such a system phosphorylated Rb protein; hTERT, human telomerase reverse transcriptase; permits the inhibition or overexpression of the organ-specific gene mTOR, mammalian target of rapamycin; NF-κB, nuclear factor kappa B; using Cre recombinase, a site-specific DNA recombinase that tar- PTEN, phosphatase and tensin homolog; VEGF, vascular endothelial growth gets DNA regions flanked by loxP sequences. Tables 33.2A and 2B factor; HIF-1α, hypoxia-inducible factor-1α; MMP, Matrix metalloproteinases. Adapted from Kelloff GJ, Lippman SM, Dannenberg AJ, et al. Progress in list representative organ-specific chemical and transgenic mouse chemoprevention drug development: the promise of molecular biomarkers models that may be used for cancer risk–reducing agent testing. for prevention of intraepithelial neoplasia and cancer—a plan to move forward. Clin Cancer Res 2006;12:3661–3697 and derived from Hanahan D, Weinberg RA. The hallmarks of cancer. Cell 2000;100:57–70. CLINICAL DEVELOPMENT OF CANCER RISK–REDUCING AGENTS

The assays measure the ability of potential cancer risk–reducing agents to reverse transformation in normal epithelial cells exposed Special Features of Cancer Risk–Reducing to carcinogens. For example, after treatment with a carcinogen such Agent Development as 7,12-demethylbenz(a)anthracene, MMOCs develop lesions sim- ilar to alveolar nodules that are considered precancerous in mouse The clinical efficacy assessment of cancer risk–reducing mammary glands in vivo.11 Pretreatment of organ cultures before agents employs phased testing (phase I to III) models used for carcinogen exposure measures the effect of cancer risk–reducing development of drugs14 but with crucial differences in study design agents in the initiation stage of carcinogenesis, whereas treatment and end points. Special features for the clinical development of

TABLE 33.2A Chemical Carcinogenesis Models Used for Screening of Cancer Risk–Reducing Agents for Common Epithelial Neoplasms in Animals

Organ Site Species Carcinogen End Point Colon Rat, mouse Azoxymethane (AOM) Aberrant crypts, adenomas, adenocarcinomas Lung Mouse N—butyl-N-(4-hydroxylbutyl)nitrosamine (NNK); Adenomas, adenocarcinomas benzo[a]pyrene; cigarette smoke Hamster Methylnitrosourea (MNU) Squamous cell carcinomas Mouse N-nitroso-tris-chloroethylurea Squamous cell carcinomas Breast Rat Dimethylbenz[a]anthracene (DMBA); MNU Adenocarcinomas, adenomas Prostate Rat MNU + testosterone Adenocarcinomas Bladder Rat, mouse N-butyl-N-(4-hydroxybutyl)nitrosamine (OH-BBN) Transitional cell carcinoma Pancreas Hamster N-nitrobis-(2-oxopropyl)amine (BOP) Ductal carcinomas Head and neck Rat 4-nitroquinoline-1-oxide (4-NQO) Tongue squamous cell carcinomas Mouse DMBA Squamous cell carcinomas Esophagus Rat Dimethyalbenz[a]anthracene Squamous cell carcinomas Rat Esophagogastroduodenal anastomosis + iron Adenocarcinomas

Adapted from Steele VE, Lubet RA. The use of animal models for cancer chemoprevention drug development. Semin Oncol 2010;37:327–338.

tahir99 - UnitedVRG Chapter 33 Cancer Risk Reducing Agents 353

TABLE 33.2B Selected Transgenic Animal Models for Carcinogenesis Evaluation

Organ Site Genes Targeted End Point References Colon Apc, Lrig1, gp130, Stat3, Smad3, Wnt-␤-catenin, Adenomas and adenocarcinomas (258, 259) villin, TGFBR2, Kras, Ink4a Lung KrasG12D, KrasG12Vgeo, PTEN, BrafV600E, cRaf, Adenomas and adenocarcinomas (260) Egfr L858R±T790M, PIK3CA, EMLA4-ALK fusion Rb. P53 Small cell Breast Mouse mammary tumor virus long terminal repeat DCIS, adenocarcinomas (261) promoter (MMTV) driven BRCA1, p53, ER␣, aromatase, TGF␣, Her2/neu, wnt, PELP-1, AIB-1 Prostate Probasin promotor driving SV40 large T antigen Prostate intraepithelial neoplasia, (262) (TRAMP/LADY), c-Myc, TMPRSS-ERG, Akt, neuroendocrine tumors (TRAMP); Wnt-␤-catenin, androgen receptor adenocarcinomas (c-Myc, TMPRESS-ERG, Akt, Wnt, androgen receptor) Nkx3.1, FGFR1, TGF, PTEN Adenocarcinomas Pancreas KrasG12D alone, LSL-Kras, PDX-1, R26Notch, Tif1␥ Pancreatic intraepithelial neoplasia (263) Combined PDX-1, LSL-Kras, LSL-Trp53; combined Pancreatic adenocarcinoma PDX-1, Brca2, LSL-Kras Trp53; KrasG12D on Mist1 locus; PDX-1, KrasG12D+ Ink4a/Arf or Smad4; Ptf1a, KrasG12D, TGFBR2

Note: Most models are mouse models. Such models will permit efficacy testing of specific pathway targets using single agent interventions, combinations, or multimechanism-based natural products.

cancer risk–reducing agents create the following challenges to be neoplasia has served and continues to serve as a biomarker for inva- overcome: (1) the need for large therapeutic index (doses associ- sive malignancy (Table 33.3). Although many advocate the use of ated with potential toxicity of an intervention need to substantially intraepithelial neoplasia-based biomarkers as regulatory surrogate exceed doses aimed at delaying or reversing transformation) for end points, others caution that intraepithelial neoplasias may not use in individuals who are asymptomatic yet may benefit from an serve as sufficiently robust surrogate biomarkers for cancer inci- CANCER PREVENTION AND SCREENING extended (years) treatment course; (2) the long latency to malig- dence or mortality.24 nant transformation (an assessment of effectiveness based on the In order to be useful as end points for cancer risk–reducing agent reduction in cancer incidence requires studies lasting for years and efficacy testing as regulatory end points, any biomarker must have involving thousands of participants); (3) adherence (once-daily statistical accuracy, precision, and effectiveness of results24 that dem- dosing regimens using interventions that have sufficiently long onstrate prediction of a hard disease end point—cancer incidence or half-lives may minimize the impact of a missed dose yet maintain mortality. An independent validation data set must address defined the biologic impact on the physiologic target; minimal toxicity standards of validation that minimize bias in the study design and the and strong psychological commitment to preventive goals also en- populations studied.25 The biomarker must be generalizable to the hance adherence15); and (4) complex risk assessment for cancer specific clinical or screening population (Table 33.4). (individuals with highly penetrant but infrequent, germ-line ge- 16,17 netic susceptibility to breast and colon cancers are excellent Phases of Cancer Risk–Reducing Agent candidates for cancer risk–reducing agents and are likely to accept some toxicity for reduced cancer risk). For individuals at more Development modestly increased risk (e.g., long-term, current smokers; persons with a family history of cancer; women with mammographically Phase I cancer risk–reducing agent trials define an optimal cancer dense breasts), quantitative risk assessment algorithms may be use- risk–reducing agent dose. An optimal cancer risk–reducing agent ful in the future to identify optimal cancer risk–reducing agents. dose is one that is usually nontoxic, scheduled once daily, and mod- The refinement of cancer risk calculators for breast,18 colon,19 and ulates a tissue, cellular, or serum biomarker of drug activity (e.g., prostate cancer20 promises to appropriately select high-risk individ- the dose of aspirin that inhibits prostaglandin production in a target uals for cancer risk–reducing agents such that anticipated benefits tissue site). The definition of a maximum tolerated dose is not an exceed potential risks. essential end point of a phase I cancer risk–reducing agent trial. Higher, yet nontoxic doses may lower cancer risk–reducing agent efficacy. For example,β -carotene at high doses has pro-oxidant ac- Biomarkers as Cancer Risk–Reducing Agent tivity and may enhance the carcinogenesis process, whereas at low Targets and Efficacy End Points doses, it is a potent antioxidant and differentiating agent.26 Phase II cancer risk–reducing agent trials begin to define cancer A biomarker is a characteristic that is measured and evaluated as risk–reduction efficacy. These short-term (6 months to 1 year) treat- an indicator of normal biologic processes, pathogenic processes, ment periods gather evidence of risk reduction by assessing drug or pharmacologic responses to therapeutic interventions.21 A sur- effects on tissue, cellular, or blood surrogate markers of carcino- rogate end point for cancer prevention assumes that a measured genesis. Phase IIa trials are nonrandomized, biomarker modulation biologic feature will predict the presence or future development of trials. Phase IIb trials are randomized, placebo-controlled trials of a cancer outcome.22 Biomarkers enable a reduction in the size and several hundred subjects testing, for example, whether a risk-reduc- duration of an intervention trial by replacing a rare or distal end ing agent reduces recurrence of a previously resected intraepithe- point with a more frequent, proximate end point.23 Intraepithelial lial neoplastic lesion as the primary end point. Cellular dynamic 354 Cancer Prevention and Screening

TABLE 33.3 Using such a definition, a clinical trial with an end point of re- duction in adenoma recurrence is considered a phase III trial. Common Intraepithelial Neoplasias Other investigators define phase III cancer risk–reducing agent Epithelium Intraepithelial Neoplasia References trials as randomized, controlled clinical trials with a cancer inci- dence or mortality end point. This controversy causes confusion Colon and rectum Adenoma (264) in the literature. For the purpose of clarity in this textbook, the Lower esophagus Barrett esophagus (265) latter definition of phase III trial is used—a prospective, random- ized, controlled clinical trial with a cancer incidence or mortality Upper esophagus Squamous dysplasia (266, 267) end point. Randomized, controlled clinical trials with a surrogate Skin: Squamous/ Actinic keratosis (268) biomarker end point such as an intraepithelial neoplasia (e.g., ad- basal cell enoma) are defined as phase IIb cancer risk–reducing agent trials. Skin: Pigmented Dysplastic nevus (269) Cervix Cervical intraepithelial (270) MICRONUTRIENTS neoplasia Head and neck Leukoplakia/oral epithelial (271) Definition dysplasia Prostate Prostate intraepithelial (272) Micronutrients comprise a large, diverse group of molecules typi- neoplasia (PIN), intraductal cally ingested as part of the diet that play roles in normal human carcinoma of the prostate biology. This group of compounds has been investigated exten- Lung Bronchial dysplasia (273) sively as cancer risk–reducing agents in purified forms (i.e., supple- ments), as components of multiagent cocktails, and occasionally, Pancreas Pancreatic intraepithelial (274) as components of food extracts/other mixtures. Although neoplasia are not micronutrients per se, they are related to retinol ( vitamin A) and share certain properties with carotenoids, which are diet de- rived and, therefore, included along with micronutrients. (e.g., proliferation, apoptotic index), biochemical, or molecular (e.g., p53, cyclin D) end points may be used as secondary end Retinoids, Carotenoids, and Antioxidant points. Preoperative or window of opportunity trials enroll subjects Nutrients for brief study periods prior to obtaining tissue by a planned resec- tion of an invasive neoplasm. Such designs permit the exploration of biomarker modulation in the invasive neoplasm and in contigu- Overview and Mechanisms 27 ous epithelial fields proximal and distal to the invasive neoplasm. Retinoids are the natural derivatives and synthetic analogs of vita- Phase III cancer risk–reducing agent trials define reduction in a min A.32 Cancer risk–reduction intervention studies have evaluated hard cancer end point such as cancer incidence or mortality. Such the parent compound (retinol, typically given as retinyl acetate or trials, using large, higher risk populations in a randomized, double- palmitate), naturally occurring retinoids such as all-trans–retinoic blinded intervention, are designed to identify a standard of preven- acid (ATRA) and 13-cis-retinoic acid (13cRA), and also synthetic tive care for a given risk population. For example, trials of tamoxifen retinoids such as etretinate and fenretinide (4-hydroxy[phenyl]ret- 28,29 for the reduction of breast cancer incidence, finasteride for the re- inamide [4HPR]). These agents have been of interest for cancer 30 duction of prostate cancer incidence, and β-carotene for the reduc- risk reduction for decades. Mechanistically, retinoids have been 31 tion of lung cancer incidence serve as examples of well-conducted, shown to modulate cellular growth and differentiation, as well as definitive phase III cancer risk–reducing agent clinical trials. apoptosis.32 A large body of research indicates that retinoids have Some investigators consider randomized, controlled clinical activity in the promotion and progression phases of carcinogenesis, trials with an end point sufficient for regulatory review as phase III. including extensive evidence of efficacy in the setting of premalig- nant lesions, leading to their evaluation in human Phase III trials. Nuclear retinoic acid receptors mediate many of the retinoid- signaling effects; however, retinoids interact with other signaling TABLE 33.4 pathways, such as estrogen signaling in breast cancer.33 Characteristics of Biomarkers for Use as Carotenoids are a group of naturally occurring plant pigments, End Points in Cancer Risk–Reducing Agent only some of which are found in appreciable levels in the human Efficacy Assessment diet and human tissues, including beta-carotene, alpha-carotene, lycopene, lutein, and β-cryptoxanthin.34 Of these, the most widely ■ Variability of expression between phases of the studied carotenoids for cancer risk reduction are beta-carotene and lycopene. Beta-carotene has the highest pro–vitamin A activity of carcinogenesis process β ■ Detected early in the carcinogenesis process the carotenoids, but alpha-carotene and -cryptoxanthin also pos- ■ Genetic progression or protein pathway based sess pro–vitamin A activity. Other carotenoids, such as lycopene, do ■ not possess vitamin A activity but are known to have potent antioxi- Target of modulation by preventive interventions 35 ■ dant activity, particularly with regard to singlet oxygen quenching. Changes in biomarker linked to reduction in incident cancer of 36 epithelial target Furthermore, eccentric cleavage products of beta-carotene, as ■ Changes in biomarker linked to clinical benefit well as other non-pro–vitamin A carotenoids such as lycopene (e.g., ■ apocarotenals, apocarotenoic acids) appear to be biologically active Can be quantified directly or via closely related activity such as 37 a downstream target or upstream kinase and may also act via retinoid-signaling pathways. ■ Measurable in an accessible biosample (preferably urine, Because of the known antioxidant function of carotenoids, they serum, saliva, stool, or breath) are often studied for risk-reducing efficacy in combination with ■ High throughput, technically feasible, analytical procedure with other antioxidant nutrients, especially vitamins E and C and sele- strong quality assurance/quality control procedures nium (sometimes as a cocktail, versus placebo). Thus, we will con- ■ Cost-effective sider this group of nutrients first, followed by other micronutrients that are thought to act via different mechanisms and/or pathways.

tahir99 - UnitedVRG Chapter 33 Cancer Risk Reducing Agents 355

Epidemiology Supplemental beta-carotene has also been studied as a single agent and in combination with other agents for the prevention A large body of literature indicates that people who consume of second primary cancers of the mouth and throat (Table 33.6). greater amounts of carotenoids from foods (primarily fruits and One trial of beta-carotene alone observed no harm or benefit47; an- vegetables) and people with higher serum or plasma levels of vari- 38 other observed nonsignificantly fewer second head and neck can- ous carotenoids have a lower risk for various cancers. In particu- cers but more lung cancers48; and a third gave beta-carotene with lar, according to the systematic review of the literature conducted α-tocopherol (400 IU per day).49 In the third trial, beta-carotene by the World Cancer Research Fund/American Institute for Can- 38 was discontinued early due to adverse findings from lung cancer cer Research, foods containing carotenoids are “probably” as- prevention trials (see the following); however, after a median fol- sociated with lower risks of cancers of the mouth/pharynx/larynx low-up of 6.5 years, all-cause mortality was increased, which the and lung; foods containing beta-carotene are “probably” associated authors attributed to the supplemental α-tocopherol. As will be with lower risks of cancers of the esophagus; and foods containing discussed later, adverse effects of antioxidant nutrients are not lim- lycopene are “probably” associated with lower risks of prostate can- ited to head and neck cancer patients; potential mechanisms for cer. Preformed retinol intake is inconsistently associated with risks adverse effects are discussed further. of various cancers but that, in part, likely reflects confounding, be- cause dietary sources of preformed retinol primarily include foods Clinical Efficacy in the Lung (Lower Airway) of animal origin (e.g., liver, eggs, milk). Vitamin C in the diet comes primarily from the consumption of fruits and vegetables; Reversal of Metaplasia or Dysplasia. Active smokers and therefore, vitamin C and carotenoids often trend together in epi- recent quitters have multiple preinvasive metaplastic and dys- demiologic findings. Vitamin E, in contrast, is found in different plastic lesions in the pulmonary tree. Most of these lesions re- foods, especially nuts, seeds, and vegetable oils; intake and blood solve upon smoking cessation, but some remain and progress concentrations are somewhat inconsistently associated with can- 39 to invasive neoplasms. Unfortunately, micronutrient or retinoid cer risk. Selenium, being a trace mineral, is difficult to measure interventions have not demonstrated preventive efficacy in most in the diet, but higher selenium status has been associated with a rigorous trials in patients with early lesions. For example, a US lower risk of certain cancers, although the results are not entirely 39 trial randomized 755 asbestos workers to receive beta-carotene consistent. (50 mg per day) and retinol (25,000 IU every other day) versus placebo; sputum atypia was not reduced after 5 years.50 As an- Preclinical In Vivo Models other example, eligible smokers with lung metaplasia or dysplasia In preclinical models, retinoids induce differentiation as well as were randomized to 6 months of 13cRA or placebo. The extent arrest proliferation33 of various cancers, making them attractive of metaplasia decreased similarly (in approximately 50% of sub- agents for cancer risk reduction.40 The International Agency for jects) in both study arms.51 Only smoking cessation was associated Research on Cancer reviewed the preclinical research involving with a significant reduction in the metaplasia index during the beta-carotene, concluding that there was “sufficient” evidence of 6-month intervention. cancer preventive activity, particularly involving mouse skin tumor models and the hamster buccal pouch model.41 Notably, there was Prevention of Invasive Neoplasms. Large, phase III efficacy AND SCREENING CANCER PREVENTION inconsistent evidence of efficacy in respiratory tract models. Lyco- trials of beta-carotene plus other micronutrients for primary pre- pene has been evaluated in numerous cell culture systems and in a vention of lung cancer have been completed, as summarized in variety of models of prostate carcinogenesis, including chemically Table 33.6. The Alpha-Tocopherol, Beta-carotene (ATBC) Trial induced, orthotopic implantation, transgenic, and xenotransplan- involved 29,133 men from Finland who were heavy cigarette tation, with mixed evidence of efficacy.42 Evidence, primarily from smokers at entry.52 In a two-by-two factorial design, participants cell culture studies, suggests that lycopene metabolites may be at were randomized to receive either supplemental α-tocopherol, least partially responsible for anticarcinogenic activity.37 beta-carotene, the combination, or placebo. Unexpectedly, par- ticipants receiving beta-carotene (alone or in combination with Clinical Trials: Retinoids, Carotenoids, and α-tocopherol) had a statistically significant 18% increase in lung Antioxidant Nutrients cancer incidence and an 8% increase in total mortality relative to participants receiving placebo. α-Tocopherol had no effect. Retinoids, carotenoids, and antioxidant nutrients have been evalu- The finding of an increased incidence of lung cancer in the ated in the setting of preneoplasia/neoplasia in many different beta-carotene–supplemented smokers was replicated in the Car- organ sites, as will be discussed. Of the many clinical trials of reti- otene and Retinol Efficacy Trial (CARET), a large randomized noids and carotenoids/other antioxidant nutrients, key trials with trial of supplemental beta-carotene plus retinol versus placebo in cancer incidence/recurrence as primary outcomes are tabulated asbestos workers and smokers.53 This trial was terminated early, (see Tables 33.5 and 33.6) and reviewed. but, at the time of termination, overall lung cancer incidence was increased by 28% in the supplemented subjects and total mortal- Clinical Efficacy in the Upper Airway ity was also increased by 17%. In contrast, the Physicians’ Health Study (PHS) of supplemental beta-carotene versus placebo in Many trials of cancer risk–reducing agents have been done in 22,071 male US physicians reported no significant effect—posi- the setting of squamous cell carcinomas of the head and neck, in tive or negative—of 12 years of supplementation of beta-carotene large part because of the substantial clinical problem of relatively on total cancer, lung cancer, or cardiovascular disease (see Table high rates of recurrences and second primary tumors in curatively 33.6).54 Two other trials involving supplemental beta-carotene treated cancer patients. Early work demonstrated that high-dose alone (the Women’s Health Study55) or with other antioxidant nu- 13cRA (50 to 100 mg/m2 per day) produced no significant differ- trients (the Medical Research Council/British Heart Foundation ences in disease recurrence (local, regional, or distant) but signifi- Heart Protection Study56) on overall cancer incidence also failed cantly lowered the rate of second primary invasive neoplasms,43 to observe efficacy. with the benefits persisting for at least 5 years.44 Substantial retinoid toxicity, however, including skin dryness and peeling, cheilitis, con- Prevention of Second Primary Invasive Neoplasms. junctivitis, and hypertriglyceridemia, was evident in a large propor- EUROSCAN was a multicenter trial employing a two-by-two fac- tion of patients. Subsequent trials thus used lower doses of retinoids torial design to test retinyl palmitate and N-acetylcysteine (also a (13cRA or a synthetic retinoid, etretinate), but failed to show effi- compound with known antioxidant activity) in preventing second cacy in reducing second primary tumor formation (Table 33.5).45,46 primary invasive neoplasms in patients with early stage cancers 356 Cancer Prevention and Screening

TABLE 33.5 Larger, Randomized Trials of Retinoids in Human Cancer Risk Reduction with Cancer Outcomesa,b

Population Drug (Dose) End Point Outcomes References United States, prior 13cRA (50–100 mg/m2/d) Second primary tumor Significant reduction in second (43, 44) HNSCC primary tumors at 32 and 55 mos; however, substantial toxicity France, prior HNSCC Etretinate (50/25 mg/d) Second primary tumor No difference (46) United States, prior 13cRA (30 mg/d) Second primary tumor No difference (45) HNSCC Europe, prior HNSCC, Vitamin A Second primary tumor No difference (57) NSCLC (300,000/150,000 IU/d) +/− N-acetylcysteine United States, Prior 13cRA (30 mg/d) Second primary tumor No difference, but second primary (58) NSCLC tumors were lower in nonsmokers on drug but higher in smokers on drug Italy, prior breast cancer 4HPR (200 mg/d) Contralateral breast Nonsignificant reduction; pre- (60) No treatment cancer menopausal women did better but opposite in postmenopausal women United States, prior BCC 13cRA (10 mg/d) Second basal cell No difference (65) carcinoma United States, prior actinic Retinol (25,000 IU/d) Skin cancer incidence Reduction in squamous cell (67) keratosis carcinomas but not basal cell carcinomas United States, prior BCC/ 13cRA (5–10 mg/d) Second skin cancer No significant difference for either (66) SCC of the skin Retinol (25,000 IU/d) agent Netherlands, renal Acitretin (30 mg/d) Skin cancer Significant reduction (64) transplant patients United States, aggressive 13cRA (1 mg/kg/d) + Second primary tumors No effect (275) SCC of the skin interferon alpha and tumor recurrences United States, prior Megadose vitamins Recurrence Significant reduction in recurrence (73) bladder TCC (40,000 IU retinol/d) versus RDA vitamins United States, prior 4HPR (200 mg/d) Recurrence No difference (72) bladder TCC a Trials of retinoids that also included beta-carotene are listed under Table 33.2 only. b Versus placebo unless otherwise indicated. HNSCC, head and neck squamous cell carcinoma; NSLC, non–small-cell lung cancer; BCC, basal cell carcinoma; TCC, transitional cell carcinoma; RDA, recommended dietary allowance. of the head and neck or lung. None of the interventions reduced randomized trial of fenretinide (versus no treatment) for 5 years to second airway primary invasive neoplasms.57 The Lung Intergroup prevent contralateral breast cancer in women aged 30 to 70 years Trial randomized patients with surgically resected lung cancer to with a history of resected early breast cancer and no prior adjuvant 13cRA versus placebo and found no significant differences be- therapy.60 The intervention produced no significant overall effect, tween the two arms in second primary tumors.58 Notably, smoking although fenretinide reduced contralateral and ipsilateral breast status modified the effect of the 13cRA intervention, which was cancer rates in premenopausal women, with an opposite (adverse) harmful in current smokers yet beneficial in former smokers. trend observed in postmenopausal women. The reduced inci- Thus, phase III trials of both carotenoids/antioxidants and reti- dence of second breast cancer in premenopausal patients persisted noids indicate that these agents overall do not reduce the risk of with longer follow-up.61 developing invasive lung cancers, nor do they prevent the develop- Retinoid X receptor (RXR)–selective retinoids are also being ment of second primary invasive neoplasms. However, the find- evaluated in preclinical and clinical studies. Ongoing work sug- ing that former smokers seemed to benefit from both 13cRA58 and gests that combination treatment may represent a promising new beta-carotene53 is intriguing. Mechanistic work suggests this inter- strategy to suppress both estrogen receptor–negative and estrogen action is real rather than chance (see the following), suggesting receptor–positive breast tumors, and the combination of retinoids that (1) risk-reduction in smokers is especially challenging,59 and with antiestrogens may be particularly effective.62 (2) trials in former smokers may merit consideration. Clinical Efficacy in the Skin Clinical Efficacy in the Breast Retinoids have been widely studied for cancer risk– reducing Moon et al.40 first showed that fenretinide was a promising cancer efficacy in skin. Early work was done in patients who have risk–reducing agent for the breast, having a high therapeutic index substantial skin cancer risk either due to xeroderma pigmentosum and synergistic interaction with tamoxifen in mammary carcino- or medication-induced immunosuppression for transplants. For genesis model studies. This laboratory work led to a large-scale example, 13cRA reduced skin cancer by 63% in patients with

tahir99 - UnitedVRG Chapter 33 Cancer Risk Reducing Agents 357

TABLE 33.6 Randomized Trials of Antioxidant Nutrients in Human Cancer Risk Reduction with Cancer Outcomesa

Population Drug (Dose) End Point Outcomes References United States, prior head/ Beta-carotene (50 mg/d) Second primary head Nonsignificant reduction in second (48) neck cancer and neck cancers head and neck cancer, increase in lung cancer Italy, prior head/neck Beta-carotene (75 mg/d) Second primary head No effect on second primary tumors, (47) cancer for 3 mos with 1 mo off and neck cancers nonsignificant decrease in death Canada, prior head/neck 30 mg beta-carotene/d + Deaths Beta-carotene discontinued, mortality (49) cancer 400 IU vitamin E/d increased at end of trial Finland, male smokers 20 mg beta-carotene/d Lung cancer Lung cancer increased with beta- (52) +/− 50 mg vitamin E/d carotene, no effect vitamin E United States, smokers Beta-carotene (30 mg/d) Lung cancer Lung cancer increased with beta- (53) and asbestos workers + retinol (25,000 IU/d) carotene + vitamin A United States, resected Selenized yeast, 200 μg/d Second primary cancer No effect (276) stage I non–small-cell lung versus placebo cancer United States, male Beta-carotene (50 mg Total cancer No effect (54) physicians every other day) United States, female Beta-carotene (50 mg/ All cancers No effect (55) health professionals every other day) United Kingdom, adults 20 mg beta-carotene/d + Total cancers No effect (56) at risk for coronary heart 600 mg vitamin E/d + 250 disease mg vitamin C/d United States, prior skin 50 mg beta-carotene/d Second skin cancer No effect (68) cancer Australia Beta-carotene (30 mg/d) Incident squamous cell No effect (69) skin cancer incident

basal cell skin cancer CANCER PREVENTION AND SCREENING United States, prior skin 200 μg selenium/d Second skin cancer No effect, with longer follow-up (70) cancer became adverse Linxian County, China, 15 mg beta-carotene/d Stomach cancer death Significant decrease in stomach (75) general population + 30 mg vitamin E/d + Esophageal cancer death cancer death; no effect on esophageal 50 μg selenium/d cancer death Linxian County, China, Multivitamin/multimineral Stomach cancer death No effect (78) esophageal dysplasia + 15 mg beta-carotene/d Esophageal cancer death United States, males 200 μg selenium/d +/− Prostate cancer No effect; with longer follow-up (84) 400 IU vitamin E/d incidence vitamin E became adverse United States, male 500 mg vitamin C/d + Prostate cancer No effect (86) physicians II 400 IU vitamin E every incidence, total cancer other day incidence a All versus placebo.

xeroderma pigmentosum; however, severe, acute mucocutaneous Supplemental beta-carotene (30 mg per day) also did not prevent toxicity with the 13cRA occurred.63 Also, the preventive effect of basal cell carcinoma or squamous cell carcinoma of the skin in an the retinoid was lost after stopping retinoid therapy. In renal trans- Australian trial.69 plant patients, acitretin (30 mg per day) reduced the numbers of Clark et al.70 randomized patients with a history of nonmel- premalignant lesions, the number of patients with skin cancer, and anoma skin cancer to 200 μg per day selenium or placebo.70 the cumulative number of skin cancers.64 Selenium did not reduce the incidence of second skin cancers; In lower risk populations, low-dose 13cRA (10 mg per day)65 a further report of this trial with longer follow-up71 indicated that and retinol or 13cRA alone did not reduce the recurrence of basal there was instead a significant increase in total nonmelanoma skin or squamous cell skin cancers,66 although retinol alone reduced cancer (hazard ratio [HR] = 1.17; 95% confidence interval [CI], squamous but not basal cell carcinomas in patients with prior 1.02 to 1.34) and squamous cell skin cancer (HR = 1.25; 95% CI, actinic keratoses.67 1.03 to 1.51). Beta-carotene (50 mg per day), in a randomized trial did not reduce the recurrence of nonmelanoma skin cancers.68 Consis- Clinical Efficacy in the Bladder tent with findings in the lung, the risk was increased by 44% in current smokers randomized to beta-carotene but not in never A trial of fenretinide (200 mg per day orally for 12 months) ver- smokers randomized to beta-carotene as compared with placebo. sus placebo was conducted for preventing tumor recurrence in 358 Cancer Prevention and Screening patients with nonmuscle-invasive bladder transitional cell car- prostate cancer. The largest trial to date of these nutrients is the Se- cinoma after transurethral resection with or without adjuvant lenium and Vitamin E Cancer Prevention Trial (SELECT), which intravesical bacillus Calmette-Guérin; recurrence rates were simi- tested selenium and vitamin E in a two-by-two factorial design for lar in both groups.72 Another trial randomized 65 patients with the primary prevention of prostate cancer. Despite preliminary biopsy-confirmed transitional cell carcinoma of the bladder to a indications of prostate cancer risk–reducing efficacy for selenium multivitamin (recommended dietary allowance [RDA] levels) (from a trial of selenium for skin cancer70) and vitamin E (from a alone or supplemented with 40,000 IU retinol, 100 mg pyridoxine, trial of vitamin E to prevent lung cancer83), there was no evidence 2,000 mg ascorbic acid, 400 U of α-tocopherol, and 90 mg zinc.73 of efficacy.84 With extended follow-up, the nonsignificant adverse The 5-year estimate of tumor recurrence was 91% in the RDA arm effect of vitamin E became significantly adverse.85 Negative/neu- versus 41% in the higher-dose nutrient arm (p = 0.0014). tral findings also were reported for vitamins E and C and prostate and total cancer in the PHS II randomized controlled trial.86 Clinical Efficacy in the Cervix The carotenoid lycopene has generated much interest with re- gard to prostate cancer risk, and several intervention trials have Randomized trials include four with beta-carotene (alone or with been conducted based on lycopene supplements. These studies other antioxidant nutrients), and five with retinoids. Only one of have been small, short term, based on intermediate end points, and these trials, involving ATRA,74 found a significant treatment ef- often lack adequate control groups. The use of a tomato sauce– fect. This trial administered a 0.372% ATRA solution by collagen based intervention is arguably a better approach to evaluate, based sponge in a cervical cap delivery system. There was a higher com- on animal data indicating that tomato powder (which includes plete response rate in the ATRA group (43%) than the placebo lycopene along with other phytochemicals), but not lycopene group (27%; p = 0.041) among the 141 patients with moderate alone, was effective at inhibiting prostate carcinogenesis.87 dysplasia; no significant differences in dysplasia regression rates between the two study arms were detected in patients with se- Mechanisms for Ineffective Retinoid and Carotenoid vere dysplasia. The investigators experienced substantial losses to Cancer Risk–Reducing Activity follow-up in this patient population. There are now a number of trials demonstrating that supplemen- Clinical Efficacy in the Esophagus and Stomach tal beta-carotene/retinoids given to current smokers can produce increases rather than reductions in cancer incidence. In tobacco Certain regions of China (Huixian and Linxian) have strikingly users, beta-carotene and other carotenoids may produce oxidative high incidence rates of esophageal and gastric cancers. Two trials carotenoid breakdown products that alter retinoid metabolism and were done in Linxian County; one was a general population trial signaling pathways, along with pro-oxidation.88 For retinoids such that tested the efficacy of four different nutrient combinations at as 13cRA, smoking may induce genetic and epigenetic changes in inhibiting the development of esophageal and gastric cancers.75 the lung that affect retinoid activity; for example, tobacco smoking Those who were given the combination of beta-carotene, vitamin can affect RAR-β expression.58 The adverse effects of supplemen- E, and selenium had a 13% reduction in total cancer deaths, a 4% tal nutrients are not limited to smokers; α-tocopherol increased reduction in esophageal cancer deaths, and a 21% reduction in rather than reduced prostate cancer in SELECT (which had rela- gastric cancer deaths (see Table 33.6). None of the other nutrient tively few smokers) and selenium increased prostate cancer among combinations reduced gastric or esophageal cancer deaths signifi- men without a baseline selenium deficiency.89 This may be a con- cantly in this trial. The treatment benefit has been shown to persist sequence of the relatively high doses used in SELECT, 77 but cer- for 10 years postintervention, with greater efficacy seen in partici- tainly calls into question the notion that reducing oxidative stress pants under age 55 years.76 This finding stands in contrast to most is a pivotal cancer risk–reduction strategy, even in nonsmokers. other antioxidant nutrient supplement intervention trials, suggest- It has become clear that reactive oxygen species (ROS), such as ing that the applicability of these results for populations with ad- hydrogen peroxide, can act as important physiologic regulators of equate nutritional status and for other tumor sites may be limited.77 intracellular signaling pathways.90,91 Data in mouse models have The other Linxian trial evaluated a multivitamin/multimineral shown that vitamin E accelerates lung tumor growth by disrupting preparation plus beta-carotene (15 mg per day) in residents with the ROS–p53 axis, potentially by removing oxidative damage to esophageal dysplasia.78 There was no clear evidence of efficacy, DNA, which can serve as a potent stimulus for p53 activation.92 although confidence intervals were wide. Although some of the large cancer risk–reducing trials may have failed in their primary objective, they may indirectly contribute to a Clinical Efficacy in the Colon/Rectum clearer understanding of cancer biology, leading to the recognition that the role of oxidative stress and ROS in human disease is much Of the randomized trials aimed at the prevention of recurrent more nuanced than originally hypothesized.93 colorectal adenomas with micronutrients that have been com- As for the retinoids, these agents are generally too toxic to be pleted, some used beta-carotene alone79 or with other nonmi- used as single agents for risk-reducing efficacy; however, a major cronutrient interventions.80 Others evaluated beta-carotene with area of ongoing research is examining retinoids (low doses) given and without supplemental vitamins C and E.81 None of the trials in combination with other agents, especially those that regulate observed benefit with supplementation. A subsequent report from the epigenome, such histone deacetylase (HDAC) inhibitors.33 one trial noted that alcohol intake and cigarette smoking modified the efficacy of beta-carotene.82 Among nonsmokers and nondrink- Folic Acid and Other B Vitamins ers, beta-carotene was associated with a significant decrease in the risk of one or more recurrent adenomas (relative risk [RR] = 0.56). Overview and Mechanisms. Folate is a water-soluble B Among persons who smoked and also drank more than one alco- vitamin found in foods, whereas folic acid is the synthetic form holic drink per day, beta-carotene significantly increased the risk found in supplements and fortified foods. Adequate folate is critical of recurrent adenoma (RR = 2.07). for DNA methylation, repair, and synthesis.94,95 The methylation status of genes can play a key role in gene silencing and gene ex- Clinical Efficacy in the Prostate pression, lending plausibility to the idea that folate could be a key nutrient in regulating cell growth and proliferation. Because oxidative stress may play a role in the etiology of pros- tate cancer, several antioxidant nutrients, including vitamin E, Epidemiology. Epidemiologic studies have linked low folate selenium, and lycopene, have been of interest for preventing intake with higher risk of several cancers, most notably colorectal

tahir99 - UnitedVRG Chapter 33 Cancer Risk Reducing Agents 359 cancer.96 Long-term use of multivitamin supplements, which are association between vitamin D status and cancer risk, as systemati- a major source of folate and other B vitamins, has been associated cally reviewed by the Agency for Healthcare Research and Quality with a reduction in the risk of colon cancer in some studies, includ- (AHRQ).114 The evidence is inconsistent for most cancer sites, ing recent (postfortification) findings.97–99 Supporting an antican- with the exception of studies showing that individuals with lower cer role of folate is that genotypes for methylene tetrahydrofolate blood vitamin D levels have a higher risk of colorectal cancer or reductase, an enzyme known to be involved in folate metabolism, adenoma. Although some observational studies have reported that predict the risk of colon cancer dependent on folate intake or sta- higher serum vitamin D is associated with lower breast cancer risk, 100 114,115 tus. Vitamin B6 has been less studied in relation to cancer than the association is inconsistent. Also, there are some studies folate, but some epidemiologic studies suggest that vitamin B6 may suggesting high serum vitamin D is associated with increases in be important for colorectal cancer.101,102 A higher risk of cancer certain cancers, particularly pancreatic cancer.116 related to deficiencies of these vitamins has been suggested for alcohol drinkers.103 Clinical Trials: Calcium and Vitamin D

Clinical Trials: Folic Acid and B Vitamins. Risk-reducing ef- Clinical Efficacy in the Colon. Baron et al.117 randomized subjects ficacy for supplemental folic acid has been primarily evaluated in with a recent history of colorectal adenomas to either calcium car- the setting of prevention of recurrent colorectal adenomas (i.e., in bonate (1,200 mg per day of elemental calcium) or placebo. Results patients with prior adenomas). Of six randomized trials of folic acid, showed significant benefit for the calcium arm (adjusted RR = 0.81; two small trials reported suggestions of benefit of folic acid supple- 95% CI, 0.67 to 0.99; p = 0.04). In a smaller, similar study of cal- mentation.104,105 However, benefits were not observed in two much cium gluconolactate and carbonate (2 g elemental calcium daily), larger trials, the Aspirin/Folate Polyp Prevention Study (AFPPS) the adjusted odds ratio (OR) for adenoma recurrence was 0.66 (95% (dose: 1 mg of folic acid daily)106 and the United Kingdom Colorec- CI, 0.38 to 1.17; p = 0.16) for calcium treatment,118 and while not tal Adenoma Prevention (ukCAP) trial (dose: 500 μg of folic acid statistically significant, it was similar to the data of Baron et al.117 daily).107 AFPPS found indications of an increased risk for advanced In the largest trial of calcium and vitamin D with primary can- lesions and multiple adenomas with prolonged treatment and fol- cer end points (e.g., colon, breast), the US Women’s Health Initia- low-up. A third large trial, the Nurses Health Study/Health Profes- tive (WHI) evaluated the combination of 400 IU of vitamin D per sionals Follow-up Study (NHS/HPFS) folic acid polyp prevention day plus 1,000 mg of calcium per day in 36,282 postmenopausal trial, showed no overall risk reduction.108 The most recent trial, women. For colon cancer, there was no benefit observed,119 al- done in a Chinese population >50 years of age,109 reported that 1 though the mean baseline intake of calcium was already very high mg folic acid per day reduced sporadic colorectal adenomas when (more than 1,151 mg per day). With regard to vitamin D as a single compared to no intervention (not a placebo-controlled study). One agent, there was also no suggestion of benefit for colon cancer in- possible explanation for the discrepancy of the Chinese trial versus cidence in a 5-year British trial of vitamin D (100,000 IU every North American and European trials is the baseline plasma folate 4 months) that reported colon cancer incidence,120 although this status. In the Chinese trial, the mean baseline folate concentration was not a primary end point. of 5 ng/mL109 was half of the reported 10 ng/mL in a United States

106 AND SCREENING CANCER PREVENTION trial, where folate fortification of the food supply occurs. Clinical Efficacy in the Breast. Vitamin D has received consider- able attention for a possible role in the prevention of breast can- Calcium and Vitamin D cer,121 although no trials have yet investigated vitamin D as a single agent for breast cancer risk reduction. The large WHI trial gave a Overview and Mechanisms. There are two major forms of vi- combination of calcium and vitamin D, as noted previously, and tamin D: ergocalciferol (D2) and cholecalciferol (D3). Vitamin D2 there was no significant effect of this combination on breast cancer is absorbed through dietary sources such as fortified milk products, risk (HR, 0.96; 95% CI, 0.85 to 1.09).122 Lappe et al.123 conducted and D3 is synthesized via ultraviolet (UV) B light isomerization of 110 a trial that examined the relation between calcium plus vitamin 7-dehydrocholestrol in the epidermis. Vitamin D3 is converted D (1,100 IU per day) supplementation (versus calcium alone or to calcitriol (1, 25-[OH]2 D3) in a two-step process requiring both placebo) in 1,179 healthy postmenopausal women in Nebraska.123 hepatic and renal hydroxylation. Calcitriol binds to the vitamin D Although fracture was the primary outcome of the trial, total can- receptor, which translocates to the nucleus and binds to multiple cer incidence was reportedly lower in the calcium plus vitamin D gene promoter sites. Through this mechanism, vitamin D regulates group, although the number of end points was very small (n = 50 cytoplasmic signaling pathways that impact cellular differentiation total cancers observed during the follow-up). An ongoing random- and growth through proteins such as Ras and mitogen-activated ized trial of vitamin D and omega-3 fatty acids (the VITamin D protein kinase (MAPK), protein lipase A, prostaglandins, cyclic and OmegA-3 TriaL [VITAL]) among 20,000 participants is ex- adenosine monophosphate (AMP), protein kinase A, and phos- 110 pected to provide more definitive data on a possible role of vitamin phatidyl inositol 3 kinase. 1,25(OH)2D3 regulates cellular pro- D in the prevention of breast and other cancers.124 liferation and apoptosis. For example, 1,25(OH)2D3 can induce cleavage of caspase 3, poly (ADP-ribose) polymerase (PARP), and MAPK, leading to apoptosis. 1,25(OH)2D3 inhibits the expression Summary and Conclusion: Micronutrients and phosphorylation of Akt, a key regulator of cellular prolifera- tion. The differentiation properties of 1,25(OH)2D3 are mediated Certain agents, including the retinoids, beta-carotene, folic acid, through transcriptional activation of the CDK inhibitor p21. The calcium plus vitamin D, vitamin E, and selenium, have received effects of vitamin D on multiple signal transduction pathways op- substantial attention for a possible role in reducing the risk of cancer erational in cancer cells are reviewed by Deeb et al.111 in humans. As reviewed herein, some of the trials have observed sta- tistically significant reductions in the risk of the primary end point Epidemiology. Observational epidemiologic studies have (e.g., retinoids in skin carcinogenesis models, calcium in colorectal shown a relatively consistent inverse association between low adenomas, antioxidant nutrients in Linxian, China, for gastric can- calcium intake, including that from supplements, and increased cer prevention), whereas others have observed statistically signifi- colorectal and colon cancer risk.112,113 Vitamin D exposure is typi- cant increases in the risk of the primary end points (beta-carotene cally assessed by measuring 25(OH)vitamin D in plasma because and retinoid lung cancer prevention trials in smokers, vitamin E exposure is derived not only from diet and supplements, but also and prostate cancer, selenium and nonmelanoma skin cancer). from cutaneous synthesis following dermal exposure to UV radia- Considering the completed trials, there is clear evidence against the tion. A large number of observational studies have evaluated the general use of nutrient supplements for cancer prevention, which is 360 Cancer Prevention and Screening the conclusion also reached by the World Cancer Research Fund/ expression and reduced colony formation in vitro, while reducing American Institute for Cancer Research.38 Note that there is no evi- PPARδ, leading to reduced resistance to apoptosis.130 Selective dence that food sources of these nutrients increase risk. cyclooxygenase 2 (COX-2) inhibitors inhibit Akt signaling and Having noted that, there are other key themes emerging from this induce apoptosis of human colorectal and prostate cancer cells growing body of research. One such theme is that nutrient supple- in vitro in a COX-2–independent manner via the inhibition of mentation may be of benefit to some but not all. One such popula- phosphoinositide-dependent kinase-1 (PDK-1). NSAIDs inhibit tion that may benefit includes persons who are low in the nutrient nuclear factor kappa B (NF-κB) at pharmacologic concentrations of interest at baseline.77 This was initially suggested in the Linxian and key cellular proliferation signaling intermediates such as acti- Country trial (done in a micronutrient-deficient population), with vator protein 1 (AP-1) and other intermediates of the MAPK path- growing support from subgroup analyses of several completed trials.77 way.130 The impact of NSAIDs on carcinogenic events driven by However, the hypothesis that nutrient supplementation can reduce these upstream pathways in humans as opposed to preliminary in cancer risk in subgroups selected based on inadequate nutritional vitro or in vivo models remains unclear. status has, to date, not been formally evaluated in intervention trials. Another consistent theme is that lifestyle factors (e.g., smoking) and genetics (polymorphisms) may determine who is most likely Epidemiology to benefit from supplementation. Trial data will likely be increas- ingly mined to identify genetic profiles associated with both bet- Pooled analyses of 34 controlled trials of aspirin 75 mg to 100 mg ter outcomes (risk prediction) and response to intervention.125,126 daily (69,224 participants), conducted primarily for cardiovascular Ultimately, a more personalized approach to cancer risk reduction disease reduction, observed reduced cancer deaths (OR, 0.63; 95% may emerge, consistent with the movement toward a more person- CI, 0.49 to 0.82). Most of the benefit occurred after 5 years follow- alized approach for cancer treatment. up.131 In a pooled analysis of 150 case control and 45 cohort studies, Finally, nearly all of these trials initiate intervention with older in addition to a reduced risk of death from colorectal cancer (OR, adults (who are more likely to develop cancer end points during 0.58; 95% CI, 0.44 to 0.78), chronic and frequent (once daily or the follow-up); but, animal models suggest that the timing of ex- more) use of aspirin also reduced the risk of death from esophageal posure may likely be quite relevant. For example, folic acid may (OR, 0.58; 95% CI, 0.44 to 0.76), gastric (OR, 0.61; 95% CI, 0.40 protect against initiation, but may also promote the proliferation to 0.93), and breast (OR, 0.81; 95% CI, 0.72 to 0.93) cancers.132 An of existing neoplasms.127 Thus, the dose, form (food versus supple- analysis of 662,624 men and women enrolled in the American Can- ment), timing, and nutritional and lifestyle characteristics may all cer Society’s Cancer Prevention Study II found that aspirin taken be relevant in affecting the efficacy of risk-reducing interventions at least 16 times per month over a 6-year period conferred a 40% involving nutrients and related substances. Further research, draw- reduced risk of colorectal cancer mortality.133 Both the 46,363 male ing upon newer tools now available through the field of nutritional patients of the Health Professional Study134 and the 82,911 patients genomics, will be needed to gain greater clarity on the heteroge- of the Nurse’s Health Study135 suggest that prolonged use (>10 years) neous biologic effects observed in nutrient-based risk reduction. of 325 mg of aspirin twice weekly or more reduces colorectal cancer risk (RR, 0.77; 95% CI, 0.67 to 0.88, from the Nurse’s Health Study). Daily NSAID intake is associated with a 40% reduction (OR, 0.56; 136 ANTI-INFLAMMATORY DRUGS 95% CI, 0.43 to 0.73) in the risk of esophageal adenocarcinoma.

Mechanism Evidence in Preclinical In Vivo Carcinogenesis Models Nonsteroidal anti-inflammatory drugs (NSAIDs) represent a class of drugs that reduce cellular inflammation through multiple NSAIDs, including aspirin, indomethacin, piroxicam, sulindac, mechanisms, the most prominent of them being the modulation of ibuprofen, and ketoprofen, suppress colonic tumorigenesis in- eicosanoid metabolism.128 Eicosanoids are metabolites of dietary duced chemically (1,2-dimethylhydrazine or its metabolites) or fatty acids, primarily linoleic acid. Linoleic acid is metabolized transgenically (Min+).137,138 The selective COX-2 inhibitors were to arachidonic acid, which is stored in the lipid membrane and, the most efficacious colon tumorigenesis inhibitors in both chemi- once mobilized from the membrane, further metabolized by pros- cal and transgenic rodent models.139,140 In preclinical models, taglandin-H synthases (PGHS) 1 and 2 to PGD2, PGE2, PGF2α, NSAIDs affect the onset and progression of cancers in the stomach, PGI2, or thromboxane A2 (TxA2) by specific synthases. Leukotriene skin, breast, lung, prostate, and urinary bladder, although the evi- pathways involve the conversion of arachidonic acid to leukotriene dence is more limited than for colon cancers.141 A4 by 5-lipoxygenase and subsequent hydrolysis of leukotriene A4 to other downstream leukotrienes. Newly formed prostaglandins function primarily through binding to prostaglandin receptors (EP Clinical Trials receptors), releasing coupled G-proteins to elicit responses in the same or neighboring cells.129 Key clinical trials of NSAIDs for the prevention of colorectal can- Prostaglandins (PG) play crucial roles in controlling cellular cer are summarized in Table 33.7. Sulindac reduced the size and proliferation, apoptosis, cellular invasiveness, and angiogenesis number of preexisting adenomas in patients with familial adeno- 129 and in modulating immunosuppression. Because PGE2 is the matous polyposis but did not suppress the development of new most abundant PG in tumors, reducing local concentrations of adenomas,142 whereas the selective COX-2 inhibitor, celecoxib, 129 PGE2 may be a pivotal cancer preventive strategy. suppressed the development of new adenomatous polyps in patients PGHS-independent mechanisms of NSAID action may, at with familial adenomatous polyposis.143 Although these results are least in part, explain NSAID preventive efficacy.130 A diverse promising, reports of invasive neoplasms developing in familial ad- group of NSAIDs inhibit apoptosis via multiple mechanisms. enomatous polyposis patients being treated with sulindac144 raise Among the more prominent of these mechanisms is the inhibition the question of whether NSAIDs preferentially alter the formation of cyclic guanosine monophosphate (cGMP) phosphodiesterase or regression of those adenomas less likely to progress to invasive activity, attenuation of beta-catenin mRNA through suppress- adenocarcinomas, as compared to those more likely to progress. ing transcription of the CTNNB1 gene, and activation of c-Jun Randomized, double-blinded placebo controlled trials of N-terminal kinase 1.130 NSAIDs activate peroxisome prolifera- NSAIDs as cancer risk–reducing agents for colorectal adenocar- tor–activated receptor (PPAR)γ, leading to increased E-cadherin cinoma (see Table 33.7) have confirmed that aspirin suppresses

tahir99 - UnitedVRG Chapter 33 Cancer Risk Reducing Agents 361

TABLE 33.7 Summary of Clinical Trials of Nonsteroidal Anti-Inflammatory Drugs as Colorectal Cancer Risk–Reducing Agents

Population Drug (Dose), Duration Phase Endpoint Outcome References Gene Associated Familial adenomatous Sulindac (300–400 mg/d, IIb Polyp regression Colorectal and duodenal polyps (277, 278) polyposis (FAP) divided doses) regressed in ∼50% (279) Hereditary nonpolyposis Aspirin 600 mg/d, III Cancer ≥2 yr, hazard ratio (HR) colon cancer (280, 281) colon cancer (Lynch resistant starch 0.41; 95% CI, 0.19–0.86; all cancers syndrome) Incidence rate ratio 0.37; 95% CI, 0.18–0.78; no effect of starch Sporadic Risk Previous adenomatous Aspirin (40, 81, 325, 650 I, IIa Dose-biomarker Aspirin dose of 81 mg daily sufficient (282–284) polyps, healthy subjects mg once per day) to suppress colorectal mucosal

prostaglandin E2 Previous adenomatous Sulindac (300 mg), 4 mos IIb Polyp regression Sulindac did not significantly decrease (146) polyps the number or size of polyps

Previous adenomatous Piroxicam (7.5 mg), 2 yr IIb Polyp recurrence Colorectal mucosal PGE2 reduced in (145) polyps piroxicam treated arm, unacceptable toxicity Prior colorectal cancer Aspirin (325 mg once per IIb Polyp recurrence Aspirin use associated with delayed (285) day), 3 yr development of adenomatous polyps Previous adenomatous Aspirin (81 mg once per IIb Polyp recurrence Low-dose aspirin reduced the (286) polyps day or 325 mg once per recurrence of adenomatous polyps day) and/or folate, 3 yr Previous adenomatous Celecoxib and rofecoxib IIb Polyp recurrence Celecoxib and rofecoxib reduced the (148–150) polyps recurrence of adenomatous polyps, unacceptable toxicity CANCER PREVENTION AND SCREENING

adenoma recurrence in patients previously treated for adenomas Minimal prospective cancer risk reduction data are avail- or for cancer. Neither sulindac nor piroxicam alone suppressed able at other epithelial organ sites. Ketorolac, given as a 1% rinse adenoma formation in high-risk, sporadic populations at tolerable solution, did not reduce the size or histology of leukoplakia le- doses.145,146 Sulindac, in combination with difluoromethylorni- sions.158 Celecoxib reduces the Ki67 labeling index and increases thine, has potent colorectal anticarcinogenesis effects.147 Selective the expression of nuclear survivin without significantly changing the COX-2 inhibitors (celecoxib, rofecoxib) reduce the recurrence cytoplasmic survivin in bronchial biopsies of smokers.159 Cancer pre- of adenomas by one-third in all patients previously treated for vention trials of aspirin as interventions for delaying progression from adenomas and by one-half in patients with previously resected large intraepithelial neoplasias in other epithelial sites remain ongoing for (≥1 cm) adenomas,148–150 but they are too toxic as cancer risk–reduc- the lower esophagus.136 No prospective, randomized trials or data ing agents due to their cardiovascular toxicity.151,152 Although most are available for breast, prostate, or gynecologic cancer prevention. NSAIDs (piroxicam, indomethacin) have sufficient gastrointestinal (GI) toxicity to reduce their acceptability as cancer risk– reducing agents,153,154 the long-term administration of low-dose aspirin in vas- EPIGENETIC TARGETING AGENTS cular prevention trials demonstrates acceptable GI toxicity.155 (SELECTIVE ESTROGEN RECEPTOR Up to 40% of individuals screened for colorectal neoplasms will MODULATORS, 5훂-STEROID REDUCTASE have an adenomatous polyp detected and removed, yet only 10% of these lesions will progress to invasive neoplasms. To date, pro- INHIBITORS, POLYAMINE INHIBITORS) spective NSAID trials of only 2 to 3 years cannot substitute for can- cer incidence or mortality end points. Given the 10-year latency Posttranslational pathway targets remain a fertile source of chemo- between adenoma formation and a cancer event, prospective tri- preventive strategies. Phase III data support cancer risk–reduction als sufficiently powered to detect colorectal cancer incidence end agent efficacy of selective estrogen receptor modulators (SERM) α points are unlikely in the future.156 Alternatively, a follow-up of pa- and 5 -steroid reductase inhibitors for breast and prostate cancer tients randomized on trials of aspirin in the prevention of vascular prevention, respectively. Inhibitors of the polyamine pathway may events in the 1980s and 1990s offers secondary analysis opportuni- be useful preventives for colorectal cancer. ties. In a pooled analysis of three prospective vascular end point cohort studies, 20-year low-dose aspirin treatments reduced can- Selective Estrogen Receptor Modulators cer deaths from all solid tumors (OR, 0.69; CI, 0.54 to 0.88) and from lung and esophageal adenocarcinomas (OR, 0.66; CI, 0.56 Mechanism to 0.77).155 Despite this, the U.S. Preventive Services Task Force (USPSTF) does not recommend the use of aspirin or NSAIDs as SERMs function as estrogen receptor (ER) agonists and antagonists cancer risk–reducing agents for normal risk populations, preferring depending on the SERM structure and target tissue. Predominant adherence to colorectal cancer screening recommendations (fecal ERα receptors occur in the human uterus, cortical bone, and the occult blood testing and endoscopy).153,154,157 liver; whereas predominant ERβ receptors occur in blood vessels, 362 Cancer Prevention and Screening cancellous bone, the whole brain, and immune cells.160,161 During in bone but reduced estrogen agonist activity in the uterus. Ral- carcinogenesis, the amount of ERα increases while the amount of oxifene was studied for the treatment and prevention of osteopo- ERβ decreases in breast tissues.162 Ideally, a desirable SERM for rosis in a large, pivotal trial (the Multiple Outcomes of Raloxifene cancer prevention will function as an antiestrogen in the breast Evaluation [MORE]) and was found to reduce the rate of vertebral 169 and uterus, but a partial estrogen agonist in skeletal, cardiovascu- fracture as compared to placebo in postmenopausal women. lar, central nervous system (CNS), GI tract, and vaginal tissues. In addition, an ideal SERM will not have procoagulant effects and Lasofoxifene and Arzoxifene. Lasofoxifene and arzoxifene will not cause perimenopausal symptoms such as hot flashes.162 are third-generation SERMs developed as more potent blockers of bone resorption with the goal of reducing the risk of fractures, Tamoxifen. Tamoxifen is a triphenylethylene compound devel- breast cancer, and heart disease while minimizing the SERM- oped for the treatment of ER-positive breast cancer in the 1960s and induced risk of endometrial hyperplasia in postmenopausal 1970s.163,164 Tamoxifen inhibits the initiation and promotion phases women. Both agents proved potent in vitro and in preliminary 170–173 of breast carcinogenesis in the dimethylbenzanthracene chemical clinical trials for bone fracture prevention. carcinogenesis model.164,165 When tamoxifen binds to ERβ, which then binds to an AP-1 type gene promoter, it functions as an estrogen Selective Estrogen Receptor Modulators as Risk- agonist. When bound to ERα, which binds to an estrogen response Reducing Agents for Breast Cancer Prevention element (ERE) target gene promoter, tamoxifen functions as an estrogen antagonist.162,166 Tamoxifen has estrogen antagonist effects Efficacy. Table 33.8 summarizes the phase III data for SERM- in the human breast; partial estrogen agonist effects in bone, the based breast cancer–risk reduction. In a systematic review of cardiovascular system, and CNS; and predominant estrogen agonist MEDLINE and Cochrane databases through December, 2012, effects in the uterus, liver, and vagina. The estrogen agonist effects the USPSTF identified seven trials of tamoxifen or raloxifene that in the liver and uterus result in tamoxifen’s toxicities of thromboem- showed a reduced incidence of invasive breast cancer by 7 to 9 cases bolism and endometrial cancer, respectively. The clinical finding in 1,000 women over 5 years compared to placebo.174 Tamoxifen is that tamoxifen reduces the incidence of contralateral second pri- more effective than raloxifene; it reduces breast cancer incidence mary breast cancers during adjuvant treatment regimens catalyzed more than raloxifene by 5 cases in 1,000 women. Both drugs reduce the push for its development as a cancer risk–reduction agent.167,168 the incidence of ER-positive breast cancer, but neither reduces the risk of ER-negative breast cancer. Neither drug reduced breast Raloxifene. The benzothiophene structure of raloxifene confers cancer–specific or all cause mortality rates. Based on benefit–risk a different tissue-specific ER-binding profile than the triphenyl- models, women with estimated 5-year risks of breast cancer of 3% ethylene tamoxifen. Raloxifene has greater estrogen agonist activity or greater are likely to benefit from treatment.175 Using similar

TABLE 33.8 Phase III, Randomized, Controlled Clinical Trials of SERMs for the Prevention of Breast Cancer

Drug and Daily Treatment Overall Outcome Study Dose N = Duration (Years) Entry Criteria HR (95% CI) References NSABP P-1 Tamoxifen 20 mg 13,388 5 Gail model: 5 yr predicted risk 0.52 (0.42–0.64) (28,287) Placebo of ≥1.66% IBIS-I Tamoxifen 20 mg 7, 1 3 9 5 >Twofold relative risk 0.72 (0.58–0.90) (288) Placebo Marsden Tamoxifen 20 mg 2,471 8 Family history 0.87 (0.63–1.21) (289) Placebo Italian Tamoxifen 20 mg 5,408 5 Normal risk, hysterectomy 0.67 (0.59–0.76) (290) Placebo NSABP P-2 Raloxifene 60 mg 19,747 5 Gail model: 5 yr predicted risk RR Raloxifene (29) (STAR) Tamoxifen 20 mg of ≥1.66% versus tamoxifen 1.02 (0.81–1.28) MORE/CORE Raloxifene 60 mg 7,705 5 Normal risk, postmenopausal 0.42 (0.29–0.60) (291,292) Placebo with osteoporosis Raloxifene 120 mg 6,511 Placebo RUTH Raloxifene 60 mg 10,101 5 Normal risk, postmenopausal 0.67(0.47–0.96) (179) Placebo with risk of coronary heart disease PEARL Lasofoxifene 0.5 mg 8,856 5 Normal risk, postmenopausal, 0.25 mg: 0.82 (170) Lasofoxifene 0.25 mg with osteoporosis (0.45–1.49) Placebo 0.5 mg: 0.21 (0.05–0.55) GENERATIONS Arzoxifene 20 mg 9,354 4 Normal risk, postmenopausal, 0.42 (0.25–0.68) (172) Placebo with osteoporosis

Table and data adapted from Cuzick J, Sestak I, Bonanni B, et al. Selective oestrogen receptor modulators in prevention of breast cancer: an updated meta- analysis of individual participant data. Lancet 2013;381:1827–1834.

tahir99 - UnitedVRG Chapter 33 Cancer Risk Reducing Agents 363 analysis methods as the USPSTF, the American Society of Clinical International Breast Cancer Intervention Study II (IBIS-II), com- Oncology recommends the use of tamoxifen (20 mg per day orally paring anastrozole with placebo, are similar to those reported for for 5 years) or raloxifene (60 mg per day orally for 5 years) “in pre- exemestane.185 Compared to exemestane, anastrozole decreases menopausal women who are age ≥35 years with a 5-year projected the incidence of ductal carcinoma in situ (DCIS), whereas ex- absolute breast cancer risk ≥1.66% according to the NCI Breast emestane does not. Neither aromatase inhibitor increased survival Cancer Risk Assessment Tool (or equivalent measures), or with compared to placebo controls. The American Society of Clinical lobular carcinoma in situ.”176 Tamoxifen reduces the risk of in situ Oncology recommends exemestane for breast cancer prevention (preinvasive) breast neoplasms (lobular carcinoma in situ, ductal in addition to tamoxifen and raloxifene. carcinoma in situ) by 50%.177,178 The reduction during treatment persists for at least 5 years after treatment.177,178 Raloxifene does not Use Counseling. Despite the widespread evidence of breast reduce the risk of in situ breast neoplasms. cancer preventive efficacy for tamoxifen and raloxifene, only 3% Data from two trials designed to evaluate the safety and effi- to 20% of eligible high-risk women agree to take tamoxifen for cacy of lasofoxifene (PEARL)170,171 and arzoxifene (GENERA- primary prevention.186 The low willingness of eligible women to TIONS)172,173 as bone fracture preventives have been analyzed for take tamoxifen for 5 years demonstrates the issue of risk benefit breast cancer–risk reduction. Their effect at reducing breast can- for cancer risk–reducing agents. Women with high short-term cer incidence was captured in secondary analyses (see Table 33.8). risk (5 year Gail risk of >3%)—for example, those with ER- Neither lasofoxifene nor arzoxifene have been evaluated in phase positive atypical hyperplasia, lobular carcinoma in situ, and the III randomized controlled breast cancer prevention trials. Arzoxi- majority of non–high-grade ductal carcinoma in situ lesions— fene development has been discontinued in the United States. have an acceptable risk to benefit ratio and are the most likely to benefit from a 5-year cancer risk–reducing agent intervention Toxicity Profiles. Tamoxifen causes a twofold increase in the with a SERM.175,176 The toxicity profile of aromatase inhibitors risk of endometrial adenocarcinoma (RR, 2.13; 95% CI, 1.36 to differs from SERMs. Although aromatase inhibitors may have 3.32) and is related to more benign gynecologic conditions, uter- a more favorable risk to benefit profile than SERMs, long-term ine bleeding, and surgical procedures than the placebo controls, outcomes and toxicity experience for aromatase inhibitor risk- whereas raloxifene did not increase the risk for endometrial can- reducing agent intervention are not available to date. In the cer or uterine bleeding.174 Tamoxifen causes a twofold increase in National Surgical Adjuvant Breast and Bowel Project, tamoxi- thromboembolic events (RR, 1.93; 95% CI, 1.41 to 2.64), whereas fen-treated women with a BRCA2 mutation but not a BRCA1 raloxifene causes a 60% increase in risk of venous thromboembo- mutation had reduced cancer incidence,187 but subsequent data lism (RR, 1.60; 95% CI, 1.15 to 2.23).174 Raloxifene does not differ from another group have found reduced cancer risk in women from tamoxifen in risk of fractures, other cancers, or cardiovascular with both BRCA mutations.188 Data remain insufficient to rec- events.177 Raloxifene’s lower risk of endometrial adenocarcinomas ommend the use of SERMs for risk reduction in women with compared to tamoxifen needs to be weighed against the increased BRCA mutations. risk of stroke seen in in the MORE/CORE trials (see Table 33.8).179 Raloxifene’s effectiveness in the community may also be compro- 훂 5 -Steroid Reductase Inhibitors CANCER PREVENTION AND SCREENING mised by its poor bioavailability (2%) due to rapid phase II enzyme 180 metabolism in the gut and liver, whereas tamoxifen is more Mechanism bioavailable and has active metabolites that permit a prolonged drug effect. Missed raloxifene doses may potentially compromise Prostate cancers require androgens to proliferate and evade apop- efficacy and prevention outcomes in widespread, community use. tosis. The primary nuclear androgen responsible for the mainte- nance of epithelial function is dihydrotestosterone. The testes and Aromatase Inhibitors. In adjuvant clinical trials for breast adrenal gland synthesize dihydrotestosterone by the conversion of cancer, aromatase inhibitors (anastrozole, exemestane, letrozole) testosterone by 5α-steroid reductase types 1 and 2 isozymes. Dihy- given after 5 years of tamoxifen enhance the reduction of breast drotestosterone binds to intracellular androgen receptors to form a cancer recurrence in the contralateral breast compared to complex that binds to DNA hormone response elements control- tamoxifen alone.181 In a phase I cancer risk–reducing agent trial, ling cellular proliferation and apoptosis. Finasteride, a selective, letrozole reduced the Ki-67 proliferation index of breast epithelial competitive inhibitor of type 2 5α-steroid reductase,189 inhibits cells aspirated from high-risk women.182 Exemestane reduced the proliferation in the transformed prostate cell. In the 3,2′-dimethyl- overall risk of ER-positive invasive breast cancer (Table 33.9). It 4-aminobiphenyl (DMAB), methylnitrosourea (MNU), and testos- did not reduce the risk of noninvasive breast neoplasms or ER-neg- terone chemical carcinogenesis models in rats, finasteride reduces ative breast cancer.183 Exemestane has no increased risk of venous prostate tumor incidence by close to six-fold. Finasteride appears thromboembolism, endometrial cancer, fracture, or cataract,183 to be more effective in the promotion phase of prostate carcino- but losses in bone mineral density and cortical thickness of the genesis.190 Dutasteride inhibits both 5α-steroid reductase inhibi- distal tibia and radius occurred after 2 years of treatment despite tor190 types 1 and 2 isoforms and has similar anticarcinogenesis calcium and vitamin D supplementation.184 The results of the activity in preclinical models to finasteride.

TABLE 33.9 Phase III, Randomized, Controlled Clinical Trials of Aromatase Inhibitors for the Prevention of Breast Cancer

Treatment Overall Outcome Study Drug and Daily Dose N = Duration (Years) Entry Criteria HR (95% CI) References MAP.3 Exemestane 25 mg 4,560 5 Gail model 5 yr predicted risk of 0.35 (0.18–0.70) (183) Placebo ≥2.3% IBIS-II Anastrozole 1 mg 3,851 5 RR twofold higher than general 0.47 (0.32–0.68) (185) Placebo population or Tyrer-Cuzick 10-yr risk >5% 364 Cancer Prevention and Screening

TABLE 33.10

Phase III, Randomized, Controlled Clinical Trials of 5훂-Steroid Reductase Inhibitors for the Prevention of Prostate Cancer

Treatment Overall Outcome Study Drug and Daily Dose N = Duration (Years) Entry Criteria HR (95% CI) References PCPT Finasteride 5 mg 18,880 7 Age ≥55 y, PSA ≤3 ng/mL 0.70 (0.65–0.76) (30, 193) Placebo REDUCE Dutaseride 0.5 mg 6,729 4 Age 50–75 y, PSA 2.5–10.0 ng/mL, RR = 0.77 (192) Placebo core biopsies within 6 mos (0.70–0.85)

PSA, prostate specific antigen.

Cancer Risk–Reducing Agent Activity or additive activity with retinoids, butylated hydroxyanisole, tamox- ifen, piroxicam, and fish oil has been demonstrated with low con- Randomized, placebo-controlled cancer incidence end point risk- centrations of DFMO.194 reducing agent clinical trials demonstrated that finasteride and dutasteride reduced the incidence of prostate cancer by approxi- Clinical Trials mately 22% (Table 33.10).30,191,192 Patients who are treated with either drug yet progress to transformed neoplasms develop more In phase I prevention trials, DFMO at a dose of 0.5 mg/m2 per tumors of a high Gleason grade (7 to 10) compared to the placebo day reduced tissue polyamines in the colon and skin196,197 and arm (22%). After 18 years of follow-up, no significant differences causes regression of cervical intraepithelial neoplasia when used in overall survival or survival after prostate cancer diagnosis were topically,198 but does not reduce tissue polyamines or other bio- found in the finasteride-treated group compared to the placebo- markers of cellular proliferation in the human breast.199 As a single treated group.193 Sexual function side effects (e.g., erectile dys- agent, DFMO has anticarcinogenic activity for nonmelanoma skin function, loss of libido, gynecomastia) were more common in the cancers, primarily basal cell carcinoma. In combination with an finasteride- or dutasteride-treated groups.30,192 NSAID (sulindac), DFMO reduced adenoma recurrences, suggest- The 5α-steroid reductase inhibitors, finasteride and dutaste- ing a synergistic reduction of colorectal cancer risk (Table 33.11). ride, prevent or delay carcinogenesis progression in the prostate, Preliminary data suggest some cancer risk–reducing agent activity yet progression of high-grade lesions is unaffected. Use of finaste- for the lower esophagus and the prostate (see Table 33.11).200 ride for a period of 7 years reduced the incidence of prostate cancer but did not significantly affect mortality.193 Increasing the diagno- sis of low-grade prostate cancer through prostate-specific antigen Statins (PSA) testing or intervention with a drug with a minimal toxicity profile without reducing mortality is of no benefit and “all forms of Mechanism 193 therapy cause considerable burden to the patient and to society.” Statins are hydroxyl-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors that inhibit the conversion of HMG-CoA to mevalonate, a cholesterol precursor. The statins are a class of SIGNAL TRANSDUCTION MODIFIERS medications with similar structures but with variable moieties that can result in hydrophilic forms (e.g., pravastatin, rosuvastatin) and Both cancer therapy and cancer prevention have investigated drugs lipophilic forms (e.g., lovastatin, simvastatin, fluvastatin, atorvas- that modify specific targets in signal transduction pathways. Although tatin).201 Statins decrease the risk of cancer in preclinical studies the emphasis in drug development has focused on cancer treatment, by inhibiting RAS- and RHO-mediated cell proliferation, upregu- interventions aimed at modulating signal transduction pathways lating cell cycle inhibitors (e.g., p21 and p27), and inducing apop- promise new approaches to interventions in the carcinogenic process. tosis of transformed cells and the inhibition of angiogenesis.202 Because of the complexity of signaling systems, the inhibition of sin- gle targets may not be effective or may cause unacceptable toxicity. Evidence in Preclinical In Vivo Carcinogenesis Models Difluoromethylornithine Lipophilic statins delay progression of pancreatic intraepithelial Mechanism neoplasias and the growth of pancreatic carcinoma xenografts. Atorvastatin alone and in combination with NSAIDs reduced co- Polyamines (spermidine, spermine, and the diamine, putrescine) lonic adenoma and adenocarcinoma incidence and multiplicity by are required to maintain cellular growth and function.194 In mam- half in rodent transgenic and chemical carcinogenesis models. Lo- malian cells, polyamine inhibition by genetic mutation or phar- vastatin reduced lung adenoma multiplicity but not incidence.201 maceutical agents is associated with virtual cessation in cellular growth. Difluoromethylornithine (DFMO) is an enzyme-activated Clinical Trials irreversible inhibitor of ornithine decarboxylase (which is trans- activated by the c-MYC oncogene and cooperates with the RAS Although several large trials of pravastatin or simvastatin on car- oncogene in malignant transformation).195 diovascular disease risk with cancer as secondary end points have shown no benefit for reducing cancer risk with follow-ups between Evidence in Preclinical In Vivo Carcinogenesis 18 months to 4 years, these trials were not adequately powered to 201 Models examine cancer end points. Several case control studies evaluat- ing statin effects have shown a significant association with lower Extensive preclinical data has found that DFMO prevents tumor risk of colorectal adenocarcinoma with odds ratios ranging from promotion in a variety of systems, including skin, mammary, 0.53 to 0.91 for arzoxifene. A secondary analysis of a celecoxib pre- colon, cervical, and bladder carcinogenesis models.194 Synergistic vention trial demonstrated no statin protection against colorectal

tahir99 - UnitedVRG Chapter 33 Cancer Risk Reducing Agents 365

TABLE 33.11 Summary of Clinical Trials of Difluoromethylornithine as a Cancer Risk–Reducing Agent

Dose per Day, Population Duration Phase Endpoint Outcome References Low risk bladder cancer 1 gm versus placebo × III Bladder cancer Did not prevent or delay (293) (Ta. T1. Grades 1 or 2) 1 year recurrence recurrence Prostate risk: men with 500 mg versus placebo IIb Prostate volume, 10-fold reduction of prostate (294) family history prostate × 1 yr polyamines, PSA size increase over 1 yr cancer, age 35–70 yr compared to placebo, PSA reduction not significant Nonmelanoma skin 500 mg/m2 versus III New nonmelanoma Lower rate of basal cell (295, 296) cancer placebo × 4–5 y skin cancers carcinomas per year (0.28 versus 0.40); persistent reduction in nonmelanoma skin cancers, not statistically significant Colon adenomas DFMO: 500 mg IIb Adenoma recurrence RR for adenoma recurrence for (147) Sulindac: 150 mg DFMO/sulindac treatment 5 versus placebo × 3 y 0.30 (0.18–0.49)

PSA, prostate specific antigen. neoplasms.203 The Women’s Health Initiative (prospective longitu- Metformin dinal cohort of 159,319 women) found that lovastatin was associ- ated with a lower risk of developing colorectal cancer (HR = 0.62; Mechanism 95% CI, 0.39 to 0.99).204 Prospective longitudinal studies have shown mixed results. The PHS reported statin use was inversely Metformin, an oral antidiabetic drug in the biguanide class, is 212 associated with prostate cancer (adjusted RR, 0.51),205 whereas the first-line drug of choice for the treatment of type 2 diabetes. the Nurse’s Health Study showed no association with risk of breast Cancers are more common in diabetics and obese individuals cancer.206 Interventional trials to determine statin preventive ef- than their normal weight and normoglycemic counterparts, lead- CANCER PREVENTION AND SCREENING ficacy for colon and breast cancer are ongoing.201 Statins may be ing to the hypothesis that elevated serum insulin concentrations 213,214 effective risk-reducing agents in individuals with the A/A variant of promote cancer risk. Insulin and insulin-like growth factors the predominant T/T genotype of rs12654264 of the HMG-CoA (IGF1 and 2) stimulate cellular DNA synthesis, proliferation, and reductase gene.207 tumor growth through phosphoinositide-3 kinase (PI3K), mamma- lian target of rapamycin (mTOR), and the RAS-MAPK signaling pathways.213 Metformin activates the adenosine monophosphate- Bisphosphonates activated protein kinase (AMPK) via LKB1, a protein-threonine kinase that has tumor-suppressor activity.201 Metformin anticar- Mechanism cinogenesis activity appears to be broad and includes downregu- Bisphosphonates are pyrophosphate analogs with a central phos- lation of erbB-2 and epidermal growth factor receptor (EGFR) phorus-carbon-phosphorus bond that resists bone degradation expression, inhibiting the phosphorylation of erbB family mem- preventing bone loss and fractures. Second- and third-generation bers, IGF1R, Akt, mTOR, and STAT3 in vivo. Low doses of met- formin inhibit the self-renewal/proliferation of cancer stem cells in amino bisphosphonates (pamidronic, alendronic, risedronic, iban- 215,216 dronic, and zoledronic acids) inhibit farnesyl diphosphate synthase breast, colon, and pancreatic models. downstream of HMG-CoA reductase, leading to decreased post- translational prenylation of GTP-binding proteins such as RAS Evidence in Preclinical In Vivo Carcinogenesis and Rho. Amino bisphosphonates inhibit cell proliferation, angio- Models genesis, and cell cycle arrest while inducing apoptosis.201 Metformin reduces tobacco carcinogen–induced tumors in mice, Evidence in In Vivo Preclinical and Clinical Models and pancreatic premalignant and malignant tumors in ham- sters.201 However, metformin’s anticarcinogenic activity appears HER2-transgenic mice treated with zoledronic acid had increased dependent on the dose and the induced carcinogenesis process. tumor-free survival and overall survival. Zoledronic acid suppressed Metformin promoted carcinogenesis in MNU-induced rat breast bone, lung, and liver metastases when treated prior to an injection cancers, MMTV-Neu ER-negative breast cancers, OH-BBN in- of breast cancer cells.201 The short-term use of bisphosphonates is duced bladder cancer, and Min+ mouse intestinal tumors using associated with reduced breast cancer incidence in case control nonobese rodents.217 Metformin cancer risk–reducing agent ef- studies208,209 and prospective cohort studies (HR = 0.68; 95% CI, fects may be limited to obesity- and diabetes-associated carcino- 0.52 to 0.88 in a prospective cohort study).210 Randomized trials genesis mechanisms. of amino-bisphosphonates in postmenopausal women with breast cancer treated for 1 year have found a reduction of breast can- Clinical Trials cer risk in the contralateral breast (HR = 0.39; 95% CI, 0.18 to 0.88).211 Case control data suggesting a bisphosphonate-associated Two large retrospective cohort studies have shown that metformin reduction in colorectal cancer risk have not been confirmed by therapy is associated with a reduced risk of solid tumors by 25% to prospective cohort studies (i.e., Women’s Health Initiative, Nurse’s 30%.201 The Women’s Health Imitative observed a lower incidence Health Study).201 of invasive breast cancer in metformin-treated women with type 2 366 Cancer Prevention and Screening diabetes mellitus.218 Phase II window of opportunity randomized immunochemoprevention for epithelial targets for which an etio- trials have shown reduced proliferation and increased apoptosis in logic agent can be identified. resected tissue of breast cancer patients.201 Helicobacter pylori Diet-Derived Natural Products Intestinal-type gastric adenocarcinoma arises through a multistep Mechanism process that begins with chronic gastritis initiated by H. pylori, progressing through gastric mucosal atrophy, intestinal metapla- Polyphenolic phytochemicals, such as curcumin, resveratrol, epi- sia to dysplasia, and ultimately, to adenocarcinoma.246 H. pylori gallocatechin gallate (EGCG), genistein, and ginger, are attractive infects 50% of the world’s population.240 Infection occurs early in as cancer risk–reducing agents for their low toxicity and multimech- life, remains quiescent, and may be associated with chronic gas- anism anticarcinogenic properties. They have anti- inflammatory tritis of variable intensity but with minimal symptoms. Although activity, in part through scavenging of ROS, modulation of pro- the majority of H. pylori organisms remain in the gastric mucous tein kinase signal transduction pathways (e.g., STAT-3, HER2/ layer, 10% adhere to the gastric mucosa through adhesion BabA, neu, MAPK, and Akt), and downstream inhibition of eicosanoid an outer membrane protein that binds to the Lewis-B histo-blood synthesis potentially due to upstream inhibition of NF-κB and group antigen.240 Progression to atrophic gastritis and peptic ulcer PPAR or direct blockade or inhibition of eicosanoid-metabolizing disease (occurs in 10% to 15% of infected individuals) requires 219–221 enzymes. Curcumin, and presumably other polyphenolics, other bacterial and host cofactors.245,246 Infection with H. pylori downregulate stem cell driver signaling systems Wnt, Hedgehog, is associated with an OR of 2.7 to 6.0 for gastric cancer; CagA and Notch with subsequent reductions of breast, pancreatic, and increases this risk by 20- to 40-fold. The risk of developing gastric 222,223 colonic stem cell self-renewal. adenocarcinoma with an H. pylori infection is estimated to be 1% Omega-3 fatty acids (derived from marine products) compete to 3%.245,246 with omega-6 fatty acid substrates for eicosanoid-metabolizing en- The eradication of H. pylori with antibiotics and anti-inflam- zymes with subsequent tissue reduction of these inflammatory me- matory agents—for example, amoxicillin, metronidazole, and 224,225 diators. These fatty acids have other diverse anticarcinogenic bismuth subsalicylate—increases the rate of regression of non- mechanisms (e.g., G-protein inhibition, changes in membrane metaplastic gastric atrophy and intestinal metaplasia in geographi- physical characteristics that alter transmembrane signaling pro- cally diverse regions.247,248 A combination of a 2-week course of a tein dynamics) that make them attractive as cancer risk–reducing proton pump inhibitor (omeprazole) and an antibiotic (amoxicil- 226,227 agents. lin) reduced the risk of gastric cancer in a high-risk population in Whole berries, black raspberries, and strawberries contain China (OR = 0.61; (95% CI, 0.36 to 0.96) for 14.7 years after the mixtures of multiple anticarcinogenic compounds such as el- treatment.249 The sequence of giving proton pump inhibitors and 228 lagic acid, anthocyanins, and tocopherols. Research-grade antibiotic therapy does not alter the treatment outcome. In addi- berries are grown in a standardized cultivation environment and tion to contributing to gastric cancer risk, H. pylori infections may assayed for key components to ensure year-to-year reproducibil- also contribute to pancreatic cancer risk.250 Because H. pylori in- ity despite yearly climatologic variation. Berries have potent sta- fections are so widespread, mass eradication campaigns in high-risk bilization of methylation properties in addition to the expected regions are being considered.251 However, complicating this is that anti-inflammatory and antioxidative properties associated with the H. pylori infections have also been associated with a reduced risk of 229,230 prominent components. both esophageal adenocarcinoma and gastric cardia carcinoma.252

Preclinical and Clinical Anticarcinogenesis Efficacy. Di- verse diet-derived natural products have moderate-to-strong an- MULTIAGENT APPROACHES TO CANCER ticarcinogenic effects in both chemical and transgenic rodent RISK REDUCTION carcinogenesis models (Table 33.12). Phase I clinical trials of cur- cumin detected little parent compound in plasma or tissues, raising In the transition to molecularly targeted interventions, combina- the possibility of biologically active conjugates or deconjugation at 231–233 tions of targets that logically address critical carcinogenic pathways the target site. Resveratrol’s plasma bioavailability exceeds may have greater efficacy than single agents. For example, pre- that of curcumin and ginger, and partitions into human colon 219,234,235 viously demonstrated interactive signaling of EGFRs and COX-2 tissue at 10-fold concentrations compared to plasma. No experiments in Min+ mice253 demonstrates cancer preventive natural products have been studied in large prospective, cancer in- synergism. Combining atorvastatin with selective or nonselec- cidence risk–reduction trials. Using intraepithelial biomarker end tive COX inhibitors enhanced the inhibition of azoxymethane- points in human phase II trials, berry formulations reduce esopha- induced colon carcinogenesis in F344 rats and reduced the dose geal dysplasia and oral leukoplakia.236,237 Curcumin reduces the 238 of the combined drugs required to achieve a reduction of colon number of colon aberrant crypt foci in human smokers. carcinogenesis.254 DFMO plus sulindac inhibited adenoma formation in a phase IIb trial of 375 patients with a prior history of adenomas followed ANTI-INFECTIVES for 36 months (see Table 33.11).147 Cardiovascular-adverse out- comes were higher in DFMO/sulindac-treated patients who had Many infectious agents are known causes of human cancers, in- preexisting high baseline cardiovascular risk; however, the cardio- cluding the human hepatitis viruses, hepatitis B virus (HBV) vascular-adverse events were similar to placebo in moderate or low and hepatitis C virus (HCV) for hepatocellular carcinoma239; cardiovascular risk patients.255 Using IGF-1 as a biomarker, Guer- Helicobacter pylori for gastric adenocarcinoma240; human pap- rieri-Gonzaga et al.256 showed that the combination of low-dose illoma viruses (HPV) for cervical, anal, vulva, penis, and oral tamoxifen with low-dose fenretinide is safe but not synergistic. cavity and pharynx carcinomas241; herpes virus-8 for Kaposi sar- As more data accumulate from in vivo models, combined drugs coma242; Epstein-Barr virus for Burkitt and other lymphomas243; aimed at specific targets in coordinated signaling pathways will liver flukes for cholangiocarcinoma244; and schistosomes for blad- enter clinical biomarker-based trials. Optimal doses, toxicity, and der carcinoma.245 The success of the HPV vaccine at reducing biomarker modulation data will select those combinations useful the incidence of intraepithelial neoplasia of the cervix (reviewed for risk reduction trials and, ultimately, generalized use in at-risk in Chapter 72) is one example that demonstrates the potential of populations.

tahir99 - UnitedVRG Chapter 33 Cancer Risk Reducing Agents 367

TABLE 33.12 Selected Diet-Derived Natural Products with Cancer Risk–Reducing Activity

In Vivo Anticarcinogenesis Nutritional Extract Source Mechanisms Efficacy Human Trials References

Curcumin ([1E,6E]- Turmeric, rhizome Inhibits: PGE2 synthesis Colon, breast, skin Phase I: Poor (219, 238) 1,7-bis-[4-hydroxy- of Curcuma via direct binding to bioavailability due to 3-methoxyphenyl]- longa COX-2 and through biotransformation in 1,6-heptadiene- inhibition of NF-κB; gut, enterohepatic 3,5-dione/ angiogenesis. ErbB2 cycling of metabolites; diferuloylmethane) transduction; PI3K-Akt Phase IIa: reduced transduction; Inhibits aberrant crypt foci stem cell self renewal Agonist: vitamin D receptor Resveratrol Grapes, Inhibits: Carcinogen Colorectal, breast, Phase I: 1 g dose (220, 234) (3,5,4′-trihydroxy- mulberries, activation via inhibition pancreas, skin, generated peak trans-stilbene) peanuts, of phase I isozyme, and prostate concentration ∼2 μM, and Cassia eicosanoids via direct conjugates 10-fold quinquangulata binding; NF-κB; Nrf2. higher; resveratrol plants Acts as a caloric tissue concentrations restriction mimetic, 10-fold higher than activates the histone plasma; deacetylase SIRT1 and Phase IIa: Small AMPK reduction in IGF-1 and IGFBP-1 Ginger (gingerols, Rhizome of Induces apoptosis Colon, breast, skin, Phase I: 2 g dose (235, 297, 298) paradols, Zingiber via caspase-3 oral cavity, liver nontoxic; shagaols) officinale mechanisms; Inhibits Phase IIa: Small

NF-κB activation and reductions in PGE2, downstream COX-2 increased Bax in

expression; reduces upper colon crypt CANCER PREVENTION AND SCREENING iNOS expression and ornithine decarboxylase activity Green tea Green tea extract Inhibits: PI3K-Akt Lung, prostate, skin, Phase IIa: 500– (221) (epigallocatechin transduction, IGF-1, colorectal 1,000 mg/m2 × 12 gallate, other IGFBP-3; NK-κB; wk reduced oral catechins) catenin reduces premalignant lesions methylation via in 50%; inhibition of DNA Phase IIb: 2.5 g × 1 yr methyltransferase 1 reduced colorectal adenoma recurrence by 50% Omega-3 fatty acids Fish oil Reduction of Colon, breast, Phase II: 4–7 mg/d (226, 299) (eicosapentaenoic inflammation via prostate reduced colon acid; eicosanoid reduction; adenomas in familial docosahexaenoic direct binding to G adenomatous acid) receptor proteins; polyposis; ongoing PPAR activation; trials for sporadic; induction of anti- extensive case control inflammatory lipid studies mediators (resolvins, protectins, maresins) Berries Black raspberries, Reduction of methylation Esophageal Phase IIb: Freeze dried (228, 229, strawberries via inhibition of squamous cell, strawberries reduced 236, 237) methyltransferases, colon, skin esophageal dysplasia; re-regulated Wnt; Phase IIa: Blackberry gel inhibits NF-κB; inhibits reduced leukoplakia cyclooxygenases; inhibits proliferation

IGFBP-1, insulin growth factor binding protein-1; iNOS, inducible isoform of nitric oxide synthase. 368 Cancer Prevention and Screening

SELECTED REFERENCES

The full reference list can be accessed at lwwhealthlibrary.com/oncology. 74. Meyskens FL Jr, Surwit E, Moon TE, et al. Enhancement of regression of cervical intraepithelial neoplasia II (moderate dysplasia) with topically ap- 4. Lippman SM, Levin B, Brenner DE, et al. Cancer prevention and the Amer- plied all-trans-retinoic acid: a randomized trial. J Natl Cancer Inst 1994;86: ican Society of Clinical Oncology. J Clin Oncol 2004;22:3848–3851. 539–543. 8. Steele VE, Boone CW, Lubet RA, et al. Preclinical drug development para- 75. Blot WJ, Li JY, Taylor PR, et al. Nutrition intervention trials in Linxian, digms for chemopreventives. Hematol Oncol Clin North Am 1998;12:943–961. China: supplementation with specific vitamin/mineral combinations, can- 22. Schatzkin A, Freedman LS, Schiffman MH, et al. Validation of intermediate cer incidence, and disease-specific mortality in the general population.J Natl end points in cancer research. J Natl Cancer Inst 1990;82:1746–1752. Cancer Inst 1993;85:1483–1492. 25. Pepe MS, Feng Z, Janes H, et al. Pivotal evaluation of the accuracy of a bio- 77. Mayne ST, Ferrucci LM, Cartmel B. Lessons learned from randomized clini- marker used for classification or prediction: standards for study design. J Natl cal trials of micronutrient supplementation for cancer prevention. Annu Rev Cancer Inst 2008;100:1432–1438. Nutr 2012;32:369–390. 28. Fisher B, Costantino J, Wickerham D, et al. Tamoxifen for prevention of 78. Li JY, Taylor PR, Li B, et al. Nutrition intervention trials in Linxian, China: breast cancer: report of the National Surgical Adjuvant Breast and Bowel multiple vitamin/mineral supplementation, cancer incidence, and disease- Project P-1 study. J Natl Cancer Inst 1998;90:1371–1388. specific mortality among adults with esophageal dysplasia. J Natl Cancer Inst 29. Vogel VG, Costantino JP, Wickerham DL, et al. Update of the National 1993;85:1492–1498. Surgical Adjuvant Breast and Bowel Project Study of Tamoxifen and Ral- 81. Greenberg ER, Baron JA, Tosteson TD, et al. A clinical trial of antioxidant oxifene (STAR) P-2 Trial: Preventing breast cancer. Cancer Prev Res (Phila) vitamins to prevent colorectal adenoma. Polyp Prevention Study Group. 2010;3:696–706. N Engl J Med 1994;331:141–147. 30. Thompson IM, Goodman PJ, Tangen CM, et al. The influence of finasteride 83. Heinonen OP, Albanes D, Virtamo J, et al. Prostate cancer and supplementa- on the development of prostate cancer. N Engl J Med 2003;349:215–224. tion with alpha-tocopherol and beta-carotene: incidence and mortality in a 31. Omenn G, Goodman G, Thornquist M, et al. Effects of a combination of controlled trial. J Natl Cancer Inst 1998;90:440–446. beta carotene and vitamin A on lung cancer and cardiovascular disease. 84. Lippman SM, Klein EA, Goodman PJ, et al. Effect of selenium and vitamin N Engl J Med 1996;334:1150–1155. E on risk of prostate cancer and other cancers: the Selenium and Vitamin E 38. World Cancer Research Fund, American Institute for Cancer Research. Cancer Prevention Trial (SELECT). JAMA 2009;301:39–51. Food, Nutrition, Physical Activity, and the Prevention of Cancer: A Global 99. Gibson TM, Weinstein SJ, Pfeiffer RM, et al. Pre- and postfortification intake Perspective. Washington, DC: AICR; 2007. of folate and risk of colorectal cancer in a large prospective cohort study in 43. Hong WK, Lippman SM, Itri LM, et al. Prevention of second primary tu- the United States. Am J Clin Nutr 2011;94:1053–1062. mors with 13cRA in squamous-cell carcinoma of the head and neck. N Engl 106. Cole BF, Baron JA, Sandler RS, et al. Folic acid for the prevention of colorec- J Med 1990;323:795–801. tal adenomas: a randomized clinical trial. JAMA 2007;297:2351–2359. 45. Khuri FR, Lee JJ, Lippman SM, et al. Randomized phase III trial of low-dose 107. Logan RF, Grainge MJ, Shepherd VC, et al. Aspirin and folic acid for the pre- isotretinoin for prevention of second primary tumors in stage I and II head vention of recurrent colorectal adenomas. Gastroenterology 2008;134:29–38. and neck cancer patients. J Natl Cancer Inst 2006;98:441–450. 108. Wu K, Platz EA, Willett WC, et al. A randomized trial on folic acid supple- 48. Mayne ST, Cartmel B, Baum M, et al. Randomized trial of supplemental mentation and risk of recurrent colorectal adenoma. Am J Clin Nutr 2009; beta-carotene to prevent second head and neck cancer. Cancer Res 2001; 90:1623–1631. 61:1457–1463. 114. Chung M, Balk EM, Brendel M, et al. Vitamin D and Calcium: A Systematic 49. Bairati I, Meyer F, Jobin E, et al. Antioxidant vitamins supplementation and Review of Health Outcomes. Evidence Report No. 183 (Prepared by the Tufts mortality: a randomized trial in head and neck cancer patients. Int J Cancer Evidence-based Practice Center). Rockville, MD: Agency for Healthcare Re- 2006;119:2221–2224. search and Quality; 2009. 52. The Alpha-Tocopherol Beta Carotene Cancer Prevention Study Group. The 117. Baron J, Beach M, Mandel JS, et al. Calcium supplements for the preven- effect of vitamin E and beta carotene on the incidence of lung cancer and tion of colorectal adenomas. The Calcium Polyp Prevention Study Group. other cancers in male smokers. N Engl J Med 1994;330:1029–1035. N Engl J Med 1999;340:101–107. 57. van Zandwijk N, Dalesio O, Pastorino U, et al. EUROSCAN, a randomized 119. Wactawski-Wende J, Kotchen JM, Anderson GL, et al. Calcium plus vita- trial of vitamin A and N-acetylcysteine in patients with head and neck cancer min D supplementation and the risk of colorectal cancer. N Engl J Med or lung cancer. For the European Organization for Research and Treatment 2006;354:684–696. of Cancer Head and Neck and Lung Cancer Cooperative Groups. J Natl 122. Chlebowski RT, Johnson KC, Kooperberg C, et al. Calcium plus vitamin D Cancer Inst 2000;92:977–986. supplementation and the risk of breast cancer. J Natl Cancer Inst 2008; 58. Lippman SM, Lee JJ, Karp DD, et al. Randomized phase III intergroup trial 100:1581–1591. of isotretinoin to prevent second primary tumors in stage I non-small-cell 132. Rothwell PM, Price JF, Fowkes FG, et al. Short-term effects of daily aspi- lung cancer. J Natl Cancer Inst 2001;93:605–618. rin on cancer incidence, mortality, and non-vascular death: analysis of the 60. Veronesi U, De Palo G, Marubini E, et al. Randomized trial of fenretinide to time course of risks and benefits in 51 randomised controlled trials. Lancet prevent second breast malignancy in women with early breast cancer. J Natl 2012;379:1602–1612. Cancer Inst 1999;91:1847–1856. 133. Thun MJ, Namboodiri MM, Heath C Jr. Aspirin use and reduced risk of fatal 62. Uray IP, Brown PH. Chemoprevention of hormone receptor-negative breast colon cancer. N Engl J Med 1991;325:1593–1596. cancer: new approaches needed. Recent Results Cancer Res 2011;188:147–162. 135. Chan AT, Giovannucci EL, Meyerhardt JA, et al. Long-term use of aspi- 63. Kraemer KH, DiGiovanna JJ, Moshell AN, et al. Prevention of skin cancer rin and nonsteroidal anti-inflammatory drugs and risk of colorectal cancer. in xeroderma pigmentosum with the use of oral isotretinoin. N Engl J Med JAMA 2005;294:914–923. 1988;318:1633–1637. 139. Kawamori T, Rao C, Seibert K, et al. Chemopreventive effect of celecoxib, 65. Tangrea JA, Edwards BK, Taylor PR, et al. Long-term therapy with low-dose a specific cyclooxygenase-2 inhibitor on colon carcinogenesis. Cancer Res isotretinoin for prevention of basal cell carcinoma: a multicenter clinical 1998;58:409–412. trial Isotretinoin-Basal Cell Carcinoma Study Group. J Natl Cancer Inst 142. Giardiello FM, Yang VW, Hylind LM, et al. Primary chemoprevention of 1992;84:328–332. familial adenomatous polyposis with sulindac. N Engl J Med 2002;346: 66. Levine N, Moon TE, Cartmel B, et al. Trial of retinol and isotretinoin in skin 1054–1059. cancer prevention: a randomized, double-blind, controlled trial. Southwest 143. Steinbach G, Lynch PM, Phillips RK. The effect of celecoxib, a cyclooxy- Skin Cancer Prevention Study Group. Cancer Epidemiol Biomarkers Prev genase-2 inhibitor, in familial adenomatous polyposis. N Engl J Med 2000; 1997;6:957–961. 342:1946–1952. 67. Moon TE, Levine N, Cartmel B, et al. Effect of retinol in preventing squa- 147. Meyskens FL Jr, McLaren CE, Pelot D, et al. Difluoromethylornithine plus mous cell skin cancer in moderate-risk subjects: a randomized, double-blind, sulindac for the prevention of sporadic colorectal adenomas: a randomized controlled trial. Southwest Skin Cancer Prevention Study Group. Cancer placebo-controlled, double-blind trial. Cancer Prev Res (Phila) 2008;1:32–38. Epidemiol Biomarkers Prev 1997;6:949–956. 149. Bertagnolli MM, Eagle CJ, Zauber AG, et al. Celecoxib for the prevention of 68. Greenberg ER, Baron JA, Stukel TA, et al. A clinical trial of beta carotene to sporadic colorectal adenomas. N Engl J Med 2006;355:873–884. prevent basal-cell and squamous-cell cancers of the skin. The Skin Cancer 151. Bresalier RS, Sandler RS, Quan H, et al. Cardiovascular events associated Prevention Study Group. N Engl J Med 1990;323:789–895. with rofecoxib in a colorectal adenoma chemoprevention trial. N Engl J Med 69. Green A, Williams G, Neale R, et al. Daily sunscreen application and beta- 2005;352:1092–1102. carotene supplementation in prevention of basal-cell and squamous-cell car- 156. Rothwell PM. Aspirin in prevention of sporadic colorectal cancer: current cinomas of the skin: a randomised controlled trial. Lancet 1999;354:723–729. clinical evidence and overall balance of risks and benefits. Recent Results 70. Clark LC, Combs GF Jr, Turnbull BW, et al. Effects of selenium supple- Cancer Res 2013;191:121–142. mentation for cancer prevention in patients with carcinoma of the skin. A 157. U.S. Preventive Services Task Force. Routine aspirin or nonsteroidal anti- randomized controlled trial. Nutritional Prevention of Cancer Study Group. inflammatory drugs for the primary prevention of colorectal cancer: U.S. JAMA 1996;276:1957–1963. Preventive Services Task Force recommendation statement. Ann Intern Med 71. Duffield-Lillico AJ, Slate EH, Reid ME, et al. Nutritional Prevention of 2007;146:361–364. Cancer Study Group. Selenium supplementation and secondary preven- 161. Kuiper GG, Carlsson B, Grandien K, et al. Comparison of the ligand binding tion of nonmelanoma skin cancer in a randomized trial. J Natl Cancer Inst specificity and transcript tissue distribution of estrogen receptors alpha and 2003;95:1477–1481. beta. Endocrinology 1997;138:863–870.

tahir99 - UnitedVRG Chapter 33 Cancer Risk Reducing Agents 369

164. Jordan VC. Tamoxifen (ICI46,474) as a targeted therapy to treat and prevent 234. Gescher A, Steward WP, Brown K. Resveratrol in the management of human breast cancer. Br J Pharmacol 2006;147:S269–S276. cancer: how strong is the clinical evidence? Ann N Y Acad Sci 2013;1290:12–20. 169. Ettinger B, Black DM, Mitlak BH, et al. Reduction of vertebral fracture risk 236. Stoner GD, Wang LS. Chemoprevention of esophageal squamous cell carci- in postmenopausal women with osteoporosis treated with raloxifene: results noma with berries. Top Curr Chem 2013;329:1–20. from a 3-year randomized clinical trial. Multiple Outcomes of Raloxifene 238. Carroll RE, Benya RV, Turgeon DK, et al. Phase IIa clinical trial of cur- Evaluation (MORE) Investigators. JAMA 1999;282:637–645. cumin for the prevention of colorectal neoplasia. Cancer Prev Res (Phila) 175. Moyer VA. Medications for risk reduction of primary breast cancer in women: 2011;4:354–364. U.S. Preventive Services Task Force recommendation statement. Ann Intern 249. Ma JL, Zhang L, Brown LM, et al. Fifteen-year effects of Helicobacter py- Med 2013;159:698-708. lori, garlic, and vitamin treatments on gastric cancer incidence and mortality. 176. Visvanathan K, Hurley P, Bantug E, et al. Use of pharmacologic interven- J Natl Cancer Inst 2012;104:488–492. tions for breast cancer risk reduction: American Society of Clinical Oncology 250. Risch HA, Lu L, Kidd MS, et al. Helicobacter pylori seropositivities and risk of clinical practice guideline. J Clin Oncol 2013;31:2942–2962. pancreatic carcinoma. Cancer Epidemiol Biomarkers Prev 2014;23:172–178. 177. Vogel VG, Costantino JP, Wickerham DL, et al. Carcinoma in situ outcomes 255. Zell JA, Pelot D, Chen WP, et al. Risk of cardiovascular events in a random- in National Surgical Adjuvant Breast and Bowel Project Breast Cancer Che- ized placebo-controlled, double-blind trial of difluoromethylornithine plus moprevention Trials. J Natl Cancer Inst Monogr 2010;2010:181–186. sulindac for the prevention of sporadic colorectal adenomas. Cancer Prev Res 178. Cuzick J, Sestak I, Bonanni B, et al. Selective oestrogen receptor modulators (Phila) 2009;2:209–212. in prevention of breast cancer: an updated meta-analysis of individual partici- 256. Guerrieri-Gonzaga A, Robertson C, Bonanni B, et al. Preliminary results pant data. Lancet 2013;381:1827–1834. on safety and activity of a randomized, double-blind, 2 × 2 trial of low-dose 179. Barrett-Connor E, Mosca L, Collins P, et al. Effects of raloxifene on cardio- tamoxifen and fenretinide for breast cancer prevention in premenopausal vascular events and breast cancer in postmenopausal women. N Engl J Med women. J Clin Oncol 2006;24:129–135. 2006;355:125–137. 257. Kelloff GJ, Lippman SM, Dannenberg AJ, et al. Progress in chemopreven- 183. Goss PE, Ingle JN, Ales-Martinez JE, et al. Exemestane for breast-cancer tion drug development: the promise of molecular biomarkers for preven- prevention in postmenopausal women. N Engl J Med 2011;364:2381–2391. tion of intraepithelial neoplasia and cancer—a plan to move forward. Clin 185. Cuzick J, Sestak I, Forbes JF, et al. Anastrozole for prevention of breast can- Cancer Res 2006;12:3661–3697. cer in high-risk postmenopausal women (IBIS-II): an international, double- 258. Washington MK, Powell AE, Sullivan R, et al. Pathology of rodent models blind, randomised placebo-controlled trial. Lancet 2014;383:1041–1048. of intestinal cancer: progress report and recommendations. Gastroenterology 187. King MC, Wieand S, Hale K, et al. Tamoxifen and breast cancer incidence 2013;144:705–717. among women with inherited mutations in BRCA1 and BRCA2: National 264. Winawer SJ, Zauber AG, Ho MN, et al. Prevention of colorectal cancer by Surgical Adjuvant Breast and Bowel Project (NSABP-P1) Breast Cancer Pre- colonoscopic polypectomy. The National Polyp Study Workgroup. N Engl J vention Trial. JAMA 2001;286:2251–2256. Med 1993;329:1977–1981. 192. Andriole GL, Bostwick DG, Brawley OW, et al. Effect of dutasteride on the 267. Taylor P, Li B, Dawsey S, et al. Prevention of esophageal cancer: the nutrition risk of prostate cancer. N Engl J Med 2010;362:1192–1202. intervention trials in Linxian, China. Linxian Nutrition Intervention Trials 193. Thompson IM Jr, Goodman PJ, Tangen CM, et al. Long-term survival of Study Group. Cancer Res 1994;54:2029s–2031s. participants in the prostate cancer prevention trial. N Engl J Med 2013;369: 278. Giardiello FM, Hamilton SR, Krush AJ, et al. Treatment of colonic and rec- 603–610. tal adenomas with sulindac in familial adenomatous polyposis. N Engl J Med 198. Meyskens FL Jr, Surwit E, Moon TE, et al. Enhancement of regression of 1993;328:1313–1316. cervical intraepithelial neoplasia II (moderate dysplasia) with topically ap- 281. Burn J, Gerdes AM, Macrae F, et al. Long-term effect of aspirin on cancer plied all-trans-retinoic acid: randomized trial. J Natl Cancer Inst 1994;86: risk in carriers of hereditary colorectal cancer: an analysis from the CAPP2 539–543. randomised controlled trial. Lancet 2011;378:2081–2087. 200. Jeter JM, Alberts DS. Difluoromethylornithine: the proof is in the poly- 282. Ruffin MT, Krishnan K, Rock CL, et al. Suppression of human colorectal amines. Cancer Prev Res (Phila) 2012;5:1341–1344. mucosal prostaglandins: determining the lowest effective aspirin dose. J Natl

201. Gronich N, Rennert G. Beyond aspirin-cancer prevention with statins, met- Cancer Inst 1997;89:1152–1160. AND SCREENING CANCER PREVENTION formin and bisphosphonates. Nat Rev Clin Oncol 2013;10:625–642. 286. Baron JA, Cole BF, Sandler RS, et al. A randomized trial of aspirin to prevent 203. Bertagnolli MM, Hsu M, Hawk ET, et al. Statin use and colorectal adenoma colorectal adenomas. N Engl J Med 2003;348:891–899. risk: results from the adenoma prevention with celecoxib trial. Cancer Prev 288. Cuzick J, Forbes JF, Sestak I, et al. Long-term results of tamoxifen prophy- Res (Phila) 2010;3:588–596. laxis for breast cancer—96-month follow-up of the randomized IBIS-I trial. 208. Rennert G. Bisphosphonates: beyond prevention of bone metastases. J Natl J Natl Cancer Inst 2007;99:272–282. Cancer Inst 2011;103:1728–1729. 290. Veronesi U, Maisonneuve P, Rotmensz N, et al. Tamoxifen for the prevention 215. Zhu P, Davis M, Blackwelder A, et al. Metformin selectively targets tumor of breast cancer: late results of the Italian Randomized Tamoxifen Prevention initiating cells in erbB-2 overexpressing breast cancer models. Cancer Prev Trial among women with hysterectomy. J Natl Cancer Inst 2007;99:727–737. Res (Phila) 2014;7:199–210. 303. Dontu G, Liu S, Wicha MS. Stem cells in mammary development and 218. Chlebowski RT, McTiernan A, Wactawski-Wende J, et al. Diabetes, met- carcinogenesis: implications for prevention and treatment. Stem Cell Rev formin, and breast cancer in postmenopausal women. J Clin Oncol 2012; 2005;1:207–213. 30:2844–2852. 308. Califano J, van der Riet P, Westra W, et al. Genetic progression model for 221. Lambert JD. Does tea prevent cancer? Evidence from laboratory and human head and neck cancer: implications for field cancerization.Cancer Res 1996; intervention studies. Am J Clin Nutr 2013;98:1667S–1675S. 56:2488–2492. 222. Kakarala M, Brenner DE, Korkaya H, et al. Targeting breast stem cells with 309. Califano J, Westra WH, Meininger G, et al. Genetic progression and clonal the cancer preventive compounds curcumin and piperine. Breast Cancer Res relationship of recurrent premalignant head and neck lesions. Clin Cancer Treat 2010;122:777–785. Res 2000;6:347–352. 226. Laviano A, Rianda S, Molfino A, et al. Omega-3 fatty acids in cancer. Curr 312. Barrett MT, Sanchez CA, Prevo LJ, et al. Evolution of neoplastic cell Opin Clin Nutr Metab Care 2013;16:156–161. lineages in Barrett oesophagus. Nat Genet 1999;22:106–109. 34 Cancer Screening

Otis W. Brawley and Howard L. Parnes

INTRODUCTION they do not provide any information regarding a test’s efficacy or effectiveness. Cancer screening refers to a test or examination performed on an ■ Sensitivity is the proportion of persons designated positive by asymptomatic individual. The goal is not simply to find cancer at the screening test among all individuals who have the disease: an early stage, nor is it to diagnose as many patients with cancer true positive (TP)/(TP + false negative [FN]). as possible. The goal of cancer screening is to prevent death and ■ Specificity is the proportion of persons designated negative by suffering from the disease in question through early therapeutic the screening test among all individuals who do not have the intervention. disease: true negative TN/(TN + false positive [FP]). The assumption that early detection improves outcomes can ■ Positive predictive value is the proportion of individuals with a be traced back to the concept that cancer inexorably progresses positive screening test who have the disease: (TP)/(TP + FP). from a small, localized, primary tumor to local–regional spread, ■ Negative predictive value is the proportion of individuals with to distant metastases and death. This linear model of disease pro- a negative screening test who do not have the disease: (TN)/ gression predicts that early intervention would reduce cancer (TN + false negative [FN]).2 mortality. Cancer screening was an element of the “periodic physical ex- For a given screening test, sensitivity and specificity are in- amination,” as espoused by the American Medical Association in versely related. For example, as one lowers the threshold for con- the 1920s.1 It consisted of palpation to find a mass or enlarged sidering a serum prostate-specific antigen (PSA) level to represent lymph nodes and auscultation to find a rub or abnormal sound. a positive screen, the sensitivity of the test increases and more can- Today, screening has grown to include radiologic testing, the mea- cers will be detected. This increased sensitivity comes at the cost surement of serum markers of disease, and even molecular testing. of decreased specificity (i.e., more men without cancer will have A positive screening test leads to further diagnostic testing, which positive screenings tests and, therefore, will be subjected to unnec- might lead to a cancer diagnosis. essary diagnostic procedures). The intuitive appeal of early detection accounts for the em- Some screening tests, such as mammograms, are more subjec- phasis that has long been placed on screening. However, it is not tive and operator dependent than others. For this reason, the sen- widely understood that screening tests are always associated with sitivity and specificity of screening mammography varies among some harm (e.g., anxiety, financial costs) and may actually cause radiologists. For a given radiologist, the lower his or her threshold substantial harm (e.g., invasive follow-up diagnostic or therapeutic for considering a mammogram to be suspicious, the higher the procedures). Because screening is, by definition, done in healthy sensitivity and lower the specificity will be for them. However, people, all early detection tests should be carefully studied and mammography can have both a higher sensitivity and higher spec- their risk–benefit ratio determined before they are adopted for ificity in the hands of a more experienced versus a less experienced widespread usage. radiologist. Screening is a public health intervention. However, some As opposed to sensitivity and specificity, the PPV and NPV draw a distinction between screening an individual within the of a screening test are dependent on disease prevalence. PPV doctor–patient relationship and mass screening, a program aimed is also highly responsive to small increases in specificity. As at screening a large population. The latter may involve advertis- shown in Table 34.2, given a disease prevalence of 5 cases per ing campaigns to encourage people to be screened for a particu- 1,000 (0.005), the PPV of a hypothetical screening test increases lar cancer at a shopping mall or at a community event, such as dramatically as specificity goes from 95% to 99.9%, but only state fair. marginally as sensitivity goes from 80% to 95%. Given a dis- Screening may be either opportunistic (i.e., a patient sees a ease prevalence of only 1 per 10,000 (0.0001), the PPV of the health-care provider who chooses to screen or not to screen) or same test is poor even at high sensitivity and specificity. The programmatic. Programmatic refers to a standardized approach positive association between breast cancer prevalence and age with algorithms for screening and follow-up as well as recall of pa- is the major reason why screening mammography is a better test tients for regular routine screening with quality control measures. (higher PPV) for women aged 50 to 59 than for women 40 to Programmatic screening is usually more effective. 49 years of age.

PERFORMANCE CHARACTERISTICS ASSESSING SCREENING TESTS AND OUTCOMES The degree to which a screening test can discriminate between individuals with and without a particular disease is described by its performance characteristics. These include the a test’s sen- Screening Test Results sitivity, specificity, positive predictive value (PPV), and nega- tive predictive value (NPV) (Table 34.1). It should be noted Lead time bias occurs whenever screening results in an earlier that these measures relate to the accuracy of a screening test; diagnosis than would have occurred in the absence of screening.

370

tahir99 - UnitedVRG Chapter 34 Cancer Screening 371

TABLE 34.1 Lead Time Bias

Performance Characteristics of a Screening Test Diagnosis due Death due to symptoms to cancer Sensitivity is the proportion designated positive by the screening test among all individuals who have the disease. A TP TP + FN Specificity is the proportion designated negative by the screening Diagnosis due Death due test among all those who do not have the disease. to screening to cancer TN TN + FP Positive predictive value is the proportion of individuals with a Lead time positive test who have the disease. B TP TP + FP Diagnosis due Death due to cancer Negative predictive value is the proportion of individuals with a to screening or other causes negative test negative who do not have the disease. TN TN + FN Lead time Prolonged life due TP, true positive, the condition present and the test is positive; FN, false to screening negative, the condition is present and the test is negative; FP, false positive, C the condition is absent and the test is positive; TN, true negative, the condition is absent and test is negative. Figure 34.1 Survival is the time from cancer diagnosis to death. (A) Lead time bias occurs when screening results in an earlier diagnosis. Without screening, a patient is diagnosed with cancer due to symptoms. (B) With screening, the patient is often diagnosed earlier. When screening and treatment do not prolong life, the screened patient can Because survival is measured from the time of diagnosis, an earlier have a longer survival solely due to the earlier diagnosis. The survival diagnosis, by definition, increases survival. Unless an effective in- increase is pure lead time bias. (C) When screening and treatment are tervention is available, lead time bias has no impact on the natural beneficial, the patient is diagnosed before the onset of symptoms and history of a disease and death will occur at the same time it would the patient lives beyond the point in which death would have occurred have in the absence of early detection (Fig. 34.1). without screening. Length bias is a function of the biologic behavior of a can-

cer. Slower growing, less aggressive cancers are more likely to be CANCER PREVENTION AND SCREENING detected by a screening test than faster growing cancers, which but are not biologically destined to harm the patient (see are more likely to be diagnosed due to the onset of symptoms Fig. 34.2). between scheduled screenings (interval cancers). Length bias has There are two categories of overdiagnosis: the detection of his- an even greater effect on survival statistics than lead time bias tologically definedcancers not destined to metastasize or harm the (Fig. 34.2). patient, and the detection of cancers not destined to metastasize or Overdiagnosis is an extreme form of length bias and rep- cause harm in the life span of the specific patient. The importance resents pure harm. It refers to the detection of tumors, often of this second category is illustrated by the widespread practice through highly sensitive modern imaging modalities and other in the United States of screening elderly patients with limited life diagnostic tests, that fulfill the histologic criteria formalignancy expectancies, who are thus unlikely to benefit from early cancer diagnosis. Overdiagnosis occurs with many malignancies, including lung, 3 TABLE 34.2 breast, prostate, renal cell, melanoma, and thyroid cancers. Neu- roblastoma provides one of the most striking examples of overdi- Positive Predictive Value Given Varying Sensitivity agnosis.4 Urine vanillylmandelic acid (VMA) testing is a highly and Specificity and Prevalence sensitive screening test for the detection of this pediatric disease. After screening programs in Germany, Japan, and Canada showed marked increases in the incidence of this disease without a con- Sensitivity % Prevalence 0.005 comitant decline in mortality, it was noticed that nearby areas that 80 90 95 did not screen had similar death rates with lower incidence.4,5 It is Specificity % 95 7 8 9now appreciated that screen-detected neuroblastomas have a very good prognosis with minimal or no treatment. Many actually re- 99 29 31 32 gress spontaneously. 99.9 80 82 83 Stage shift—i.e., a cancer diagnosis at an earlier stage than Sensitivity % would have occurred in the absence of screening—is necessary, Prevalence 0.0001 but not sufficient, for a screening test to be effective in terms of 80 90 95 reducing mortality. Both lead time bias and length bias contrib- Specificity % 95 0.2 0.2 2.0 ute to this phenomenon. Although it is tempting to speculate that diagnosis at an earlier stage must confer benefit, this is not 99 0.8 0.9 0.9 necessarily the case. For example, a substantial proportion of 99.9 0.7 8.0 9.0 men treated with radical prostatectomy for what appears to be a localized prostate cancer relapse after undergoing surgery. Con- PPV improves dramatically in response to small changes in specificity. Changes in specificity influence PPV much more than changes in sensitivity. versely, some men who are treated with definitive therapy would Note the influence of prevalence on PPV. Screening tests do not perform as never have gone on to develop metastatic disease in the absence well in populations with a low prevalence of disease. of treatment. 372 Cancer Prevention and Screening

Length Bias and Cancer Screening Screen Screen Screen Screen

Tumor burden Figure 34.2 Length bias and cancer causes death screening. The red line is indicative of a fast-growing tumor that is not amenable to regular screening. The blue line is Tumor burden indicative of a fast-growing tumor that causes symptoms can be diagnosed by screening or later by symptoms; death may possibly be prevented by treatment. The green line is a slower growing but potentially deadly cancer that can be detected by symptoms or several screenings and treated, possibly preventing death. The orange line is indicative of a very slow growing tumor that would never cause death and would never need treatment Tumor begins despite being screen detected. This is Time classic overdiagnosis.

Selection bias occurs when enrollees in a clinical study dif- drop out of the study. Both drop-ins and drop-outs reduce the sta- fer from the general population. In fact, people who voluntarily tistical power of a clinical trial. participate in clinical trials tend to be healthier than the general In the United States, it is now considered standard to obtain in- population, perhaps due to a greater interest in health and health- formed consent before randomization takes place. However, there care research. Screening studies tend to enroll individuals health- have been several published studies that randomized participants ier than the general population. This so-called healthy volunteer from rosters of eligible subjects such as census lists. In these tri- effect6,7 can introduce a powerful bias if not adequately controlled als, informed consent was obtained after randomization and only for by randomization procedures. among those randomized to the screening arm of the study. Those randomized to the control arm were not contacted, and indeed, Assessing Screening Outcomes did not know they were in a clinical trial. They were followed through national death registries. Although the study was analyzed The usual primary goal of cancer screening is to reduce mortality on an intent-to-screen basis, this method can still introduce biases. from the disease in question (a reduction in disease-specific mor- For example, only patients on the intervention arm had access to tality). Screening studies generally do not have sufficient statistical the screening facility and staff for counseling and treatment if diag- power to assess the impact of screening for a specific malignancy nosed; those in the control group were more likely to be treated in on overall mortality. (Lung cancer screening provides an exception the community as opposed to high-volume centers of excellence to this rule; see the following.) As discussed previously, the fact that and were less likely to be treated with surgery and more likely to a screening test increases the percentage of people diagnosed with be treated with hormones alone than those on the screened arm. early stage cancer and decreases that of late stage cancer (stage The study arms would also tend to differ in their knowledge of the shift) is not equivalent to proof of mortality reduction. Further, disease, which may contribute to an overestimate of the benefits of 9 due to the healthy volunteer effect, case control and cohort studies a screening test. cannot provide definitive evidence of mortality benefit. Prospec- Virtually every screening test is a balance between known tive, randomized clinical trials are required to address this issue. In harms and potential benefits. The most important risk of screening such trials, volunteers are randomized to be screened or not and is the detection and subsequent treatment of a cancer that would are then followed longitudinally to determine if there is a differ- never have come to clinical detection or harmed the patient in ence in disease-specific or overall mortality. the absence of screening (i.e., overdiagnosis and overtreatment). A reduction in mortality rates or in the risk of death is often Treatment can cause emotional and physical morbidity and even 10 stated in terms of relative risk. However, this method of reporting death. Even when screening has a net mortality benefit, there may be misleading. It is preferable to report both the relative and can be considerable harm. For example, in the recent random- absolute reduction in mortality. For example, the European Ran- ized trial of spiral lung computed tomography (CT) scan, approxi- domized Study of Screening for Prostate Cancer (ERSPC) showed mately 27,000 current smokers and former smokers were given that screening reduced the risk of prostate cancer death by 20%. three annual low-dose CT scans. More than 20% had a positive However, this translates into only 1 prostate cancer death averted screening CT scan, necessitating further testing. About 1,000 sub- per 1,000 men screened (5 prostate cancer deaths per 1,000 sequently underwent invasive diagnostic procedures and 16 deaths 11 men not screened versus 4 prostate cancer deaths per 1,000 men were reported within 60 days of the procedure. It is not known screened) and a relatively modest lifetime reduction in the abso- how many of these deaths were directly related to the screening. lute risk of prostate cancer death of only 0.6%, from 3.0% to 2.4%.8 It can be dangerous to extrapolate estimates of benefit from one population to another. In particular, studies showing that a radiographic test is beneficial to average risk individuals may not PROBLEMS WITH RANDOMIZED TRIALS mean that it is beneficial to a population at high risk, and vice versa. For example, women at high risk for breast cancer due to It is important to acknowledge that even prospective, randomized an inherited mutation of a DNA repair gene may be at higher risk trials can have serious methodologic shortcomings. For example, for radiation-induced cancer from mammographies compared imbalances caused by flaws in the randomization scheme can prej- to the general population; a screening test (e.g., spiral lung CT udice the outcome of a trial. Other flaws include so-called drop-in scan), shown to be efficacious in a high-risk population of heavy or contamination, in which some participants on the control arm smokers may result in net harm if applied to a low or average risk get the intervention. Patients on the intervention arm may also population.

tahir99 - UnitedVRG Chapter 34 Cancer Screening 373

SCREENING GUIDELINES AND To date, no study has shown that BSEs decrease mortality. RECOMMENDATIONS BSEs have been studied in two large randomized trials. In one, approximately 266,000 Chinese women were randomized to re- A number of organizations develop cancer screening recommen- ceive intensive BSE instruction with reinforcements and remind- dations or guidelines. These organizations use varying methods. ers compared to a control group receiving no instruction on BSE. The Institute of Medicine (IOM) has released two reports to es- At 10 years of follow-up, there was no difference in mortality, but tablish standards for developing trustworthy clinical practice the intervention arm had a significantly higher incidence of be- guidelines and conducting systematic evidence reviews that nign breast lesions diagnosed and breast biopsies preformed. In serve as their basis.12,13 The U.S. Preventive Services Task Force the second study, 124,000 Russian women were randomized to (USPSTF) and the American Cancer Society (ACS) are two orga- monthly BSEs versus no BSEs. There was no difference in mortal- nizations that issue respected and widely used cancer guidelines ity rates, despite the BSE group having a higher proportion of early (Table 34.3). Both have changed their methods to comply with the stage tumors and a significant increase in the proportion of cancer IOM standards. patients surviving 15 years after diagnosis. The USPSTF is a panel of experts in prevention and evidence- Ultrasonography is primarily used in the diagnostic evaluation based medicine.14 They are primary care providers specializing of a breast mass identified by palpation or mammography. There is in internal medicine, pediatrics, family practice, gynecology and little evidence to support the use of ultrasound as an initial screen- ing test. This modality is highly operator dependent and time obstetrics, nursing, and health behavior. The task force process 19 begins by conducting an extensive structured scientific evidence consuming, with a high rate of FP findings. An MRI is used for review. The task force then develops recommendations for pri- screening women at elevated breast cancer risk due to BRCA1 and mary care clinicians and health-care systems. They adhere to some BRCA2 mutations, Li-Fraumeni syndrome, Cowden disease, or a of the highest standards for recommending a screening test. They very strong family history. MRI is more sensitive but less specific than mammography, leading to a high FP rate and more unnec- are very much concerned with the question, “Does the evidence 20 supporting a screening test demonstrate that the benefits outweigh essary biopsies, especially among young women. The impact of its harms?” MRI breast screening on breast cancer mortality has not yet been The ACS guidelines date back to the 1970s. The current pro- determined. cess for making guidelines involves commissioning academics to Thermography, an infrared imaging technology, has some ad- vocates as a breast cancer screening modality despite a lack of evi- do an independent systematic evidence review. A single general- 21 ist group digests the evidence review, listens to public input, and dence from several small cohort studies. Nipple aspirate cytology and ductal lavage have also been suggested as possible screening writes the guidelines. The ACS panel tries to clearly articulate the 22 benefits, limitations, and harms associated with a screening test.15 methods. Both should be considered experimental at this time.

Effectiveness of Breast Cancer Screening BREAST CANCER Breast cancer screening has been associated with a dramatic rise AND SCREENING CANCER PREVENTION Mammographies, clinical breast examinations (CBE) by a health- in breast cancer incidence. At the same time, there has been a care provider, and breast self-examinations (BSE) have long been dramatic decrease in breast cancer mortality rates. However, in advocated16 for the early detection of breast cancer. In recent the United States and Europe, incidence-by-stage data show a dra- years, ultrasound, magnetic resonance imaging (MRI), and other matic increase in the proportion of early stage cancers without a technologies have been added to the list of proposed screening concomitant decrease in the incidence of regional and metastatic modalities. cancers.23 These findings are at odds with the clinical trials data Mammographic screening was first advocated in the 1950s. and raise questions regarding the extent to which early diagnosis is The Health Insurance Plan (HIP) Study was the first prospective, responsible for declining breast cancer mortality rates. randomized clinical trial to formally assess its value in reducing From 1976 to 2008, the incidence of early-stage breast cancer death from breast cancer. In this study, started in 1963, about for American women aged 40 and older increased from 112 to 61,000 women were randomized to three annual mammograms 234 per 100,000. This is a rise of 122 cases per 100,000, whereas with clinical breast examination versus no screening, which was the absolute decrease in late-stage cancers was only 8 cases per the standard practice at that time. HIP first reported that mammog- 100,000 (from 102 to 94 cases per 100,000). These data raise raphy reduced breast cancer mortality by 30% at about 10 years questions regarding the magnitude of benefit, as well as the po- after study entry. With 18 years of follow-up, those in the screening tential risks, of breast cancer screening. The discrepancy between arm had a 25% lower breast cancer mortality rate.16 the magnitude of the increase of early disease and the decrease of Nine additional prospective randomized studies have been late-stage cancer and cancer mortality suggests that a proportion published. These studies provide the basis for the current consen- of invasive breast cancers diagnosed by screening represents over- sus that screening women 40 to 75 years of age does reduce the diagnosis. These data suggest that overdiagnosis accounts for up to relative risk of breast cancer death by 10% to 25%. The 10 studies 31% of all breast cancers diagnosed by screening.24 Others have demonstrate that the risk–benefit ratio is more favorable for estimated that up to 50% of breast cancers detected by screening women over 50 years of age. Mammography has also been shown mammography are overdiagnosed cancers. In an exhaustive review to be operator dependent, with better performance characteristics of the screening literature, a panel of experts concluded that over- (higher sensitivity and specificity and lower FP rates) reported by diagnosis does exist and estimated it to be 11% to 19% of breast high-volume centers (Table 34.4). cancers diagnosed by screening.25 It is important to note that every one of these studies has some A confounding factor with regard to the mortality benefits of flaws and limitations. They vary in the questions asked and their breast cancer screening is the improvement that has occurred in findings. The Canadian screening trial suggests mammographies breast cancer treatment over this period of time. The effects of and clinical breast examinations do not decrease risk of death for the advances in therapy are supported by cancer modeling studies. woman aged 40 to 49 and that mammographies add nothing to Indeed, the Cancer Intervention and Surveillance Modeling Net- CBEs for women age 50 to 59 years.17 On the other extreme, the work (CISNET), supported by the U.S. National Cancer Institute Kopparberg Sweden study suggests that mammographies are as- (NCI), has estimated that two-thirds of the observed breast cancer sociated with a 32% reduction in the risk of death for women aged mortality reduction is attributable to modern therapy, rather than 40 to 74 years.18 to screening.26 TABLE 34.3 Screening Recommendations for Normal-Risk Asymptomatic Subjects

Cancer Type Test or Procedure American Cancer Society U.S. Preventive Services Task Force Breast Self-examination Women ≥20 years: Breast self-exam is an option “D” Clinical examination Women 20–39 years: Perform every 3 years Women ≥40 years: “I” (as a stand Women ≥40 years: Perform annually alone without mammography) Mammography Women ≥40 years: Screen annually for as long as the Women 40–49 years: The decision woman is in good health should be an individual one, and take patient context/values into account (“C”) Women 50–74 years: Every 2 years (“B”) Women ≥75 years: “I” MRI Women >20% lifetime risk of breast cancer: Screen with “I” MRI plus mammography annually Women 15%–20% lifetime risk of breast cancer: Discuss option of MRI plus mammography annually Women <15% lifetime risk of breast cancer: Do not screen annually with MRI Cervical Pap test (cytology) Women ages 21–29 years: Screen every 3 years Women ages 21–65 years: Screen every 3 years (“A”) Women 30–65 years: Acceptable approach to screen with Women <21 years: “D” cytology every 3 years (see HPV test) Women >65 years, with adequate, normal prior Pap screenings: “D” Women <21 years: No screening Women >65 years: No screening following adequate negative prior screening Women after total hysterectomy for noncancerous Women after total hysterectomy for causes: Do not screen noncancerous causes: “D” HPV test Women <30 years: Do not use HPV testing Women ages 30–65 years: Screen in Women ages 30–65 years: Preferred approach to screen combination with cytology every with HPV and cytology cotesting every 5 years (see 5 years if woman desires to lengthen Pap test) the screening interval (see Pap test) Women >65 years: No screening following adequate (“A”) negative prior screening Women <30 years: “D” Women after total hysterectomy for noncancerous Women >65 years, with adequate, causes: Do not screen normal prior Pap screenings: “D” Women after total hysterectomy for noncancerous causes: “D” Colorectal Sigmoidoscopy Adults ≥50 years: Screen every 5 years Adults 50–75 years: Every 5 years in Note: For all CRC screening tests, stop screening when combination with high-sensitivity benefits are unlikely due to life-limiting comorbidity. fecal occult blood testing (FOBT) every 3 years (“A”)a Adults 76–85 years: “C” Adults ≥85 years: “D” Fecal occult blood Adults ≥50 years: Screen every year with high sensitivity Adults 50–75 years: Annually, for high- testing (FOBT) guaiac based FOBT or fecal immunochemical test (FIT) sensitivity FOBT (“A”) only Adults 76–85 years: “C” Adults ≥85 years: “D” Colonoscopy Adults ≥50 years: Screen every 10 years Adults 50–75 years: every 10 years (“A”) Adults 76–85 years: “C” Adults ≥85 years: “D” Fecal DNA testing Adults ≥50 years: Screen, but interval uncertain “I” Fecal Adults ≥50 years: Screen every year “I” immunochemical testing (FIT) CT colonography Adults ≥50 years: Screen every 5 years “I” Lung Complete skin Men and women, 55–74 years, with ≥30 pack-year “I” (“B” draft recommendation issued examination by smoking history, still smoking or have quit within past for public comment in July 2013) clinician or patient 15 years: Discuss benefits, limitations, and potential harms of screening. Only perform screening in facilities with the right type of CT scanner and with high expertise/specialists. (continued)

tahir99 - UnitedVRG Chapter 34 Cancer Screening 375

TABLE 34.3 Screening Recommendations for Normal-Risk Asymptomatic Subjects (continued)

Cancer Type Test or Procedure American Cancer Society U.S. Preventive Services Task Force Ovary CA-125 There is no sufficiently accurate test proven effective in “D” Transvaginal the early detection of ovarian cancer. For women at high “D” ultrasound risk of ovarian cancer and/or who have unexplained, persistent symptoms, the combination of CA-125 and transvaginal ultrasound with pelvic exam may be offered. Prostate Prostate-specific Starting at age 50, men should talk to a doctor about the Men, all ages: “D” antigen (PSA) pros and cons of testing so they can decide if testing is the right choice for them. If African American or have a father or brother who had prostate cancer before age 65, men should have this talk starting at age 45. How often they are tested will depend on their PSA level. Digital rectal As for PSA; if men decide to be tested, they should have No individual recommendation examination (DRE) the PSA blood test with or without a rectal exam. Skin Complete skin Self-examination monthly; clinical exam as part of routine “I” examination by cancer-related checkup clinician or patient

Note: Summary of the screening procedures recommended for the general population by the American Cancer Society and the U.S. Preventive Services Task Force. These recommendations refer to asymptomatic persons who have no risk factors for the cancer, other than age or gender. a USPSTF lettered recommendations are defined as follows: “A”: The USPSTF recommends the service, because there is high certainty that the net benefit is substantial. “B”: The USPSTF recommends the service, because there is high certainty that the net benefit is moderate or moderate certainty hatt the net benefit is moderate to substantial. “C”: The USPSTF recommends selectively offering or providing this service to individual patients based on professional judgment and patient preferences. There is at least moderate certainty that the net benefit is small. “D”: The USPSTF recommends against the service, because there is moderate or high certainty that the service has no net benefitor that the harms outweigh the benefits. “I”: The USPSTF concludes that the current evidence is insufficient to assess the balance of benefits and harms of the service.

Questions have also been raised regarding the quality of the age 40 to 49 years had a mortality benefit at 18 years of follow-up. CANCER PREVENTION AND SCREENING randomized screening trials that demonstrated the mortality ben- However, to a large extent, the mortality benefit among those aged efits of mammography and clinical breast examination because 45 to 49 years at entry was driven by breast cancers diagnosed after these trials suffered from a variety of design flaws. In some, ran- they reached age 50 years.16 domization methods were suboptimal, others reported varying Mammography, like all screening tests, is more efficient (higher numbers of participants over the years, and still others had substan- PPV) for the detection of disease in populations with higher disease tial contamination (drop-ins). Perhaps more importantly, most tri- prevalence (see Table 34.2). Mammography is, therefore, a better als were started and concluded before the widespread use of more test in women age 50 to 59 years than it is among women age 40 advanced mammographic technology, before the modern era of to 49 years because the risk of breast cancer increases with age. adjuvant therapy, and before the advent of targeted therapy. Mammography is also less optimal in women age 40 to 49 years Although randomized control trials (RCT) remain the gold stan- compared to women 50 to 59 years of age for the following reasons: dard for assessing the benefits of a clinical intervention, they cannot ■ A larger proportion have increased breast density, which can take into account improvements in both treatment and patient aware- obscure lesions (lower sensitivity). ness that occurred over time. For this reason, observational and mod- ■ Younger women are more likely to develop aggressive, fast-growing eling studies can provide important, complementary information. breast cancers that are diagnosed between regular screening visits. One systematic review of 17 published population-based and By definition, these interval cancers are not screen detected.29 cohort studies compared breast cancer mortality in groups of women aged 50 to 69 years who started breast cancer screening The USPSTF meta-analysis of eight large randomized trials at different times. Although these studies are subject to methodo- suggested a 15% relative reduction in mortality (relative risk [RR], logic limitations, only four suggested that breast cancer screening 0.85; 95% confidence interval [CI], 0.75 to 0.96) from mammog- reduced the relative risk of breast cancer mortality by 33% or more raphy screening for women aged 40 to 49 years after 11 to 20 years and five suggested no benefit from screening. The review con- of follow-up. This is equivalent to a needing to invite 1,904 women cluded that breast cancer screening likely reduces the risk of breast to screenings over 10 years to prevent one breast cancer death. cancer death by no more than 10%.27 Studies, however, show that more than half of women aged 40 to Even with these limitations, a systematic review of the data 49 years screened annually over a 10-year period will have an FP sponsored by the USPSTF concluded that regular mammography mammogram necessitating further evaluation, often including bi- reduces breast cancer mortality in women aged 40 to 74 years.28 opsy. In addition, estimates of overdiagnosis in this group range The task force also concluded that the benefits of mammography from 10% to 40% of diagnosed invasive cancers.30 are most significant in women aged 50 to 74 years. In an effort to decrease FP rates, some have suggested screening every 2 years rather than yearly. Comparing biennial with annual screening, the CISNET Model consistently shows that biennial Screening Women Age 40 to 49 screening of women ages 40 to 70 only marginally decreases the number of lives saved while halving the false positive rate.29 Nota- Experts disagree about the utility of screening women in their for- bly, the Swedish two-county trial, which had a planned 24-month ties. In the HIP Randomized Control Trial, women who entered at screening interval (the actual interval was 33 months) reported one 376 Cancer Prevention and Screening

TABLE 34.4 Randomized Controlled Trials

Study Randomization Sample Size Intervention and Age at Entry Follow-up Finding Health Insurance Individual 60,565–60,857 MMG and CBE for 3 years 18 years RR 0.77 (95% Plan, United Age 40–64 years CI: 0.61–0.97) States 1963a,b Malmo, Sweden Individual 42,283 Two-view MMG every 18–24 12 years RR 0.81 (95% 1976c,d months × 5 Age 45–69 years CI: 0.62–1.07) Ostergotland Geographic 38,405–39,034 study Three single-view MMG every 12 years RR 0.82 (95% (County E of cluster 37,145–37,936 control 2 years women, Age 40–50 years CI: 0.64–1.05) Two-County Trial) Every 33 months women, Age Sweden 1977e–g 50–74 Kopparberg Geographic 38,562–39,051 intervention Three single-view MMG every 12 years RR 0.68 (95% (County W of cluster 18,478–18,846 control 2 years women, Age 40–50 CI: 0.52–0.89) Two-County Trial) years Sweden 1977e–g Every 33 months women, Age 50–74 years Edinburgh, United Cluster by 23,266 study Initially, two-view MMG and CBE 10 years RR 0.84 (95% Kingdomh physician 21,904 control Then annual CBE with single-view CI: 0.63–1.12) practice MMG years 3, 5, and 7, Age 45–64 years NBSS-1, Canada Individual 25,214 study (100% Annual two-view MMG and CBE 13 years RR 0.97 (95% 1980i,j screened after entry CBE) for 4–5 years, Age 40–49 years CI: 0.74–1.27) 25,216 control NBSS-2, Canada Individual 19,711 study (100% Annual two-view MMG and CBE 11–16 years RR 1.02 (95% 1980i,j screened after entry CBE) versus CBE, Age 50–59 years (mean CI: 0.78–1.33) 19,694 control 13 years) Stockholm, Cluster by birth 40,318–38,525 intervention Single view MMG every 8 years RR 0.80 (95%) Sweden 1981k date group 28 months × 2 CI: 0.53–1.22) 19,943–20,978 control group Age 40–64 years Gothenberg, Complex 21,650 invited Initial two-view MMG, then single- 12–14 years RR 0.79 (95% Sweden 1982d 29,961 control view MMG every 18 months × 4 CI 0.58–1.08) Single read first three rounds, then In the evaluation double-read, Age 39–59 years phase RR 0.77 (95% CI 0.60–1.00) In follow-up phase Age Triall Individual 160,921 Invited group aged 48 and younger 10.7 years RR 0.83 (95% (53,884 invited; 106,956 not offered annual screening by CI: 0.66–1.04) invited) MMG (double-view first screen, then single mediolateral oblique view thereafter); 68% accepted screening on the first screen an 69% and 70% were reinvited (81% attended at least one screen) Age 39–41 years a Shapiro S, Venet W, Strax P, et al. Ten- to fourteen-year effect of screening on breast cancer mortality. J Natl Cancer Inst 1982;69:349–355. b Shapiro S. Periodic screening for breast cancer: the HIP Randomized Controlled Trial. Health Insurance Plan. J Natl Cancer Inst Monogr 1997:27–30. c Andersson I, Aspegren K, Janzon L, et al. Mammographic screening and mortality from breast cancer: the Malmo mammographic screening trial. BMJ 1988;297:943–948. d Nystrom L, Rutqvist LE, Wall S, et al. Breast cancer screening with mammography: overview of Swedish randomised trials. Lancet 1993;341:973–978. e Tabar L, Fagerberg CJ, Gad A, et al. Reduction in mortality from breast cancer after mass screening with mammography. Randomised trial from the Breast Cancer Screening Working Group of the Swedish National Board of Health and Welfare. Lancet 1985;1:829–832. f Tabar L, Fagerberg G, Duffy SW, Day NE. The Swedish two county trial of mammographic screening for breast cancer: recent results and calculation of benefit. J Epidemiol Community Health 1989;43:107–114. g Tabar L, Fagerberg G, Duffy SW, et al. Update of the Swedish two-county program of mammographic screening for breast cancer. Radiol Clin North Am 1992;30:187–210. h Roberts MM, Alexander FE, Anderson TJ, et al. Edinburgh trial of screening for breast cancer: mortality at seven years. Lancet 1990;335:241–246. i Miller AB, To T, Baines CJ, Wall C. The Canadian National Breast Screening Study-1: breast cancer mortality after 11 to 16 years of follow-up. A randomized screening trial of mammography in women age 40 to 49 years. Ann Intern Med 2002;137:305–312. j Miller AB, Wall C, Baines CJ, et al. Twenty five year follow-up for breast cancer incidence and mortality of the Canadian National Breast Screening Study: randomised screening trial. BMJ 2014;348–366. k Frisell J, Eklund G, Hellstrom L, et al. Randomized study of mammography screening—preliminary report on mortality in the Stockholm trial. Breast Cancer Res Treat 1991;18:49–56. l Moss SM, Cuckle H, Evans A, et al. Effect of mammographic screening from age 40 years on breast cancer mortality at 10 years’ follow-up: a randomised controlled trial. Lancet 2006;368:2053–2060.

tahir99 - UnitedVRG Chapter 34 Cancer Screening 377 of the greatest reductions in breast cancer mortality among the supplemental imaging and biopsies.37 Digital breast tomosynthesis RCTs conducted to date. (DBT) uses x-rays and a digital detector to generate cross-sectional images of the breasts. Data are limited, but compared to mam- Screening Women at High Risk mograms, DBT appears to offer increased sensitivity and a reduc- tion in the recall rates.38 Another potential supplementary imaging modality currently under investigation is three-dimensional (3-D) There is interest in creating risk profiles as a way of reducing the automated breast ultrasound, and having screening ultrasounds inconveniences and harms of screening. It might be possible to performed by technologists rather than radiologists. identify women who are at greater risk of breast cancer and refocus screening efforts on those most likely to benefit. Risk factors for breast cancer include the following: Ductal Carcinoma In Situ ■ Extremely dense breasts on mammography or a first-degree The incidence of noninvasive ductal carcinoma in situ (DCIS) has relative with breast cancer are each associated with at least a increased more than fivefold since 1970 as a direct consequence twofold increase in breast cancer risk 39 ■ of widespread screening mammographies. DCIS is a heteroge- Prior benign breast biopsy, second-degree relatives with breast neous condition with low- and intermediate-grade lesions taking a cancer, or heterogeneously dense breasts each increase risk 1.5- decade or more to progress. Nevertheless, women with this diagno- to twofold sis are uniformly subjected to treatment. A better understanding of ■ Current oral contraceptive use, nulliparity, and age at first birth 31 this entity and an increased ability to predict its biologic behavior 30 years and older increase risk 1- to 1.5-fold. may enable more judicious, personalized treatment of DCIS. Importantly, these are risk factors for breast cancer diagnosis, There is little evidence that the early detection and aggressive not breast cancer mortality. Few studies have assessed the associa- treatment of low- and intermediate-grade DCIS reduces breast tion between these factors and death from breast cancer; however, cancer mortality. The standard of care for all grades of DCIS is reproductive factors and breast density have been shown to have lumpectomy with radiation or mastectomy, followed by tamoxi- limited influence on breast cancer mortality.32,33 fen for 5 years. Interestingly, patterns of care studies indicate that Genetic testing for BRCA1 and BRCA2 mutations and other mastectomy rates are increasing,40 and that women are more often markers of breast cancer risk has identified a group of women at high choosing double mastectomies for the treatment of DCIS.41 Ge- risk for breast cancer. Unfortunately, when to begin and the optimal nomic characterization will hopefully lead to the identification of frequency of screening have not been defined. Mammography is less a subset of noninvasive cancers that can be treated less aggressively sensitive at detecting breast cancers in women carrying BRCA1 and or even observed. BRCA2 mutations, possibly because such cancers occur in younger women in whom mammography is known to be less sensitive. Harms MRI screening may be more sensitive than mammography in women at high risk, but specificity is lower. MRIs are associated The harms and disadvantages of mammography screening include with both an increase in FP and an increase in the detection of overdiagnosis, FP tests, FN tests, and the possibility of radiation- AND SCREENING CANCER PREVENTION smaller cancers, which are more likely to be biologically indolent. induced breast cancer. The impact of MRIs on breast cancer mortality with or without The fact that mammography screening has increased the inci- concomitant use of mammographies has not been evaluated in a dence of localized disease without a significant change in metastatic randomized controlled trial. disease at the time of diagnosis suggests that there is some degree of overdiagnosis. The risk of overdiagnosis is greatest at the first screen- Breast Density ing3 and varies with patient age, tumor type, and grade of disease. FP screening tests lead to substantial inconvenience and anxi- It is well established that mammogram sensitivity is lower in women ety in addition to unnecessary invasive biopsies with their atten- with heterogeneously dense or very dense breasts.29,32 However, at dant complications. In the United States, about 10% of all women this time, there are no clear guidelines regarding whether or how screened for breast cancer are called back for additional testing, and screening algorithms should take breast density into account. less than half of them will be diagnosed with breast cancer.39 The In the American College of Radiology’s Imaging Network risk of a FP mammogram is greater for women under the age of 50.37 (ACRIN)/NCI 666 Trial, breast ultrasound was offered to women FN tests delay diagnosis and provide false reassurance. They with increased mammographic breast density and, if either test was are more common in younger women and in women with dense positive, they were referred for a breast biopsy.34 The radiologists breasts.42,43 Certain histologic subtypes are also more difficult to performing the ultrasounds were not aware of the mammographic see on mammogram. Mucinous and lobular tumors and rapidly findings. Mammography detected 7.6 cancers per 1,000 women growing tumors tend to blend in with normal breast architecture.44 screened; ultrasound increased the cancer detection rate to 11.8 A typical screening mammogram provides approximately 4 mSv per 1,000. However, the PPV for mammography alone was 22.6%, of radiation. It has been estimated that annual mammographies will whereas the PPV for mammography with ultrasound was only 11.2%. cause up to 1 case of breast cancer per 1,000 women screened from It has yet to be determined whether supplemental imaging age 40 to age 80 years. Radiation exposure at younger ages causes reduces breast cancer mortality in women with increased breast a greater risk of breast cancer.45 There is also concern that ionizing density. Although it continues to be strongly advocated by some, radiation from mammographies might disproportionately increase systematic reviews have concluded that the evidence is currently the breast cancer risk for women with certain BRCA1 or BRCA2 insufficient to recommend for or against this approach.35 There mutations, because these genes are related to DNA repair.46 are also a number of barriers to supplemental imaging, includ- ing inconsistent insurance coverage, lack of availability in many Recommendations communities, concerns about cost-effectiveness (particularly with regard to MRI), and the increased FP rate associated with supple- Women at Average Risk mental imaging leading to unnecessary biopsies.36 Newer technologies may improve screening accuracy for The ACS and most other medical groups recommend that average women with dense breasts. Compared to conventional mam- risk women undergo a CBE every 3 years starting at age 20 and mography, full field digital mammography (FFDM) appears to that women 40 years of age and over should undergo CBEs and have less FPs. This could reduce the number of women needing screening mammograms annually. Women should be informed 378 Cancer Prevention and Screening of the benefits, limitations, and harms associated with breast can- mortality.55 This study would subsequently show that stool blood cer screening. A mammography will not detect all breast cancers, testing was associated with a 20% reduction in colon cancer inci- and some breast cancers detected with mammographies may still dence.51 These results were confirmed by two other randomized have a poor prognosis. The harms associated with breast cancer trials.56,57 A reduction in colon cancer–specific mortality persisted screening also include the potential for FP results, causing sub- in the Minnesota trial through 30 years of follow-up. Overall mor- stantial anxiety. When abnormal findings cannot be resolved with tality was not affected. additional imaging, a biopsy is required to rule out the possibility Rehydration increases the sensitivity of FOBT at the expense of breast cancer. A majority of biopsies are benign. Finally, some of lowering specificity.58 Indeed, rehydrated specimens have a very breast cancers detected by a mammography may be biologically high FP rate. Overall, 1% to 5% of FOBTs are positive, but only indolent, meaning they would not have caused a problem or have 2% to 10% of those with a positive FOBT have cancer. been detected in a woman’s lifetime had she not undergone a Fecal immunochemical tests (FIT) are stool tests that do not mammography. react to hemoglobin in dietary products. They appear to have The USPSTF, the American College of Physicians, and the higher sensitivity and specificity for colorectal cancer when com- Canadian Task Force on the Periodic Health Examination rec- pared to nonrehydrated FOBT tests.59 ommend routine screening beginning at age 50 years.30,47,48 For Fecal DNA testing is an emerging modality. These tests look women aged 40 to 49 years of age, these groups advise physicians for DNA sequences specific to colorectal polyps and colorectal to enter into a discussion with the patient. The physician and pa- cancer. They may have increased sensitivity and specificity com- tient should take into account individual risks and concerns before pared to FOBT. Although fecal DNA tests appear to find cancer, deciding to screen.47 the body of evidence on their ability to reduce colorectal cancer An Advisory Committee on Cancer Prevention in the Euro- mortality is limited due to a lack of study. This test has been inter- pean Union recommends that women between the ages of 50 mittently available. and 69 years be offered mammogram screening in an organized Flexible sigmoidoscopies are, of course, limited to an examina- screening program with quality assurance.49 This committee says tion of the rectum and sigmoid colon. A prospective randomized women aged 40 to 49 years should be advised of the potential trial of once-only flexible sigmoidoscopies demonstrated a 23% harms of screening and, if mammographic screening is offered, reduction in colorectal cancer incidence and a 31% reduction it should be performed with strict quality standards and double in colorectal cancer mortality after a median 11.2 years of follow- reading. up.60 In the NCI’s Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial (PLCO), there was a 21% reduction in colorectal Women at High Risk cancer incidence and a 26% reduction in colorectal cancer mor- tality with two sigmoidoscopies done 3 to 5 years apart compared The ACS has issued guidelines for women who were known or with the usual care group after a median follow-up of 11.9 years.53 likely carriers of a BRCA mutation and other rarer high-risk ge- In both studies, there was no effect on proximal lesions (i.e., right netic syndromes, or at high risk for other reasons.50 Annual screen- and transverse colon) due to the limited reach of the scope. It is ing mammographies and MRIs starting at age 30 is recommended estimated that flexible sigmoidoscopies can find 60% to 80% of for women: cancers and polyps found by colonoscopies.61 ■ With a known BRCA mutation In two meta-analyses of five randomized controlled trials of sig- ■ Who are untested but have a first-degree relative with aBRCA moidoscopies, there was an 18% relative reduction in colorectal cancer incidence and a 28% relative reduction in colorectal can- mutation 62,63 ■ Who had been treated with radiation to the chest for Hodgkin cer mortality. Participants ranged in age from 50 to 74 years. disease Follow-up ranged from 6 to 13 years. ■ Who have an approximately 20% to 25% or greater lifetime risk The colonoscopy has become the preferred screening method of breast cancer based on specialized breast cancer risk estima- of many, although there have been no prospective, randomized tion models. trials of colonoscopy screening. A positive FOBT, FIT, fecal DNA test, or sigmoidoscopy warrants a follow-up diagnostic colonoscopy. Perhaps the best support for colonoscopy screening is indirect evi- dence from the Minnesota Colon Cancer Control Study, which COLON CANCER SCREENING required that all participants with a positive stool blood test have di- agnostic imaging of the entire colon. In the Minnesota study, more Colorectal cancer screening with the rigid sigmoidoscope dates than 40% of those screened annually eventually received a colon- back to the late 1960s. The desire to examine the entire colon led oscopy. One can also make the argument that the sigmoidoscopy to the use of a barium enema and the development of fecal occult studies indirectly support the efficacy of colonoscopy screening, blood tests. With the development of fiber optics, flexible sigmoid- although it can be argued that embryologic and epidemiologic evi- oscopies and, later, colonoscopies were employed. Today, fecal dence indicate that the right and left colon are biologically distinct occult blood testing (FOBT), stool DNA testing, flexible sigmoid- and, therefore, the mortality benefits from sigmoidoscopies do not oscopies, colonoscopies, and CT colonographies and, occasion- constitute proof that a colonoscopy would similarly reduce mortal- ally, barium enemas are all used in colorectal cancer screening. ity from proximal colon lesions. MRI colonoscopy is in development. In studies involving repeat colonoscopies by a second physi- Screening examinations of the colon and rectum can find cian, 21% of all adenomas were missed, including 26% of 1 to cancer early, but also find precancerous polyps. Randomized tri- 5 mm adenomas and 2% of adenomas 10 mm or more in length.64 als have demonstrated that endoscopic polypectomies reduce the Other limitations of colonoscopies include the inconvenience of incidence of colorectal cancer by about 20%.51–53 the bowel preparation and the risk of bowel perforation (about 3 FOBT was the first colorectal screening test studied in a pro- out of 1,000 procedures, overall, with nearly all of the risk among spective randomized clinical trial. The Minnesota Colon Cancer patients who undergo colonoscopic polypectomies). The cost of Control Study randomized 46,551 adults to one of three arms: an- the procedure and the limited number of physicians who can do nual FOBTs, biennial screening, or usual care. A rehydrated guaiac the procedure are also of concern. test was used. With 13 years of follow-up, the annual screened arm A CT colonography or virtual colonoscopy allows a physician had a 33% relative reduction in colorectal cancer mortality com- to visually reproduce the endoscopic examination on a computer pared to the usual care group.54 At 18 years of follow-up, the bien- screen. A CT colonography involves the same prep as a colonos- nially screened group had a 21% reduction in colorectal cancer copy, but is less invasive. It might have a higher compliance rate.

tahir99 - UnitedVRG Chapter 34 Cancer Screening 379

In experienced hands, the sensitivity of a CT colonography for the The USPSTF issued colorectal cancer screening guidelines detection of polyps ≥6 mm appears to be comparable to that of in 2008.67 The guidelines were based on a systematic literature a colonoscopy. In a meta-analysis of 30 studies, 2-D and 3-D CT review and decision models. The task force concluded that three colonographies performed equally well.65 screening strategies appear to be equivalent for adults age 50 to The disadvantages of a CT colonography include the fact that 75 years: it requires a colonic prep and a finding on CT requires a follow-up 1. An annual FOBT with a sensitive test diagnostic colonoscopy. The rate of extracolonic findings of un- 2. A flexible sigmoidoscopy every 5 years, with a sensitive FOBT certain significance is high ∼( 15% to 30%), and each one must every 3 years be evaluated, thereby contributing to additional expense and po- 3. A colonoscopy every 10 years tential morbidity. The long-term, cumulative radiation risk of re- peated colonography screenings is also a concern. The task force recommends that patients age 76 to 85 years be evaluated individually for screening. They found “insufficient Current Recommendations evidence” to recommend CT colonographies or fecal DNA testing. Patients at Increased Risk of Colorectal Cancer The ACS, the American College of Gastroenterology, the American Gastroenterological Association, the American Soci- Patients can have higher than average risk of colorectal cancer ety for Gastrointestinal Endoscopy, and the American College of due to familial or hereditary factors and clinical conditions such as Radiology have issued joint colorectal cancer guidelines. These inflammatory bowel disease. These patients technically undergo groups consider FOBT, FIT, rigid and flexible sigmoidoscopies, surveillance and not screening. Nevertheless, there are few clinical colonoscopies, and CT colonographies to all be reasonable screen- studies to guide recommendations. Guidelines have been created ing methodologies. based on professional opinion and an understanding of the biology They recommend the following: (1) Screening modalities be of colorectal cancer (Table 34.5).68 chosen based on personal preference and access, and (2) average risk adults should begin colorectal cancer screening at age 50 years with one of the following options: OTHER CANCERS OF THE 1. Annual high sensitivity FOBT or FIT GASTROINTESTINAL TRACT 2. A flexible sigmoidoscopy every 5 years 3. A colonoscopy every 10 years There are no widely accepted screening guidelines for cancers 4. A double contrast barium enema every 5 years of the esophagus, stomach, pancreas, and liver. However, surveil- 5. A CT colonography every 5 years lance is advocated for some patients at high risk. No test is of unequivocal superiority. Patient preferences should be incorporated into screening in order to increase compliance. The Esophageal Cancer Screening

guidelines also stress that a single screening examination is far from CANCER PREVENTION AND SCREENING optimal and that patients should be in a program of regular screening. Esophageal cancer screening has centered on endoscopic examina- Although some colorectal cancers are diagnosed in persons tions for those at high risk due to chronic, severe gastroesophageal under the age of 50 years, screening persons age 40 to 49 years has reflux disease.69 Some physicians advocate routine endoscopic surveil- low yield.66 The guidelines also state that patients with less than a lance of patients with Barrett esophagus. At this time, there is no evi- 10-year life expectancy should not be screened. dence that such surveillance is effective at reducing cancer mortality.

TABLE 34.5 Colon Cancer Screening Recommendations for People with Familial or Inherited Risk

Familial Risk Category Screening Recommendation First-degree relativea affected with colorectal cancer or an Same as average risk but starting at age 40 years adenomatous polyp at age ≥60 years, or two second-degree relativesb affected with colorectal cancer Two or more first-degree relatives withcolon cancer, or a single Colonoscopy every 5 years, beginning at age 40 years or 10 years first-degree relative withcolon cancer or adenomatous polyps younger than the earliest diagnosis in the family, whichever comes first diagnosed at an age <60 years One second-degree or any third-degree relativeb,c with Same as average risk colorectal cancer Gene carrier or at risk for familial adenomatous polyposisd Sigmoidoscopy annually, beginning at age 10–12 yearse Gene carrier or at risk for HNPCC Colonoscopy, every 1–2 years, beginning at age 20–25 years or 10 years younger than the earliest case in the family, whichever comes first a First-degree relatives include patients, siblings, and children. b Second-degree relatives include grandparents, aunts, and uncles. c Third-degree relatives include great-grandparents and cousins. d Includes the subcategories of familial adenomatous polyposis, Gardner syndrome, some Turcot syndrome families, and attenuated adenomatous polyposis coli (AAPC). e In AAPC, colonoscopy should be used instead of sigmoidoscopy because of the preponderance of proximal colonic adenomas. Colonoscopy screening in AAPC should probably begin in the late teens or early 20s. HNPCC, hereditary nonpolyposis colon cancer. From Winawer S, Fletcher R, Rex D, et al. Colorectal cancer screening and surveillance: clinical guidelines and rationale: update based on new evidence. Gastroenterology 2003;124:544–560, with permission. 380 Cancer Prevention and Screening

Gastric Cancer Screening ability to identify squamous premalignancies and malignancies (from the ectodermal cervix) and glandular dysplasia and adeno- Barium-meal photofluorography, serum pepsinogen, and gastric carcinomas (from the endocervix). It is, however, more sensitive at endoscopy have been proposed as screening methods for the early detecting squamous lesions. detection of gastric cancer. There are no randomized trials evalu- The Pap test was introduced before the advent of the prospec- ating the impact of these modalities on gastric cancer mortality. tive, randomized clinical trial and, therefore, has never been so Indeed, screening with barium-meal photofluorography has been tested. However, a number of observational studies over the past studied in high-risk populations for more than 40 years without 60 years support the effectiveness of this screening test.78,79 Mul- clear evidence of benefit. tiple ecologic studies have shown an inverse correlation between Time-trend analysis and case control studies of gastric endos- the introduction of Pap testing in a given country and reductions copy have suggested a decrease in gastric cancer mortality among in both cervical cancer incidence and mortality.80 Importantly, those at high risk in screened versus unscreened individuals; how- mortality reductions in these studies have been proportional to the ever, a large observational study in a high-risk population failed to intensity of screening. In one series, more than half of women di- demonstrate a benefit.70,71 agnosed with cervical cancer either had never had a Pap test or had Although widespread gastric screenings cannot be advocated, not been screened within 5 years of diagnosis.80 there may be justification for endoscopic screenings of high-risk Cervical cytology has evolved over the years. The original Pap populations. Candidates for screening might include elderly indi- smear used an ectocervical spatula to apply a specimen (“smear”) viduals with atrophic gastritis or pernicious anemia, patients who to glass slides. It later included an endocervical brush. The smear have had partial gastrectomy,72 those with a history of sporadic was fixed, stained, and manually examined under a microscope. adenomas, and patients with familial adenomatous polyposis or That method is still used today, but a liquid-based/thin-layer system hereditary nonpolyposis colon cancer. capable of being analyzed by computer is gaining in popularity.81 Human papillomavirus (HPV) 16 and 18 are the cause of more than 70% of cervical cancers. Thirteen other HPV subtypes are Pancreatic Cancer Screening known to be associated with cervical cancer. With increasing un- derstanding of the role of HPV in cervical disease, interest in de- At this time, there are no data from prospective clinical trials to sup- veloping tests to determine the presence of HPV DNA and RNA port a role for pancreatic cancer screening. Some patients with an has grown. HPV screening can be used along with cytology (cotest- extensive family history have undergone periodic CT scanning of the ing), in response to an abnormal cytologic test (reflexive testing), or abdomen, but this approach has not been shown to reduce pancreatic as a stand-alone test. One advantage of the liquid-based/thin-layer cancer mortality. There is an ongoing search for screening biomark- tests over the older smears is that it makes reflexive testing easier to ers. There is a need to follow large cohorts prospectively after collect- 73 perform. An abnormal cytology screen can be objectively verified ing and storing biologic samples to identify biomarkers of risk. by testing for the presence of the HPV virus without calling the patient back. Liver Cancer Screening HPV testing is especially useful because of its negative predic- tive value. Although a positive test for HPV infection is not diag- Screening for liver cancer or hepatocellular carcinoma (HCC) has nostic of cervical disease, a negative HPV test strongly suggests that focused on very high-risk individuals, such as those with cirrho- the abnormal Pap does not represent a premalignant condition. sis.54 To date, trial results are unreliable due to small study sizes The utility of the HPV test is limited in younger women be- and a lack of randomization. cause one-third or more of women in their 20s have active cervical Serum alpha-fetoprotein (AFP), a fetal-specific glycoprotein anti- infections at any given time. The overwhelming majority of these gen, is an HCC tumor marker used in screening. It is not specific to infections and resultant dysplasia will regress and resolve within HCC because it may be elevated in hepatitis, pregnancy, and some 8 to 24 months. For women over the age of 30, screening for the germ cell tumors. AFP has variable sensitivity and has not been presence of HPV DNA or RNA appears to be superior to cytology tested in any randomized clinical trial with a mortality end point. in identifying women at risk for cervical dysplasia and cancer.82 An In one prospective, 16-year, population-based observational HPV infection in women over the age of 30 is more likely to be study, screening was done on 1,487 Alaska natives with chronic persistent and clinically significant.83 The risk of cervical cancer hepatis B virus (HBV) infection. The survival of those with screen- also increases with age, and most cervical cancer deaths occur in detected HCC was compared with a historical group of clinically women over 50 years of age. diagnosed HCC patients.74 With a target of AFP determination every 6 months, there was a 97% sensitivity and 95% specificity for Cytologic Terminology HCC. Such high sensitivity and specificity have not been found The terminology of the Pap smear has changed over time. The in other studies. It is not known if AFP screening decreases HCC traditional cytologic categories were mild, moderate, and severe mortality.75 dysplasia and carcinoma in situ. Mild correlated with cervical Hepatic ultrasound has been used as an additional method for intraepithelial neoplasia (CIN)1 histology on biopsy; moderate detection of HCC. This procedure is operator dependent with usually indicated CIN2; and severe dysplasia indicated CIN3 or variable sensitivity and specificity. Ultrasound screening is com- carcinoma in situ. monly used in patients with hepatitis and cirrhosis.76,77 There was some subjectivity and some overlap, especially in Interest in CT scanning has grown due to the limitations of the area of mild and moderate dysplasia. The NCI sponsored AFP and ultrasound. CT scans may be a more sensitive test for the development of the Bethesda system in 1988. This system HCC than ultrasound or AFP.75 provides an assessment of the adequacy of the cervical specimen and a way of categorizing and describing the Pap smear find- GYNECOLOGIC CANCER ings. It more effectively and uniformly communicates cytology results from the laboratory to the patient caregiver. The Bethesda system was modified in 1991 and again in 2001.84 Today, more Cervical Cancer Screening than 40 international professional societies have endorsed the Bethesda system. Dr. George Papanicolaou first introduced the Pap smear or Pap The Bethesda system recognizes both squamous and glandular test in the early 1940s. The test was widely adopted based on its cytologic abnormalities.

tahir99 - UnitedVRG Chapter 34 Cancer Screening 381

Squamous cell abnormalities include: increasing the screening interval can reduce overdiagnosis and ex- cessive treatment without decreasing screening efficacy. ■ Atypical squamous cells (ASC), which are categorized as either: In 2012, the ACS, the American Society for Colposcopy and ■ Of undetermined significance (ASC-US) Cervical Pathology (ASCCP), and the American Society for Clini- ■ Cannot exclude high-grade squamous intraepithelial lesions cal Pathology (ASCP) issued joint screening guidelines.93 These (ASC-H) guidelines recommend different surveillance strategies and op- ■ Low-grade squamous intraepithelial lesion (LSIL), which cor- tions based on a woman’s age, screening history, risk factors, and relates with histologic CIN1 choice of screening tests. The following are the recommendations ■ High-grade squamous intraepithelial lesion (HSIL), which cor- for a woman at average risk. relates with histologic CIN2, CIN3, and carcinoma in situ ■ Screening for cervical cancer should begin at 21 years of age. Glandular cell abnormalities (features suggestive of adenocar- Women aged 21 to 29 years should receive cytology screen- cinoma) include: ing (with either conventional cervical cytology smears or ■ Atypical glandular cells (AGC): endocervical, endometrial, or liquid-based cytology) every 3 years. HPV testing should not not otherwise specified be performed in this age group (although it can be used to ■ AGCs, favor neoplastic follow-up a diagnosis of ASC-US). Women under 21 years of ■ Endocervical or not otherwise specified age should not be screened regardless of their age of sexual ■ Endocervical adenocarcinoma in situ (AIS) initiation. ■ Adenocarcinoma ■ For women aged 30 to 65 years, the preferred approach is to be screened every 5 years with both HPV testing and cytology ASCs differ from normal cells but do not meet criteria for LSIL ( cotesting). It is also acceptable to continue screening every or HSIL. A small proportion of ASC-US smears are from CIN1 le- 3 years with cytology alone. sions; a smaller proportion are from CIN2 or 3. LSILs are usually ■ Women should discontinue screening after age 65 years if they due to a transient HPV infection. HSILs are more likely to be due have had three consecutive negative cytology tests or two con- to a persistent HPV infection and are more likely to progress to secutive negative HPV test results within the 10-year period cervical cancer than LSILs. before ceasing screening, with the most recent test occurring The Lower Anogenital Squamous Terminology (LAST) project within the last 5 years. of the College of American Pathology and the American Society ■ Women who have undergone a hysterectomy for noncancerous for Colposcopy and Cervical Pathology has proposed that histo- conditions do not need to undergo cervical cancer screening. logic cervical findings be described using the same terminology as ■ Women, regardless of age, should NOT be screened annually cytologic findings.85 by any screening method. Women under the age of 30 who have not received the HPV ■ Women who have received HPV vaccinations should still be vaccine have a high incidence of HPV infection86 and the highest screened according to the previously listed schedule. prevalence of CIN. However, the overwhelming majority of these HPV infections and associated CIN will spontaneously regress.87,88 Due to the high regression rates, cervical screening and treatment Screening in Low Resource Countries AND SCREENING CANCER PREVENTION in women aged 20 to 24 years appear to have little or no impact on Cytology and HPV testing is not widely available in much of the incidence of invasive cervical cancer. It is estimated that about the world. Cervical cancer remains a leading cause of death in 6% of CIN1 lesions progress to CIN3, and 10% to 20% of CIN3 many of these areas. Visual inspection of the cervix is a low-tech lesions progress to invasive cancer.89 method of screening that is now recognized as having the potential The Atypical Squamous Cells of Undetermined Significance to save thousands of lives per year. A clustered, randomized trial (ASCUS)-LSIL Triage Study (ALTS) evaluated women with in India compared one-time cervical visual inspection and im- abnormal Pap smears.90 The investigators concluded that women mediate colposcopy, biopsy, and/or cryotherapy (where indicated) with ASC-US should be tested for HPV. Those who are HPV versus counseling on cervical cancer deaths in women aged 30 to positive should receive a colposcopy. In addition, because most 59 years. After 7 years of follow-up, the age-standardized rate of women with LSIL or HSIL had an HPV infection, an immedi- death due to cervical cancer was 39.6 per 100,000 person-years ate colposcopy and a biopsy of lesions was recommended.91 HPV in the intervention group versus 56.7 per 100,000 person-years in DNA testing is very sensitive for identifying CIN2 or worse pathol- unscreened controls.94,95 This was the first prospective randomized ogy. Among women 30 to 69 years of age, the sensitivity of the Pap clinical trial to evaluate cervical cancer screening. test with HPV testing was 95% compared with 55% for the Pap test alone.92 Ovarian Cancer Screening Performance Characteristics of Cervical Cytology The sensitivity of cytology varies and is a function of the adequacy Modalities proposed for ovarian cancer screening include the of the cervical specimen. It is also affected by the age of the woman bimanual pelvic examination, serum CA-125 antigen measure- and the experience of the cytologist. The addition of HPV testing ment, and transvaginal ultrasound (TVU). The bimanual pelvic increases the number of women referred for a colposcopy. Not sur- examination is subjective and not very reproducible, but serum prisingly, sensitivity is improved by serial examinations over time CA-125 can be objectively measured. Unfortunately, CA-125 is versus a single screen. neither sensitive nor specific. It is elevated in only about half of women with ovarian cancer and may be elevated in a number of nonmalignant diseases (e.g., diverticulosis, endometriosis, cir- Screening Recommendations rhosis, normal menstruation, pregnancy, uterine fibroids).96–98 TVU has shown poor performance in the detection of ovarian Cervical screening, like other screening tests, is associated with cancer in average and high-risk women.99 There is interest in some degree of overdiagnosis as evidenced by the phenomenon the analysis of serum proteomic patterns, but this should be of spontaneous regression (see previous) and, therefore, poten- considered experimental. 100,101 tial harm from overtreatment, such as cervical incompetence, The combination of CA-125 and TVU has been assessed which may reduce fertility and the ability to carry a pregnancy to in two large, prospective randomized trials. The U.S. trial, the term. Because dysplasia takes years to progress to cervical cancer, Prostate Lung Colorectal and Ovarian trial (PLCO), enrolled 382 Cancer Prevention and Screening

78,216 women of average risk age 55 to 74 years.102,103 Participants an increased number of cancers, an increased proportion of early were randomized to receive annual examinations with CA-125 (at stage cancers, and a larger proportion of screen-diagnosed patients entry and then annually for 5 years) and TVU (at entry and then surviving more than 5 years. annually for 3 years) (n = 39,105), or usual care (n = 39,111). Par- These findings led many to advocate for mass lung cancer ticipants were followed for a maximum of 13 years, with mortality screening, whereas others called for a prospective, randomized from ovarian cancer as the main study outcome. At the conclu- trial with a lung cancer mortality endpoint.113 The Mayo Lung sion of the study, the number of deaths from ovarian cancer was Project (MLP), which began in 1971, was such a trial. More than similar in each group. There were 3.1 ovarian cancer deaths per 9,200 male smokers were enrolled and randomized to either have 10,000 women years in the screened group versus 2.6 deaths per sputum cytology collected and CXRs done every 4 months for 10,000 women years in the control group (RR = 1.18; 95% CI, 6 years or to have these same tests performed annually. 0.82 to 1.71).103 At 13 years of follow-up, there were more early stage cancers The U.K. Collaborative Trial of Ovarian Cancer Screening in the intensively screened arm (n = 99) than in the control arm (UKCTOCS) is a randomized trial assessing the efficacy of CA- (n = 51), but the number of advanced tumors was nearly identi- 125 and TVU in more than 200,000 postmenopausal women. In cal (107 versus 109, respectively).114 Despite an increase in 5-year this trial, CA-125 is being used as a first-line test and TVU as a survival (35% versus 15%) intensive screening was not associated follow-up test using a risk of ovarian cancer algorithm (ROCA).104 with a reduction in lung cancer mortality (3.2 versus 3.0 deaths per The ROCA measures changes in CA-125 over time rather than 1,000 person-years, respectively).115 using a predefined cut point.105 ROCA is believed to improve sen- The impact of screening on cancer incidence persisted through sitivity for smaller tumors without measurably increasing the FP nearly 20 years of follow-up. There were 585 lung cancers diag- rate. A mortality assessment is expected in 2015.106 nosed on the intensive screening arm versus 500 on the control No organization currently recommends screening average arm (p = 0.009) and intensive screening continued to be associ- risk women for ovarian cancer. In 2012, the USPSTF recom- ated with a significant increase in disease-specific survival. How- mended against screening for ovarian cancer, concluding that ever, a concomitant decrease in lung-cancer mortality did not there was “adequate evidence” that (1) annual screening with emerge with long-term follow-up (4.4 lung cancer deaths per TVU and CA-125 does not reduce ovarian cancer mortal- 1,000 person-years in the intensively screened arm versus 3.9 per ity and (2) screening for ovarian cancer can lead to important 1,000 person-years in the control arm).116 This suggests that some harms, mainly surgical interventions in women without ovarian lung cancers diagnosed by screening would not have resulted in cancer. 107 death had they not been detected (i.e., overdiagnosis).116 Two other large, randomized studies of CXR and sputum cytol- Women at High Risk for Ovarian Cancer ogy were conducted in the United States during the same time period. All three studies evaluated different screening schedules Although no study has shown a mortality benefit for ovarian can- rather than screening versus no screening. Paradoxically, a meta- cer screening of high-risk individuals, a National Institutes of analysis of the three studies found that more frequent screening Health (NIH) consensus panel concluded that it was prudent for was associated with an increase (albeit not statistically significant), women with a known hereditary ovarian cancer syndrome, such as rather than a decrease, in lung cancer mortality when compared BRCA1/2 mutations or HNPCC, to have annual rectovaginal pel- with less frequent screening.117 A study conducted in Czechoslo- vic examinations, CA-125 determinations, and TVU until child- vakia in the 1980s also failed to show a reduction in lung cancer bearing is completed or at least until age 35 years, at which time a mortality with CXR screening.118 prophylactic bilateral oophorectomy is recommended.108 More recently, the NCI conducted the PLCO trial at 10 sites across the United States. This was a prospective, randomized trial of nearly 155,000 men and women, aged 55 to 74 years. Participants Endometrial Cancer Screening were randomized to receive annual, single-view, posteroanterior CXRs for 4 years versus routine care. With 13 years of follow-up, no There is insufficient evidence to recommend endometrial cancer significant difference in lung cancer mortality was observed. A total screening either for women at average risk or for those at increased of 1,213 lung cancer deaths occurred on the intervention arm ver- risk due to a history of unopposed estrogen therapy, tamoxifen sus 1,230 in the control group (RR, 0.99; 95% CI, 0.87 to 1.22).119 therapy, late menopause, nulliparity, infertility or failure to ovu- Low-dose computerized tomography (LDCT) is an appealing 109 late, obesity, diabetes, or hypertension. The ACS recommends technology for lung cancer screening. It uses an average of 1.5 mSv that women be informed about the symptoms of endometrial can- of radiation to perform a lung scan in 15 seconds. A conventional cer—in particular, vaginal bleeding and spotting—after the onset CT scan uses 8 mSv of radiation and takes several minutes. The of menopause. Women should be encouraged to immediately re- LDCT image is not as sharp as the conventional image, but sen- port these symptoms to their physician. sitivity and specificity for the detection of lung lesions are similar. As in the early chest radiograph trials, a number of single-arm Women at High Risk for Endometrial Cancer LDCT studies reported a substantial increase in the number of early stage lung cancers diagnosed. These studies also demon- Women with a suspected autosomal-dominant predisposition to strated that 5-year survival rates were increased in screened com- colon cancer (e.g. Lynch syndrome), should consider undergo- pared to unscreened populations. ing an annual endometrial biopsy to evaluate endometrial histol- 110,111 These findings led to the conduct of several randomized tri- ogy, beginning at age 35 years. This is based only on expert als of LDCT for the early detection of lung cancer. The largest, opinion, given the paucity of clinical trial data. Women should be longest, and first to report a mortality end point is the National informed about the potential benefits, harms, and limitations of Lung Screening Trial (NLST). In this trial, approximately 53,000 testing for early endometrial cancer. persons were randomized to receive three annual LDCT scans or single-view posteroanterior CXRs. Eligible participants were cur- rent and former smokers between 55 and 74 years of age at the LUNG CANCER SCREENING time of randomization with at least a 30 pack-year smoking history; former smokers were eligible if they had quit smoking within the Lung cancer screening programs using chest radiographs (CXR) previous 15 years. and sputum cytology began in the late 1940s.112 An evaluation With a median follow-up of 6.5 years, 13% more lung can- of these programs showed that screening led to the diagnosis of cers were diagnosed and a 20% (95% CI, 6.8 to 26.7; p = 0.004)

tahir99 - UnitedVRG Chapter 34 Cancer Screening 383

relative reduction in lung cancer mortality was observed in the The USPSTF guidelines give LDCT a grade B recommen- LDCT arm compared to the CXR arm.11 This corresponds to rates dation, concluding that there is moderate certainty that annual of death from lung cancer of 247 and 309 per 100,000 person- screening for lung cancer with LDCT is of moderate net benefit in years, respectively.11 Another important finding from the NLST asymptomatic persons at high risk for lung cancer based on age, total was a 6.7% (95% CI, 1.2 to 13.6; p = 0.02) decrease in death from cumulative exposure to tobacco smoke, and years since quitting. any cause in the LDCT group. NLST participants were at high risk for developing lung can- cer based on their smoking history. Indeed, 25% of all participant PROSTATE CANCER SCREENING deaths were due to lung cancer. A further analysis of the NLST shows that screening prevented the greatest number of lung can- Hugh Hampton Young first advocated the early detection of pros- cer deaths among participants who were at the highest risk but tate cancer with a careful digital rectal examination (DRE) in prevented very few deaths among those at the lowest risk. These 1903. Screening for prostate cancer with the DRE and serum PSA findings provide empirical support for risk-based screening.120 was first advocated in the mid 1980s and became commonplace LDCT screening is clearly promising, but there are some no- by 1992. PSA screening is directly responsible for prostate cancer table caveats. The risk of a FP finding in the first screen was 21%. becoming the most common nonskin cancer in American men. Overall, after three CT scans, 39.1% of participants had at least one PSA is a glycoprotein produced almost exclusively by the epi- positive screening result. Of those who screened positive, the FP thelial component of the prostate gland. This protein was discov- rate was 96.4% in the LDCT group.11 Positive results require addi- ered in the late 1970s, and a serum test to measure circulating tional workup, which can include conventional CT scans, a needle levels was developed in the early 1980s. Although PSA is prostate biopsy, bronchoscopy, mediastinoscopy, or thoracotomy. These di- specific, it is not prostate cancer specific and may be elevated in agnostic procedures are associated with anxiety, expense, and com- a variety of conditions (e.g., benign prostatic hyperplasia, inflam- plications (e.g., pneumo- or hemothorax after a lung biopsy). In the mation and following trauma to the gland, the presence of prostate LDCT study arm, there were 16 deaths within 60 days of an inva- cancer). sive diagnostic procedure. Of the 16 deaths, 6 ultimately did not The PSA test has been widely advocated for prostate cancer have cancer. Although it is not known whether these deaths were screening because it is objective, easily measured, reproducible, directly caused by the invasive procedure, such findings do give noninvasive, and inexpensive. Although PSA screening increases pause. Although the radiation dose from LDCT is low, the possibil- the detection of potentially curable disease, there is substantial ity that this screening test could cause radiation-induced cancers debate about the overall utility of the test. This is because PSA is at least a theoretical concern. The possibility of this long-term screening introduces substantial lead time and length bias as well phenomenon will have to be assessed in future analyses. as being associated with a high FN and FP rates and having a low The CXR lung screening studies suggested that there is a res- positive predictive value. The prostate cancer conundrum was ervoir of biologically indolent lung cancer and that a percentage best summarized by the distinguished urologist, Willet Whitmore of screen-detected lung cancers represent overdiagnosis. The esti- when he said, “Is cure necessary for those in whom it is possible? Is mated rate of overdiagnosis in the long-term follow-up of the Mayo cure possible for those in whom it is necessary?”126 Lung Study and the other CXR studies was 17 to 18.5%.121 Simi- Observational studies suggest that the problem of prostate can- AND SCREENING CANCER PREVENTION larly, it is estimated that 18.5% of the cancers diagnosed on the cer overdiagnosis precedes the PSA era. In a landmark analysis LDCT arm of the NLST represented overdiagnosis.122 with 20-year follow-up, only a small proportion of 767 men, diag- There are estimates that widespread, high-quality screening nosed with localized prostate cancer in the 1970s and early 1980s has the potential to prevent 12,000 lung cancer deaths per year and followed expectantly, died from prostate cancer: 4% to 7% of in the United States.123 However, the NLST was performed at 33 those with Gleason 2 to 4 tumors, 6% to 11% of those with Gleason centers specifically chosen for their expertise in the screening, di- 5 disease, and 18% to 30% of men with Gleason 6 cancer.127 agnosis, and treatment of lung cancer. It is not known whether the Although obviously present in the pre-PSA era, overdiagnosis in- widespread adoption of LDCT lung cancer screening will result creased substantially after the introduction of PSA screening. This in higher complication rates and a less favorable risk–benefit ratio. is illustrated by an examination of the prostate cancer incidence Although LDCT lung cancer screening should clearly be con- and mortality rates in Washington state and Connecticut. Due to sidered for those at high risk of the disease, those at lower risk are the earlier uptake of PSA screening, the incidence of prostate can- equally likely to suffer the harms associated with screening but less cer in Washington increased to twice that of Connecticut during likely to reap the benefits. the 1990s. However, mortality rates remained similar throughout Following the announcement of the NLST results, the ACS, the decade and, in fact, have remained similar to this day. The the American College of Chest Physicians (AACP), the Ameri- Surveillance, Epidemiology, and End Results (SEER) cancer reg- can Society of Clinical Oncology (ASCO), and the National istries show that, over the last 2 decades, a larger proportion of men Comprehensive Cancer Network (NCCN) recommended that living in western Washington have been diagnosed with prostate clinicians should initiate a discussion about lung cancer screening cancer and definitively treated, without a concomitant reduction with patients who would have qualified for the trial. That is: in prostate cancer mortality compared to that of men living in Connecticut.128 ■ Age 55 to 74 years Additional evidence of the potential for overdiagnosis comes ■ At least a 30 pack-year smoking history from the unexpectedly large number of men diagnosed with pros- ■ Currently smoke or have quit within the past 15 years tate cancer in the Prostate Cancer Prevention Trial (PCPT). The ■ Relatively good health PCPT was a prospective, randomized, placebo-controlled trial of Core elements of this discussion should include the benefits, finasteride for prostate cancer prevention. Men were screened an- uncertainties, and harms associated with screening for lung can- nually during this trial, and those who had not been diagnosed cer with LDCT. Adults who choose to be screened should enter with prostate cancer after 7 years on-study were asked to undergo an organized screening program at an institution with expertise in an end-of-study prostate biopsy. Of 4,692 men on the placebo arm LDCT screening, with access to a multidisciplinary team skilled in whose prostate cancer status had been determined by biopsy or the evaluation, diagnosis, and treatment of abnormal lung lesions. transurethral resection (TURP), 24.4% were diagnosed with pros- If such a program is not available, the risks of harm due to screen- tate cancer. Given that the lifetime risk of prostate cancer mortal- ing may be greater than the benefits.124,125 The guidelines recom- ity in the United States is less than 3%, it is clear that many men mend an annual LDCT screening with the caveat that participants harbor indolent prostate cancer and, therefore, are at risk of being in NLST had only three annual screens. overdiagnosed. 384 Cancer Prevention and Screening

The unexpectantly high rate of positive end-of-study biopsies respectively. Subgroup analyses suggested mortality benefits for in men with PSA levels less than or equal to 4.0 ng/mL provided men with PSA values greater than 10 ng/mL and for those with a more accurate assessment of disease prevalence and thus a intermediate- and high-risk disease.134 more accurate assessment of PSA sensitivity than was previously possible. Of the 2,950 men on the placebo arm of the PCPT with PSA levels consistently less than or equal to 4 ng/mL who The Prospective Randomized Screening Trials underwent end-of-study biopsies, 449 (15.2%) were diagnosed with prostate cancer. Accordingly, a PSA level <4.0 ng/mL is The PLCO Cancer Screening Trial was a multicenter, phase III more likely to be a false negative. Because Sensitivity = True trial conducted in the United States by the NCI. In this trial, nearly Positives / (True Positives + False Negatives), a higher FN rate 77,000 men age 55 to 74 years were randomized to receive annual means a lower sensitivity at any given PSA threshold. This has PSA testing for 6 years or usual care. At 13 years of follow-up, a prompted some to advocate using a lower PSA threshold for nonsignificant increase in cumulative prostate cancer mortality recommending biopsies. However, although lowering the PSA was observed among men randomized to annual screening (RR, threshold from 4.0 to 2.5 ng/mL increases the sensitivity from 1.09; 95% CI, 0.87 to 1.36).135 The most important limitation of 24% to 42.8%, it reduces specificity from 92.7% to an unaccept- this trial was the high rate of PSA testing among men randomized ably low 80%.129 to the control arm. This drop-in or contamination served to reduce In the PCPT, cancer was found on end-of-study biopsies at all the statistical power of the study to detect differences in outcome PSA levels (e.g., including 10% of biopsies in men with PSA levels between the two arms. It has also been argued that, due to the high between 0.6 and 1.0 ng/mL and 6% of biopsies in men with PSA rate of PSA screening on the control arm, PLCO effectively com- levels between 0 and 0.6 were positive), suggesting a continuum pared regular prostate cancer screening to opportunistic screening of prostate cancer risk and no cut point with simultaneously high rather than comparing screening to no screening. sensitivity and high specificity. High-grade disease was also docu- The ERSPC is a multicenter trial initiated in 1991 in the mented at all PSA levels, albeit at an overall frequency of only Netherlands and Belgium; five additional European countries 2.3% of men with PSAs <4 ng/mL.130,131 joined between 1994 and 1998.136,137 The frequency of PSA test- ing was every 4 years in all countries except Sweden, in which it was every 2 years. The study results were initially reported in Does Prostate Cancer Treatment Prevent 2009 and updated in 2012.136,137 Although the overall analysis of Deaths? 182,160 men, aged 50 to 74, did not show a reduction in pros- tate cancer–specific mortality, screening was associated with a In order for screening to work, treatment has to work. The first significant decrease in prostate cancer mortality in the prespeci- prospective, randomized studies showing that any prostate can- fied core age group, 55 to 69 years, which included 162,243 men. cer treatment saves lives were published in the late 1990s. These After a median follow-up of 11 years, a 21% relative reduction studies demonstrated an overall survival benefit for the addition of of prostate cancer death (RR, 0.79; 95% CI, 0.68 to 0.91) was long-term androgen deprivation to radiation therapy in men with observed in this group. In absolute terms, prostate cancer mortal- locally advanced, high-risk prostate cancer.132 ity was reduced from 5 to 4 men per 1,000 screened and 37 men The value of surgery for localized disease was assessed by the had to be diagnosed to avert one prostate cancer death. It remains Scandinavian Prostate Cancer Group 4 study (SPCG-4). In this to be seen whether the benefits of screening will increase with trial, 695 men with clinically localized prostate cancer were pro- continued follow-up. spectively randomized to receive radical prostatectomy (RP) or The recruitment and randomization procedures of the ERSPC watchful waiting (WW). In the expectant management group, differed among countries. Notably, potential participants in Fin- hormonal therapy was given at the time of symptomatic metas- land, Sweden, and Italy were identified from population registries tases. About 60% of those enrolled had low-grade, 23% had mod- and underwent randomization before written informed consent erate-grade, 5% had high-grade tumors, and 12% had tumors of was obtained. In some trials, men on the control arm were not unknown grade. At a median follow-up of 12.8 years, the RP group aware they were in the study. Therefore, men on the intervention had significantly lower overall (RR 0.75; p = 0.007) and prostate arm in these countries were more likely to be cared for at high- cancer–specific mortality (RR 0.62; p = 0.01), with 14.6% of the volume referral centers. This may have contributed to the higher PR group and 20.7% of the WW group having died of prostate can- proportion of men on the screening arm, with clinically localized cer. The number needed to treat or prevent one prostate cancer cancer being treated with RPs.138 death was 15. The survival benefit associated with RP was similar In a separate report on 20,000 men randomized to screening before and after 9 years of follow-up and for men with low and or a control group in Göteborg, Sweden, there was a 40% (95% high-risk disease. However, a subset analysis suggested that the CI, 1.50 to 1.80) risk reduction at 14 years of follow-up.139 They mortality benefit of surgery was limited to men less than 65 years reported 293 (95% CI, 177 to 799) needed to be screened and 12 of age. An important limitation of this trial is that 75% of the study needed to be diagnosed in order to prevent one prostate cancer participants had palpable disease, only 12% had nonpalpable dis- death. Three-fourths of the men in this report and 89% of the pros- ease, and only 5% of the cancers had been screen detected. It is, tate cancer deaths were included in the published ERSPC analy- therefore, difficult to apply these data to the US prostate cancer sis. Given this, these data do not constitute independent evidence population, which is dominated by nonpalpable, screen-detected of the efficacy of prostate cancer screening. disease.133 The other site to report separately was in Finland. A total of In contrast to the SPCG-4, the Prostate Intervention versus 80,144 men were randomized to a screening or usual care arm. Observation Trial (PIVOT) was conducted in the United States At 12 years after randomization, there was no statistical difference during the early PSA era. In this study, 731 men with screen- in risk of prostate cancer death (hazard ratio [HR] = 0.85; 95% detected prostate cancer were randomized to receive RP or WW. CI, 0.69 to 1.04).140 Possible explanations as to why Sweden and Of the participants, 50% had nonpalpable disease and, using es- Finland would have such different outcomes include differences tablished criteria for PSA levels, grade, and tumor stage, 43% of in the frequency of screening (every 2 years versus every 4 years, re- men had low-risk, 36% had intermediate-risk, and 21% had high- spectively) and the higher background rate of death from prostate risk prostate cancer. With a median follow-up of 12 years, during cancer in the control group in the Goteborg cohort. Given that the which time 48.4% (354 of 731) of the study participants had died, mortality data from these two cohorts have been largely included RP was associated with statistically insignificant 2.9% and 2.6% ab- in the ERSPC analyses, they do not provide independent evidence solute reductions in overall and prostate cancer–specific mortality, of the efficacy of prostate cancer screening.

tahir99 - UnitedVRG Chapter 34 Cancer Screening 385

The decline in prostate cancer mortality in the United States with screening was moderate, whereas the quality of the evidence since the introduction of PSA screening 2 decades ago is often for harm was high. They recommended shared decision making offered as evidence supporting a mortality benefit for prostate can- for this group, in whom they have concluded the benefits may cer screening. However, prostate cancer mortality rates have also outweigh the harm. declined in many countries that have not widely adopted screen- ing.141 Thus, it is likely that improvements in treatment have con- tributed, at least in part, to the observed decline in prostate cancer SKIN CANCER SCREENING mortality. Another possible contributing factor may be the World Health Organization (WHO) algorithm for adjudicating cause of Assessments of skin cancer screening have focused on melanoma death. A change occurred just as mortality rates began to go up in end points with very little attention to screening for nonmelanoma the late 1970s, and WHO changed back to the older algorithm skin cancer. A systematic review of skin cancer screening studies in 1991 when prostate cancer mortality began declining in many examining the available evidence through mid 2005 concluded countries.142 All of these factors, including a beneficial effect from that direct evidence of improved health outcomes associated with screening, may be contributing to the declining prostate cancer skin cancer screening is lacking.146 mortality rates in the United States. No randomized, clinical trial of skin cancer screening has been attempted. However, several observational studies have suggested Screening Recommendations that melanoma screening might reduce mortality. For example, a decrease in melanoma mortality did occur after a Scottish cam- paign to promote awareness of the signs of suspicious skin lesions The topic of prostate cancer screening tends to evoke strong emo- and encourage early self-referral. However, uncontrolled, ecologic tional reactions. Although the intuitive appeal of early detection is studies such as this provide a relatively low level of evidence, be- undeniable and screening may save some lives, the magnitude of cause it is not possible to determine whether the observed mortal- the mortality reduction is relatively small, whereas the harms as- ity reduction was due to screening or other factors. sociated with screening can be substantial. Whether the potential More recently, the Skin Cancer Research to Provide Evi- benefits outweigh the known harms is a question that each man dence for Effectiveness of Screening project, or SCREEN project, must answer for himself based on his individual preferences. compared a region of Germany in which intensive skin cancer Several professional organizations in the United States, Europe, screening was performed to areas of Germany without intensive and Canada have recently reviewed the screening data and issued screening. Approximately 360,000 residents of the Schleswig- screening guidelines. All acknowledge that legitimate concerns Holstein region aged 20 years and older participated. They chose remain regarding the risk–benefit ratio of prostate cancer screen- either to be screened by a nondermatologist physician trained in ing. There is also general agreement that prostate cancer screening skin examinations or by a dermatologist. Almost 16,000 biopsies should only be done in the context of fully informed consent and were performed and 585 melanomas were diagnosed. Overall, 1 that men should know that experts do not agree as to whether the in 23 participants had an excisional skin biopsy and 620 persons benefits of screening for this disease outweigh the harms. Most rec-

needed to be screened to detect one melanoma. This screening ef- AND SCREENING CANCER PREVENTION ommend against mass screening in public meeting places, malls, fort led to a 16% and 38% increase in melanoma incidence among churches, etc. men and women, respectively, compared to 2 years earlier. The In 2009, the American Urological Association (AUA) PSA melanoma incidence rate returned to preprogram levels after the Best Practice Statement was published, which stated, “Given the program ended. Of the screen-detected melanomas, 90% were uncertainty that PSA testing results in more benefit than harm, a less than 1 mm thick. Screening was performed in 2003 to 2004, thoughtful and broad approach to PSA is critical. Patients need and melanoma mortality in this region subsequently declined. In to be informed of the risks and the benefits of testing before it is 2008, it was nearly 50% lower in both men and women compared undertaken. The risks of over-detection and over-treatment should to the rest of Germany.147,148 be included in this discussion.”143 In 2010, the ACS updated their guidelines, stating that the balance of benefits and harms related to prostate cancer early de- Recommendations of Experts tection are uncertain and the existing evidence is insufficient to support a recommendation for or against the routine use of PSA Skin cancer screening recommendations are based on expert screening.144 The ACS called for discussion and shared decision opinion, given the absence of a randomized clinical trial data and making within the physician–patient relationship. limited observational studies. The ACS recommends monthly skin The most recent 2012 USPSTF guidelines recommend self-examinations and a yearly clinical skin examination as part of a against the use of PSA screening on the basis that there is routine cancer-related checkup.149 The USPSTF finds insufficient moderate certainty that the harms of PSA testing outweigh the evidence to recommend for or against either routine skin cancer benefits and, on that basis, recommended against PSA-based screening of the general population by primary care providers or screening for all men.14 The task force did acknowledge that counseling patients to perform periodic skin self-examinations. some men will continue to request screening and some physi- The task force does recommend that clinicians “remain alert” for cians will continue to offer it. Like the ACS and AUA, they state skin lesions with malignant features when performing a physical that screening under such circumstances should respect patient examination for other purposes, particularly in high-risk individu- preferences. als. The American Academy of Dermatology recommends that In 2013, the AUA conducted a systematic review of over 300 persons at highest risk (i.e., those with a strong family history of studies. They recommended against screening men younger than melanoma and multiple atypical nevi), perform frequent self-ex- 40 years of age, and against screening average-risk men age 40 amination and seek a professional evaluation of the skin at least to 54 years, most men over 70 years of age, and men with a life once per year.150 expectancy of less than 10 to 15 years. They recommend that High-risk individuals are persons with multiple nevi or atypical screening decisions be individualized for higher risk men ages 40 moles. There is consensus they should be educated about the need to 54 years and men over 70 years of age who are in excellent for frequent surveillance by a trained health-care provider begin- health. They placed primacy on shared decision making versus ning at an early age. In the United States, Australia, and Western physician judgments about the balance of benefits and harms at Europe, Caucasian men age 50 years and over account for nearly the population level.145 Even for men aged 55 to 69 years, the half of all melanoma cases. There is some discussion that mela- AUA concluded that the quality of evidence for benefits associated noma early detection efforts should be focused on this population. 386 Cancer Prevention and Screening

REFERENCES

1. Collen MF, Dales LG, Friedman GD, et al. Multiphasic checkup evaluation 32. Barnett GC, Shah M, Redman K, et al. Risk factors for the incidence of study. 4. Preliminary cost benefit analysis for middle-aged men. Prev Med breast cancer: do they affect survival from the disease? J Clin Oncol 2008;26: 1973;2:236–246. 3310–3316. 2. Prorok PC, Kramer BS, Gohagan JK. Screening theory and study design: 33. Gierach GL, Ichikawa L, Kerlikowske K, et al. Relationship between mam- the basics. In: Kramer B, Prorok P, eds. Cancer Screening. New York: Marcel mographic density and breast cancer death in the Breast Cancer Surveillance Dekker; 1999:29–53. Consortium. J Natl Cancer Inst 2012;104:1218–1227. 3. Welch HG, Black WC. Overdiagnosis in cancer. J Natl Cancer Inst 2010; 34. Berg WA, Blume JD, Cormack JB, et al. Combined screening with ultra- 102:605–613. sound and mammography vs mammography alone in women at elevated risk 4. Yamamoto K, Hayashi Y, Hanada R, et al. Mass screening and age-specific of breast cancer. JAMA 2008;299:2151–2163. incidence of neuroblastoma in Saitama Prefecture, Japan. J Clin Oncol 35. Gartlehner G, Thaler K, Chapman A, et al. Mammography in combination 1995;13:2033–2038. with breast ultrasonography versus mammography for breast cancer screening 5. Woods WG, Gao RN, Shuster JJ, et al. Screening of infants and mortality due in women at average risk. Cochrane Database Syst Rev 2013;4:CD009632. to neuroblastoma. N Engl J Med 2002;346:1041–1046. 36. Tice JA, O’Meara ES, Weaver DL, et al. Benign breast disease, mammo- 6. Friedman GD, Collen MF, Fireman BH. Multiphasic Health Checkup graphic breast density, and the risk of breast cancer. J Natl Cancer Inst Evaluation: a 16-year follow-up. J Chronic Dis 1986;39:453–463. 2013;105:1043–1049. 7. Pinsky PF, Miller A, Kramer BS, et al. Evidence of a healthy volunteer effect 37. Kerlikowske K, Hubbard RA, Miglioretti DL, et al. Comparative effective- in the prostate, lung, colorectal, and ovarian cancer screening trial. Am J ness of digital versus film-screen mammography in community practice in Epidemiol 2007;165:874–881. the United States: a cohort study. Ann Intern Med 2011;155:493–502. 8. Boyle P, Brawley OW. Prostate cancer: current evidence weighs against 38. Haas BM, Kalra V, Geisel J, et al. Comparison of tomosynthesis plus digital population screening. CA Cancer J Clin 2009;59:220–224. mammography and digital mammography alone for breast cancer screening. 9. Autier P, Boyle P, Buyse M, et al. Is FOB screening really the answer for Radiology 2013;269:694–700. lowering mortality in colorectal cancer? Recent Results Cancer Res 2003;163: 39. Rosenberg RD, Yankaskas BC, Abraham LA, et al. Performance benchmarks 254–263. for screening mammography. Radiology 2006;241:55–66. 10. de Boer AG, Taskila T, Ojajarvi A, et al. Cancer survivors and unemployment: 40. Gomez SL, Lichtensztajn D, Kurian AW, et al. Increasing mastectomy rates a meta-analysis and meta-regression. JAMA 2009;301:753–762. for early-stage breast cancer? Population-based trends from California. J Clin 11. Aberle DR, Adams AM, Berg CD, et al. Reduced lung-cancer mortality Oncol 2010;28:e155–e157. with low-dose computed tomographic screening. N Engl J Med 2011;365: 41. Tuttle TM, Jarosek S, Habermann EB, et al. Increasing rates of contralat- 395–409. eral prophylactic mastectomy among patients with ductal carcinoma in situ. 12. Eden J, Levit L, Berg A, et al., eds. Finding What Works in Health Care: J Clin Oncol 2009;27:1362–1367. Standards for Systematic Reviews. Washington DC: The National Academies 42. Rosenberg RD, Hunt WC, Williamson MR, et al. Effects of age, breast Press; 2011. den sity, ethnicity, and estrogen replacement therapy on screening mammo- 13. Graham R, Mancher M, Wolman DM, et al. Medicine CoSfDTCPGIo. graphic sensitivity and cancer stage at diagnosis: review of 183,134 screening Washington, DC: The National Academies; 2011. mammograms in Albuquerque, New Mexico. Radiology 1998;209:511–518. 14. Moyer VA. Screening for prostate cancer: U.S. Preventive Services Task 43. Kerlikowske K, Grady D, Barclay J, et al. Effect of age, breast density, and Force recommendation statement. Ann Intern Med 2012;157:120–134. family history on the sensitivity of first screening mammography. JAMA 15. Brawley O, Byers T, Chen A, et al. New American Cancer Society pro- 1996;276:33–38. cess for creating trustworthy cancer screening guidelines. JAMA 2011;306: 44. Porter PL, El-Bastawissi AY, Mandelson MT, et al. Breast tumor character- 2495–2499. istics as predictors of mammographic detection: comparison of interval- and 16. Shapiro S. Periodic screening for breast cancer: the HIP Randomized Con- screen-detected cancers. J Natl Cancer Inst 1999;91:2020–2028. trolled Trial. Health Insurance Plan. J Natl Cancer Inst Monogr 1997:27–30. 45. Ronckers CM, Erdmann CA, Land CE. Radiation and breast cancer: a re- 17. Miller AB, Wall C, Baines CJ, et al. Twenty five year follow-up for breast view of current evidence. Breast Cancer Res 2005;7:21–32. cancer incidence and mortality of the Canadian National Breast Screening 46. Pijpe A, Andrieu N, Easton DF, et al. Exposure to diagnostic radiation and Study: randomised screening trial. BMJ 2014;348:g366. risk of breast cancer among carriers of BRCA1/2 mutations: retrospective co- 18. Tabar L, Fagerberg G, Duffy SW, et al. Update of the Swedish two-county hort study (GENE-RAD-RISK). BMJ 2012;345:e5660. program of mammographic screening for breast cancer. Radiol Clin North 47. Qaseem A, Snow V, Sherif K, et al. Screening mammography for women 40 Am 1992;30:187–210. to 49 years of age: a clinical practice guideline from the American College of 19. Moy L, Slanetz PJ, Moore R, et al. Specificity of mammography and US Physicians. Ann Intern Med 2007;146:511–515. in the evaluation of a palpable abnormality: retrospective review. Radiology 48. Tonelli M, Connor Gorber S, Joffres M, et al. Recommendations on screening 2002;225:176–181. for breast cancer in average-risk women aged 40-74 years. CMAJ 2011;183: 20. Lord SJ, Lei W, Craft P, et al. A systematic review of the effectiveness of 1991–2001. magnetic resonance imaging (MRI) as an addition to mammography and 49. Recommendations on cancer screening in the European Union. Advisory ultrasound in screening young women at high risk of breast cancer. Eur J Committee on Cancer Prevention. Eur J Cancer 2000;36:1473–1478. Cancer 2007;43:1905–1917. 50. Saslow D, Boetes C, Burke W, et al. American Cancer Society guidelines for 21. Wishart GC, Campisi M, Boswell M, et al. The accuracy of digital infrared breast screening with MRI as an adjunct to mammography. CA Cancer J Clin imaging for breast cancer detection in women undergoing breast biopsy. Eur 2007;57:75–89. J Surg Oncol 2010;36:535–540. 51. Mandel JS, Church TR, Bond JH, et al. The effect of fecal occult-blood 22. Dooley WC, Ljung BM, Veronesi U, et al. Ductal lavage for detection of screening on the incidence of colorectal cancer. N Engl J Med 2000;343: cellular atypia in women at high risk for breast cancer. J Natl Cancer Inst 1603–1607. 2001;93:1624–1632. 52. Nishihara R, Wu K, Lochhead P, et al. Long-term colorectal-cancer in- 23. Autier P, Boniol M, Middleton R, et al. Advanced breast cancer inci- cidence and mortality after lower endoscopy. N Engl J Med 2013;369: dence following population-based mammographic screening. Ann Oncol 1095–1105. 2011;22:1726–1735. 53. Schoen RE, Pinsky PF, Weissfeld JL, et al. Colorectal-cancer incidence and 24. Bleyer A, Welch HG. Effect of three decades of screening mammography on mortality with screening flexible sigmoidoscopy. N Engl J Med 2012;366: breast-cancer incidence. N Engl J Med 2012;367:1998–2005. 2345–2357. 25. Marmot MG, Altman DG, Cameron DA, et al. The benefits and harms of breast 54. Mandel JS, Bond JH, Church TR, et al. Reducing mortality from colorectal cancer screening: an independent review. Br J Cancer 2013;108:2205–2240. cancer by screening for fecal occult blood. Minnesota Colon Cancer Con- 26. Berry DA, Cronin KA, Plevritis SK, et al. Effect of screening and adjuvant trol Study. N Engl J Med 1993;328:1365–1371. therapy on mortality from breast cancer. N Engl J Med 2005;353:1784–1792. 55. Mandel JS, Church TR, Ederer F, et al. Colorectal cancer mortality: 27. Harris R, Yeatts J, Kinsinger L. Breast cancer screening for women ages effectiveness of biennial screening for fecal occult blood. J Natl Cancer Inst 50 to 69 years a systematic review of observational evidence. Prev Med 1999;91:434–437. 2011;53:108–114. 56. Hardcastle JD, Chamberlain JO, Robinson MH, et al. Randomised 28. Nelson HD, Tyne K, Naik A, et al. Screening for breast cancer: an update for controlled trial of faecal-occult-blood screening for colorectal cancer. Lancet the U.S. Preventive Services Task Force. Ann Intern Med 2009;151:727–737. 1996;348:1472–1477. 29. Mandelblatt JS, Cronin KA, Bailey S, et al. Effects of mammography screen- 57. Kronborg O, Fenger C, Olsen J, et al. Randomised study of screening for ing under different screening schedules: model estimates of potential ben- colorectal cancer with faecal-occult-blood test. Lancet 1996;348:1467–1471. efits and harms. Ann Intern Med 2009;151:738–747. 58. Ahlquist DA, Wieand HS, Moertel CG, et al. Accuracy of fecal occult blood 30. U.S. Preventive Services Task Force. Screening for breast cancer: U.S. Pre- screening for colorectal neoplasia. A prospective study using Hemoccult and ventive Services Task Force recommendation statement. Ann Intern Med HemoQuant tests. JAMA 1993;269:1262–1267. 2009;151:716–726. 59. Levin B, Brooks D, Smith RA, et al. Emerging technologies in screening 31. Nelson HD, Zakher B, Cantor A, et al. Risk factors for breast cancer for for colorectal cancer: CT colonography, immunochemical fecal occult women aged 40 to 49 years: a systematic review and meta-analysis. Ann Intern blood tests, and stool screening using molecular markers. CA Cancer J Clin Med 2012;156:635–648. 2003;53:44–55.

tahir99 - UnitedVRG Chapter 34 Cancer Screening 387

60. Atkin WS, Edwards R, Kralj-Hans I, et al. Once-only flexible sigmoidoscopy 90. Cox JT, Schiffman M, Solomon D. Prospective follow-up suggests similar screening in prevention of colorectal cancer: a multicentre randomised con- risk of subsequent cervical intraepithelial neoplasia grade 2 or 3 among trolled trial. Lancet 2010;375:1624–1633. women with cervical intraepithelial neoplasia grade 1 or negative colposcopy 61. Levin TR. Flexible sigmoidoscopy for colorectal cancer screening: valid ap- and directed biopsy. Am J Obstet Gynecol 2003;188:1406–1412. proach or short-sighted? Gastroenterol Clin North Am 2002;31:1015–1029. 91. Guido R, Schiffman M, Solomon D, et al. Postcolposcopy management 62. Littlejohn C, Hilton S, Macfarlane GJ, et al. Systematic review and meta- strategies for women referred with low-grade squamous intraepithelial le- analysis of the evidence for flexible sigmoidoscopy as a screening method for sions or human papillomavirus DNA-positive atypical squamous cells of un- the prevention of colorectal cancer. Br J Surg 2012;99:1488–1500. determined significance: a two-year prospective study. Am J Obstet Gynecol 63. Elmunzer BJ, Hayward RA, Schoenfeld PS, et al. Effect of flexible sigmoid- 2003;188:1401–1405. oscopy-based screening on incidence and mortality of colorectal cancer: a 92. Mayrand MH, Duarte-Franco E, Rodrigues I, et al. Human papillomavirus systematic review and meta-analysis of randomized controlled trials. PLoS DNA versus Papanicolaou screening tests for cervical cancer. N Engl J Med Med 2012;9:e1001352. 2007;357:1579–1588. 64. van Rijn JC, Reitsma JB, Stoker J, et al. Polyp miss rate determined by tandem 93. Saslow D, Solomon D, Lawson HW, et al. American Cancer Society, Ameri- colonoscopy: a systematic review. Am J Gastroenterol 2006;101:343–350. can Society for Colposcopy and Cervical Pathology, and American Society 65. Rosman AS, Korsten MA. Meta-analysis comparing CT colonography, air for Clinical Pathology screening guidelines for the prevention and early de- contrast barium enema, and colonoscopy. Am J Med 2007;120:203–210. tection of cervical cancer. CA Cancer J Clin 2012;62:147–172. 66. Imperiale TF, Wagner DR, Lin CY, et al. Using risk for advanced proximal 94. Sankaranarayanan R, Esmy PO, Rajkumar R, et al. Effect of visual screening colonic neoplasia to tailor endoscopic screening for colorectal cancer. Ann on cervical cancer incidence and mortality in Tamil Nadu, India: a cluster- Intern Med 2003;139:959–965. randomised trial. Lancet 2007;370:398–406. 67. U.S. Preventive Services Task Force. Screening for colorectal cancer: U.S. 95. Szarewski A. Cervical screening by visual inspection with acetic acid. Lancet Preventive Services Task Force recommendation statement. Ann Intern Med 2007;370:365–366. 2008;149:627–637. 96. Johnson CC, Kessel B, Riley TL, et al. The epidemiology of CA-125 in women 68. Winawer S, Fletcher R, Rex D, et al. Colorectal cancer screening and sur- without evidence of ovarian cancer in the Prostate, Lung, Colorectal and veillance: clinical guidelines and rationale-Update based on new evidence. Ovarian Cancer (PLCO) Screening Trial. Gynecol Oncol 2008;110:383–389. Gastroenterology 2003;124:544–560. 97. Duffy MJ, Bonfrer JM, Kulpa J, et al. CA125 in ovarian cancer: European 69. Quintero E, Castells A, Bujanda L, et al. Colonoscopy versus fecal immu- Group on Tumor Markers guidelines for clinical use. Int J Gynecol Cancer nochemical testing in colorectal-cancer screening. N Engl J Med 2012;366: 2005;15:679–691. 697–706. 98. Moss EL, Hollingworth J, Reynolds TM. The role of CA125 in clinical prac- 70. Murakami R, Tsukuma H, Ubukata T, et al. Estimation of validity of mass screen- tice. J Clin Pathol 2005;58:308–312. ing program for gastric cancer in Osaka, Japan. Cancer 1990;65:1255–1260. 99. Fishman DA, Cohen L, Blank SV, et al. The role of ultrasound evaluation 71. Kampschoer GH, Fujii A, Masuda Y. Gastric cancer detected by mass sur- in the detection of early-stage epithelial ovarian cancer. Am J Obstet Gynecol vey. Comparison between mass survey and outpatient detection. Scand J 2005;192:1214–1221. Gastroenterol 1989;24:813–817. 100. Kobayashi E, Ueda Y, Matsuzaki S, et al. Biomarkers for screening, diag- 72. Stael von Holstein C, Eriksson S, Huldt B, et al. Endoscopic screening dur- nosis, and monitoring of ovarian cancer. Cancer Epidemiol Biomarkers Prev ing 17 years for gastric stump carcinoma. A prospective clinical trial. Scand J 2012;21:1902–1912. Gastroenterol 1991;26:1020–1026. 101. Ren J, Cai H, Li Y, et al. Tumor markers for early detection of ovarian cancer. 73. Shaukat A, Mongin SJ, Geisser MS, et al. Long-term mortality after screen- Expert Rev Mol Diagn 2010;10:787–798. ing for colorectal cancer. N Engl J Med 2013;369:1106–1114. 102. Prorok PC, Andriole GL, Bresalier RS, et al. Design of the Prostate, Lung, 74. McMahon BJ, Bulkow L, Harpster A, et al. Screening for hepatocellular Colorectal and Ovarian (PLCO) Cancer Screening Trial. Control Clin Trials carcinoma in Alaska natives infected with chronic hepatitis B: a 16-year 2000;21:273S–309S. population-based study. Hepatology 2000;32:842–846. 103. Buys SS, Partridge E, Black A, et al. Effect of screening on ovarian cancer

75. Chalasani N, Horlander JC Sr, Said A, et al. Screening for hepatocellular mortality: the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer AND SCREENING CANCER PREVENTION carcinoma in patients with advanced cirrhosis. Am J Gastroenterol 1999;94: Screening Randomized Controlled Trial. JAMA 2011;305:2295–2303. 2988–2993. 104. Menon U, Gentry-Maharaj A, Hallett R, et al. Sensitivity and specificity of 76. Sherman M, Peltekian KM, Lee C. Screening for hepatocellular carcinoma multimodal and ultrasound screening for ovarian cancer, and stage distribu- in chronic carriers of hepatitis B virus: incidence and prevalence of hepa- tion of detected cancers: results of the prevalence screen of the UK Col- tocellular carcinoma in a North American urban population. Hepatology laborative Trial of Ovarian Cancer Screening (UKCTOCS). Lancet Oncol 1995;22:432–438. 2009;10:327–340. 77. Dodd GD 3rd, Miller WJ, Baron RL, et al. Detection of malignant tumors 105. Drescher CW, Shah C, Thorpe J, et al. Longitudinal screening algorithm in end-stage cirrhotic livers: efficacy of sonography as a screening technique. that incorporates change over time in CA125 levels identifies ovarian cancer AJR Am J Roentgenol 1992;159:727–733. earlier than a single-threshold rule. J Clin Oncol 2013;31:387–392. 78. Laara E, Day NE, Hakama M. Trends in mortality from cervical cancer 106. Sharma A, Apostolidou S, Burnell M, et al. Risk of epithelial ovarian cancer in the Nordic countries: association with organised screening programmes. in asymptomatic women with ultrasound-detected ovarian masses: a prospec- Lancet 1987;1:1247–1249. tive cohort study within the UK collaborative trial of ovarian cancer screen- 79. Christopherson WM, Lundin FE Jr, Mendez WM, et al. Cervical cancer ing (UKCTOCS). Ultrasound Obstet Gynecol 2012;40:338–344. control: a study of morbidity and mortality trends over a twenty-one-year pe- 107. Moyer VA. Screening for ovarian cancer: U.S. Preventive Services Task Force riod. Cancer 1976;38:1357–1366. reaffirmation recommendation statement.Ann Intern Med 2012;157:900–904. 80. Janerich DT, Hadjimichael O, Schwartz PE, et al. The screening histories 108. NIH consensus conference. Ovarian cancer. Screening, treatment, and of women with invasive cervical cancer, Connecticut. Am J Public Health follow-up. NIH Consensus Development Panel on Ovarian Cancer. JAMA 1995;85:791–794. 1995;273:491–497. 81. Sawaya GF, McConnell KJ, Kulasingam SL, et al. Risk of cervical cancer 109. Smith RA, von Eschenbach AC, Wender R, et al. American Cancer Society associated with extending the interval between cervical-cancer screenings. guidelines for the early detection of cancer: update of early detection guide- N Engl J Med 2003;349:1501–1509. lines for prostate, colorectal, and endometrial cancers. Also: update 2001— 82. Sankaranarayanan R, Nene BM, Shastri SS, et al. HPV screening for cervical testing for early lung cancer detection. CA Cancer J Clin 2001;51:38–75. cancer in rural India. N Engl J Med 2009;360:1385–1394. 110. Burke W, Petersen G, Lynch P, et al. Recommendations for follow-up care 83. Vesco KK, Whitlock EP, Eder M, et al. In: Screening for Cervical Cancer: of individuals with an inherited predisposition to cancer. I. Hereditary non- A Systematic Evidence Review for the US Preventive Services Task Force. polyposis colon cancer. Cancer Genetics Studies Consortium. JAMA 1997; Rockville, MD: Agency for Healthcare Research and Quality; 2011. 277:915–919. 84. Solomon D, Davey D, Kurman R, et al. The 2001 Bethesda System: termi- 111. Gull B, Karlsson B, Milsom I, et al. Can ultrasound replace dilation and nology for reporting results of cervical cytology. JAMA 2002;287:2114–2119. curettage? A longitudinal evaluation of postmenopausal bleeding and trans- 85. Darragh TM, Colgan TJ, Thomas Cox J, et al. The Lower Anogenital vaginal sonographic measurement of the endometrium as predictors of endo- Squamous Terminology Standardization project for HPV-associated lesions: metrial cancer. Am J Obstet Gynecol 2003;188:401–408. background and consensus recommendations from the College of American 112. Scamman CL. Follow-up study of lung cancer suspects in a mass chest X-ray Pathologists and the American Society for Colposcopy and Cervical Pathol- survey. N Engl J Med 1951;244:541–544. ogy. Int J Gynecol Pathol 2013;32:76–115. 113. Croswell JM, Ransohoff DF, Kramer BS. Principles of cancer screening: les- 86. Ho GY, Bierman R, Beardsley L, et al. Natural history of cervicovaginal pap- sons from history and study design issues. Semin Oncol 2010;37:202–215. illomavirus infection in young women. N Engl J Med 1998;338:423–428. 114. Fontana RS, Sanderson DR, Taylor WF, et al. Early lung cancer detection: 87. Holowaty P, Miller AB, Rohan T, et al. Natural history of dysplasia of the results of the initial (prevalence) radiologic and cytologic screening in the uterine cervix. J Natl Cancer Inst 1999;91:252–258. Mayo Clinic study. Am Rev Respir Dis 1984;130:561–565. 88. Richardson H, Kelsall G, Tellier P, et al. The natural history of type-specific 115. Fontana RS, Sanderson DR, Woolner LB, et al. Screening for lung cancer. human papillomavirus infections in female university students. Cancer Epi- A critique of the Mayo Lung Project. Cancer 1991;67:1155–1164. demiol Biomarkers Prev 2003;12:485–490. 116. Marcus PM, Bergstralh EJ, Zweig MH, et al. Extended lung cancer inci- 89. Melnikow J, Nuovo J, Willan AR, et al. Natural history of cervical squamous dence follow-up in the Mayo Lung Project and overdiagnosis. J Natl Cancer intraepithelial lesions: a meta-analysis. Obstet Gynecol 1998;92:727–735. Inst 2006;98:748–756. 388 Cancer Prevention and Screening

117. Manser R, Wright G, Hart D, et al. Surgery for early stage non-small cell lung 134. Wilt TJ, Brawer MK, Jones KM, et al. Radical prostatectomy versus observa- cancer. Cochrane Database Syst Rev 2005:CD004699. tion for localized prostate cancer. N Engl J Med 2012;367:203–213. 118. Kubik A, Parkin DM, Khlat M, et al. Lack of benefit from semi-annual 135. Andriole GL, Crawford ED, Grubb RL 3rd, et al. Prostate cancer screening screening for cancer of the lung: follow-up report of a randomized controlled in the randomized Prostate, Lung, Colorectal, and Ovarian Cancer Screen- trial on a population of high-risk males in Czechoslovakia. Int J Cancer 1990; ing Trial: mortality results after 13 years of follow-up. J Natl Cancer Inst 45:26–33. 2012;104:125–132. 119. Oken MM, Hocking WG, Kvale PA, et al. Screening by chest radiograph and 136. Schroder FH, Hugosson J, Roobol MJ, et al. Screening and prostate-cancer lung cancer mortality: the Prostate, Lung, Colorectal, and Ovarian (PLCO) mortality in a randomized European study. N Engl J Med 2009;360:1320–1328. randomized trial. JAMA 2011;306:1865–1873. 137. Schroder FH, Hugosson J, Roobol MJ, et al. Prostate-cancer mortality at 11 120. Kovalchik SA, Tammemagi M, Berg CD, et al. Targeting of low-dose CT years of follow-up. N Engl J Med 2012;366:981–990. screening according to the risk of lung-cancer death. N Engl J Med 2013; 138. Wolters T, Roobol MJ, Steyerberg EW, et al. The effect of study arm on 369:245–254. prostate cancer treatment in the large screening trial ERSPC. Int J Cancer 121. Kubik AK, Parkin DM, Zatloukal P. Czech Study on Lung Cancer Screen- 2010;126:2387–2393. ing: post-trial follow-up of lung cancer deaths up to year 15 since enrollment. 139. Hugosson J, Carlsson S, Aus G, et al. Mortality results from the Goteborg Cancer 2000;89:2363–2368. randomised population-based prostate-cancer screening trial. Lancet Oncol 122. Patz EF Jr, Pinsky P, Gatsonis C, et al. Overdiagnosis in low-dose computed 2010;11:725–732. tomography screening for lung cancer. JAMA Intern Med 2014;174:269–274. 140. Kilpelainen TP, Tammela TL, Malila N, et al. Prostate cancer mortal- 123. Ma J, Ward EM, Smith R, et al. Annual number of lung cancer deaths ity in the Finnish randomized screening trial. J Natl Cancer Inst 2013;105: potentially avertable by screening in the United States. Cancer 2013;119: 719–725. 1381–1385. 141. Center MM, Jemal A, Lortet-Tieulent J, et al. International variation in pros- 124. Wender R, Fontham ET, Barrera E Jr, et al. American Cancer Society lung tate cancer incidence and mortality rates. Eur Urol 2012;61:1079–1092. cancer screening guidelines. CA Cancer J Clin 2013;63:107–117. 142. Boyle P. Screening for prostate cancer: have you had your cholesterol mea- 125. Bach PB, Mirkin JN, Oliver TK, et al. Benefits and harms of CT screening sured? BJU Int 2003;92:191–199. for lung cancer: a systematic review. JAMA 2012;307:2418–2429. 143. Greene KL, Albertsen PC, Babaian RJ, et al. Prostate specific antigen best 126. Montie JE, Smith JA. Whitmoreisms: memorable quotes from Willet F. practice statement: 2009 update. J Urol 2009;182:2232–2241. Whitmore, Jr, M.D. Urology 2004;63:207–209. 144. Wolf AM, Wender RC, Etzioni RB, et al. American Cancer Society guideline 127. Albertsen PC, Hanley JA, Fine J. 20-year outcomes following conservative man- for the early detection of prostate cancer: update 2010. CA Cancer J Clin agement of clinically localized prostate cancer. JAMA 2005;293:2095–2101. 2010;60:70–98. 128. Lu-Yao G, Albertsen PC, Stanford JL, et al. Screening, treatment, and pros- 145. Carter HB. American Urological Association (AUA) guideline on prostate tate cancer mortality in the Seattle area and Connecticut: fifteen-year follow- cancer detection: process and rationale. BJU Int 2013;112:543–547. up. J Gen Intern Med 2008;23:1809–1814. 146. Wolff T, Tai E, Miller T. Screening for skin cancer: an update of the evi- 129. Thompson IM, Chi C, Ankerst DP, et al. Effect of finasteride on the sensitivity dence for the U.S. Preventive Services Task Force. Ann Intern Med 2009;150: of PSA for detecting prostate cancer. J Natl Cancer Inst 2006;98:1128–1133. 194–198. 130. Thompson IM, Pauler DK, Goodman PJ, et al. Prevalence of prostate cancer 147. Katalinic A, Waldmann A, Weinstock MA, et al. Does skin cancer screening among men with a prostate-specific antigen level < or =4.0 ng per milliliter. save lives?: an observational study comparing trends in melanoma mortality N Engl J Med 2004;350:2239–2246. in regions with and without screening. Cancer 2012;118:5395–5402. 131. Thompson IM, Ankerst DP, Chi C, et al. Operating characteristics of pros- 148. Breitbart EW, Waldmann A, Nolte S, et al. Systematic skin cancer screening tate-specific antigen in men with an initial PSA level of 3.0 ng/ml or lower. in Northern Germany. J Am Acad Dermatol 2012;66:201–211. JAMA 2005;294:66–70. 149. Smith RA, Brooks D, Cokkinides V, et al. Cancer screening in the United 132. Widmark A, Klepp O, Solberg A, et al. Endocrine treatment, with or with- States, 2013: a review of current American Cancer Society guidelines, cur- out radiotherapy, in locally advanced prostate cancer (SPCG-7/SFUO-3): an rent issues in cancer screening, and new guidance on cervical cancer screen- open randomised phase III trial. Lancet 2009;373:301–308. ing and lung cancer screening. CA Cancer J Clin 2013;63:88–105. 133. Bill-Axelson A, Holmberg L, Ruutu M, et al. Radical prostatectomy ver- 150. U.S. Preventive Services Task Force. Screening for skin cancer: U.S. Preventive sus watchful waiting in early prostate cancer. N Engl J Med 2011;364: Services Task Force recommendation statement. Ann Intern Med 2009;150: 1708–1717. 188–193.

tahir99 - UnitedVRG 35 Genetic Counseling

Ellen T. Matloff and Danielle C. Bonadies

INTRODUCTION patient’s age, childbearing status, menopausal status, risk category, ease of screening, and personal preferences and will likely change Clinically based genetic testing has evolved from an uncommon over time with the patient. The ultimate goal of cancer genetic analysis ordered for the rare hereditary cancer family to a widely counseling is to help the patient reach the decision best suited to available tool ordered on a routine basis to assist in surgical and her personal situation, needs, and circumstances. radiation decision making, chemoprevention, and surveillance There are now a significant number of referral centers across of the patient with cancer, as well as management of the entire the country specializing in cancer genetic counseling, and the family. The evolution of this field has created a need for accurate numbers are growing. However, some experts insist that the only cancer genetic counseling and risk assessment. Extensive cover- way to keep up with the overwhelming demand for counseling age of this topic by the media, including Angelina Jolie’s public will be to educate more physicians and nurses in cancer genetics. disclosure of her BRCA1+ status in May 2013, and widespread The feasibility of adding another specialized and time-consuming advertising by commercial testing laboratories have further fueled task to the clinical burden of these professionals is questionable, the demand for counseling and testing. particularly with average patient encounters of 19.5 and 21.6 min- Cancer genetic counseling is a communication process be- utes for general practitioners and gynecologists, respectively.7,8 A tween a health-care professional and an individual concerning more practical goal is to better educate clinicians in the area of cancer occurrence and risk in his or her family.1 The process, risk assessment so that they can screen their patient populations which may include the entire family through a blend of genetic, for individuals at high risk for hereditary cancer and refer them medical, and psychosocial assessments and interventions, has been on to comprehensive counseling and testing programs. Access to described as a bridge between the fields of traditional oncology genetic counseling is no longer an issue because there are now and genetic counseling.1 internet, phone, and satellite-based telemedicine services available (Table 35.1), with most major health insurance companies now The goals of this process include providing the client with an AND SCREENING CANCER PREVENTION assessment of individual cancer risk, while offering the emotional covering these services9–11 and several requiring them.12 support needed to understand and cope with this information. It also involves deciphering whether the cancers in a family are likely to be caused by a mutation in a cancer gene and, if so, which one. WHO IS A CANDIDATE FOR CANCER There are >30 hereditary cancer syndromes, many of which can GENETIC COUNSELING? be caused by mutations in different genes. Therefore, testing for these syndromes can be complicated. Advertisements by genetic Only 5% to 10% of most cancer is thought to be caused by single testing companies bill genetic testing as a simple process that can mutations within autosomal-dominant inherited cancer suscepti- be carried out by health-care professionals with no training in bility genes.13 The key for clinicians is to determine which patients this area; however, there are many genes involved in cancer, the are at greatest risk to carry a hereditary mutation. There are seven interpretation of the test results is often complicated, the risk of critical risk factors in hereditary cancer (Table 35.2). The first is result misinterpretation is great and associated with potential li- early age of cancer onset. This risk factor, even in the absence of a ability, and the emotional and psychological ramifications for the family history, has been shown to be associated with an increased patient and family can be powerful.2,3 A few hours of training by a frequency of germline mutations in many types of cancers.14 The company generating a profit from the sale of these tests does not second risk factor is the presence of the same cancer in multiple adequately prepare providers to offer their own genetic counsel- affected relatives on the same side of the pedigree. These cancers ing and testing services.4 Furthermore, the delegation of genetic do not need to be of similar histologic type in order to be caused testing responsibilities to office staff and, recently, mammography by a single mutation. The third risk factor is the clustering of can- technicians, is alarming and likely presents a huge liability for cers known to be caused by a single gene mutation in one family these ordering physicians, their practices, and their institutions.5,6 (e.g., breast/ovarian/pancreatic cancer or colon/uterine/ovarian Providers should proceed with caution before taking on the role of cancers). The fourth risk factor is the occurrence of multiple pri- primary genetic counselor for their patients. mary cancers in one individual. This includes multiple primary Counseling about hereditary cancers differs from traditional breast or colon cancers as well as a single individual with separate genetic counseling in several ways. Clients seeking cancer ge- cancers known to be caused by a single gene mutation (e.g., breast netic counseling are rarely concerned with reproductive decisions, and ovarian cancer in a single individual). Ethnicity also plays a which are often the primary focus in traditional genetic counseling, role in determining who is at greatest risk to carry a hereditary can- but are instead seeking information about their own and other rela- cer mutation. Individuals of Jewish ancestry are at increased risk tives’ chances of developing cancer.1 Additionally, the risks given to carry three specificBRCA1/2 mutations.15 The presence of a are not absolute but change over time as the family and personal cancer that presents unusually—in this case, breast cancer in a history changes and the patient ages. The risk reduction options male—represents a sixth risk factor and is important even when available are often radical (e.g., chemoprevention or prophylactic it is the only risk factor present. Finally, the last risk factor is pa- surgery), and are not appropriate for every patient at every age. thology. Certain types of cancer are overrepresented in hereditary The surveillance and management plan must be tailored to the cancer families. For example, medullary and triple negative breast

389 390 Cancer Prevention and Screening

TABLE 35.1 carry a mutation in a different gene. It is already well-established that medullary thyroid carcinoma, sebaceous adenoma or carci- How to Find a Genetic Counselor for Your Patient noma, adrenocortical carcinoma before the age of 25 years, and multiple adenomatous, hamartomatous, or juvenile colon polyps American Board of Genetic Counselors 11,20 https://abgcmember.goamp.com/Net/ABGCWcm/Find_ are indicative of other rare hereditary cancer syndromes. These Counselor/ABGCWcm/PublicDir.aspx?hkey=0ad511c0- risk factors should be viewed in the context of the entire family d9e9-4714-bd4b-0d73a59ee175 history, and must be weighed in proportion to the number of in- http://bit.ly/1kzTbk9 dividuals who have not developed cancer. The risk assessment is Directory of board-certified genetic counselors often limited in families that are small or have few female relatives; in such families, a single risk factor may carry more weight. InformedDNA A less common, but extremely important, finding is the pres- www.informeddna.com ence of unusual physical findings or birth defects that are known to (800) 975-4819 be associated with rare hereditary cancer syndromes. Examples in- A nationwide network of independent genetic counselors clude benign skin findings, autism, large head circumference20,21 that use telephone and internet technology to bring genetic and thyroid disorders in Cowden syndrome, odontogenic kerato- counseling to patients and providers. Covered by many cysts in Gorlin syndrome,22 and desmoid tumors or dental abnor- insurance companies. malities in familial adenomatous polyposis (FAP).23 These and National Society of Genetic Counselors other findings should prompt further investigation of the patient’s www.nsgc.org (click “Find a Counselor” button) family history and consideration of a referral to genetic counseling. (312) 321-6834 In this chapter, the breast/ovarian cancer counseling session For a listing of genetic counselors in your area who specialize with a female patient will serve as a paradigm by which all other in cancer. sessions may follow broadly. National Cancer Institute Cancer Genetics Services Directory www.cancer.gov/cancertopics/genetics/directory COMPONENTS OF THE CANCER GENETIC (800) 4-CANCER COUNSELING SESSION A free service designed to locate providers of cancer risk counseling and testing services. Precounseling Information cancers (where the estrogen, progesterone and Her2 receptors are Before coming in for genetic counseling, the counselee should be all negative, often abbreviated ER‐/PR‐/Her2) are overrepresented informed about what to expect at each visit, and what information in BRCA1 families,16,17 and the National Comprehensive Cancer he/she should collect ahead of time. The counselee can then begin Network (NCCN) BRCA testing guidelines now include individu- to collect medical and family history information and pathology re- als diagnosed with a triple negative breast cancer

tahir99 - UnitedVRG Chapter 35 Genetic Counseling 391 the cancer status of common children, but may also determine if models.32,33 Computer-based models are also available to help children are at increased risk for a serious recessive genetic disease determine the chance that a BRCA mutation will be found in a such as Fanconi anemia.26 Children who inherit two copies of a family.34 At first glance, many of these models appear simple and BRCA2 mutation (one from each parent) are now known to have easy to use, and it may be tempting to exclusively rely on these this serious disorder characterized by defective DNA repair and models to assess cancer risk. However, each model has its strengths high rates of birth defects, aplastic anemia, leukemia, and solid and weaknesses, and the counselor needs to understand the limi- tumors.26 Patients should be encouraged to report changes in their tations well and know which are validated, which are considered family history over time (e.g., new cancer diagnoses, genetic test- problematic, when a model will not work on a particular patient, ing results in relatives), because this may change their risk assess- or when another genetic syndrome should be considered. For ex- ment and counseling. ample, none of the existing models are able to factor in other risks A detailed family history should also include genetic diseases, that may be essential in hereditary risk calculation (e.g., a sister birth defects, mental retardation, multiple miscarriages, and infant who was diagnosed with breast cancer after radiation treatment for deaths. A history of certain recessive genetic diseases (e.g., ataxia Hodgkin disease). telangiectasia, Fanconi anemia) can indicate that healthy family The risk of a detectable mutation will also vary based on can- members who carry just one copy of the genetic mutation may be cer history and the degree of relationship to an affected family at increased risk to develop cancer.26,27 Other genetic disorders, member. For example, family members with early-onset breast such as hereditary hemorrhagic telangiectasia, can be associated cancer have a higher likelihood of testing positive than unaffected with a hereditary cancer syndrome caused by a mutation in the family members. Therefore, the risk assessment process should in- same gene—in this case, juvenile polyposis.28 clude a discussion of which family member is the best candidate for testing. Dysmorphology Screening DNA Testing Congenital anomalies, benign tumors, and unusual dermatologic features occur in a large number of hereditary cancer predisposi- DNA testing is now available for a variety of hereditary cancer tion syndromes. Examples include osteomas of the jaw in FAP, syndromes. However, despite misrepresentation by the media, palmar pits in Gorlin syndrome, and papillomas of the lips and testing is feasible for only a small percentage of individuals with mucous membranes in Cowden syndrome. Obtaining an accu- cancer. DNA testing offers the important advantage of presenting rate past medical history of benign lesions and birth defects, and clients with actual risks instead of the empiric risks derived from screening for such dysmorphology can greatly impact diagnosis, risk calculation models. DNA testing can be very expensive; full counseling, and testing. For example, BRCA1/2 testing is inappro- sequencing and rearrangement testing of the BRCA1/2 genes cur- priate in a patient with breast cancer who has a family history of rently averages $2,500, and full panel testing costs up to $7,000 thyroid cancer and the orocutaneous manifestations of Cowden per patient. Importantly, testing should begin in an affected family syndrome. member whenever possible to maximize scientific accuracy. Most insurance companies now cover cancer genetic testing in families AND SCREENING CANCER PREVENTION Risk Assessment where the test is medically indicated. One of the most crucial aspects of DNA testing is accurate re- Risk assessment is one of the most complicated components of the sult ordering and interpretation. Unfortunately, errors in ordering genetic counseling session. It is crucial to remember that risk as- and interpretation are the greatest risk of genetic testing and are sessment changes over time as the person ages and as the health very common.35 Emerging data reveal that between 30% to 50% statuses of their family members change. Risk assessment can be of genetic tests are ordered inappropriately, which is problematic broken down into three separate components. for patients, clinicians, and insurers.36–38 Recent data demonstrate that many medical providers have difficulty interpreting even ■ What is the chance that the counselee will develop the cancer basic pedigrees and genetic test results.33–35 Additional studies have observed in his/her family (or a genetically related cancer such demonstrated that an inaccurate interpretation of genetic testing as ovarian cancer due to a family history of breast cancer)? has been shown to result in inappropriate medical management ■ What is the chance that the cancers in this family are caused by recommendations, unnecessary prophylactic surgeries, a massive a single gene mutation? waste of health-care dollars, psychosocial distress, and false reassur- ■ What is the chance that we can identify the gene mutation ance for patients.2,3 in this family with our current knowledge and laboratory Interpretations are becoming increasingly complicated as more techniques? tests and gene panels become available. For example, one study Cancer clustering in a family may be due to genetic and/or envi- demonstrated that approximately 25% of high-risk families that ronmental factors, or may be coincidental because some cancers were BRCA1 and BRCA2 negative by commercially available se- are very common in the general population.29 Although inherited quencing were found to carry a deletion or duplication in one of factors may be the primary cause of cancers in some families, in these genes, or a mutation in another gene.39 others, cancer may develop because an inherited factor increases This is particularly concerning in an era in which testing com- the individual’s susceptibility to environmental carcinogens. It is panies are canvassing physicians, and now mammography techni- also possible that members of the same family may be exposed to cians, and encouraging them to perform their own counseling and similar environmental exposures due to shared geography or pat- testing. The potential impact of test results on the patient and his/ terns in behavior and diet that may increase the risk of cancer.30 her family is great and, therefore, accurate interpretation of the Therefore, it is important to distinguish the difference between a results is paramount. Professional groups have recognized this and familial pattern of cancer (due to environmental factors or chance) have adopted standards encouraging clinicians to refer patients and a hereditary pattern of cancer (due to a shared genetic muta- to genetics experts to ensure proper ordering and interpretation tion). Emerging research is also evaluating the role and clinical of genetic tests. The U.S. Preventive Services Task Force recom- utility of more common low-penetrance susceptibility genes and mends that women whose family history is suggestive of a BRCA single nucleotide polymorphisms (SNP) that may account for a mutation be referred for genetic counseling before being offered proportion of familial cancers.31 genetic testing.40 The American College of Surgeons’ Commis- Several models are available to calculate the chance that a sion on Cancer standards include “cancer risk assessment, genetic woman will develop breast cancer, including the Gail and Claus counseling and testing services provided to patients either on site 392 Cancer Prevention and Screening or by referral, by a qualified genetics professional.”4 In an effort of accuracy. Family members who do not carry the mutation found to reduce errors, some insurance companies are requiring genetic in their family are deemed true negative. Those who are found counseling by a certified genetic counselor before testing for he- to carry the mutation in their family will have more definitive in- reditary breast or colon cancer syndromes.12 formation about their risks to develop cancer. This information Results can fall into a few broad categories. It is important to can be crucial in assisting patients in decision making regarding note that a negative test result can actually be interpreted in three surveillance and risk reduction. different ways, detailed in #2, #3, and #4, which follows. If a mutation is not identified in the affected relative, it usually means that either the cancers in the family are (1) not hereditary, 1. Deleterious mutation “positive.” When a deleterious mutation or (2) caused by an undetectable mutation or a mutation in a dif- in a well-known cancer gene is discovered, the cancer risks ferent gene. A careful review of the family history and the risk fac- for the patient and her family are relatively straightforward. tors will help to decipher whether interpretation 1 or 2 is more However, with the development of multigene panels and the likely. Additional genetic testing may need to be ordered at this inclusion of many lesser known genes, the risks of detecting a point. In cases in which the cancers appear hereditary and no mu- mutation within a gene whose cancer risks are ill defined and tation is found, DNA banking should be offered to the proband for medical management options unknown is much greater. Even a time in the future when improved testing may become available. for well-known genes, the risks are not precise and should be A letter indicating exactly who in the family has access to the DNA presented to patients as a risk range.41,42 When a true mutation should accompany the banked sample. is found, it is critical to test both parents (whenever possible) to The genetic counseling result disclosure session should also in- determine from which side of the family the mutation is origi- clude a detailed discussion of which other family members would nating, even when the answer appears obvious. benefit from genetic counseling and testing and referral informa- 2. True negative. An individual does not carry the deleterious mu- tion. This can apply not only to families who have been found to tation found in her family, which ideally, has been proven to carry a deleterious mutation, but may also prove useful in other segregate with the cancer family history. In this case, the pa- families (e.g., test a higher risk relative or determine segregation of tient’s cancer risks are usually reduced to the population risks. a variant within a family). 3. Negative. A mutation was not detected, and the cancers in the The penetrance of mutations in cancer susceptibility genes is family are not likely to be hereditary based on the personal and also difficult to interpret. Initial estimates derived from high-risk family history assessment. For example, a patient is diagnosed families provided very high cancer risks for BRCA1 and BRCA2 with breast cancer at age 38 years and comes from a large fam- mutation carriers.43 More recent studies done on populations ily with no other cancer diagnoses and relatives who died at old that were not selected for family history have revealed lower pen- ages of other causes. etrances.44 Because exact penetrance rates cannot be determined 4. Uninformative. A mutation cannot be found in affected family for individual families at this time, and because precise genotype/ members of a family in which the cancer pattern appears to phenotype correlations remain unclear, it is prudent to provide be hereditary; there is likely an undetectable mutation within patients with a range of cancer risk and to explain that their risk the gene, or the family carries a mutation in a different gene. probably falls somewhere within this spectrum. This can prove If, for example, the patient developed breast cancer at age 38 challenging for genes that lack published long-term data on can- years, has a father with breast cancer, and has a paternal aunt cer associations and risks. who developed breast and ovarian cancers before age 50 years, Female carriers of BRCA1 and BRCA2 mutations have a 50% a negative test result would be almost meaningless. It would to 85% lifetime risk to develop breast cancer and between a 15% simply mean that the family has a mutation that could not be to 60% lifetime risk to develop ovarian cancer.15,42,43 It is important identified with our current testing methods or a mutation in to note that the classification “ovarian cancer” also includes can- another cancer gene. The entire family would be followed as cer of the fallopian tubes and primary peritoneal carcinoma.44,45 high risk. BRCA2 carriers also have an increased lifetime risk of male breast 5. Variant of uncertain significance. A genetic change is identi- cancer, pancreatic cancer, and possibly, melanoma.46,47 fied, the significance of which is unknown. It is possible that this change is deleterious or completely benign. It may be help- ful to test other affected family members to see if the mutation Options for Surveillance, Risk Reduction, and segregates with disease in the family. If it does not segregate, Tailored Treatment the variant is less likely to be significant. If it does, the variant is more likely to be significant. Other tools, including a splice site The cancer risk counseling session is a forum to provide coun- predictor, in conjunction with data on species conservation and selees with information, support, options, and hope. Mutation amino acid difference scores, can also be helpful in determin- carriers can be offered: earlier and more aggressive surveillance, ing the likelihood that a variant is significant. It is rarely helpful chemoprevention, and/or prophylactic surgery. Detailed manage- (and can be detrimental) to test unaffected family members for ment options for BRCA carriers are discussed in this chapter. such variants. The rates of variants of uncertain significance Surveillance recommendations are evolving with newer tech- vary greatly depending on the reporting protocols of the lab niques and additional data. At this time, it is recommended that in- and the genes analyzed. Creation of open databases through a dividuals at increased risk for breast cancer, particularly those who nationwide movement called Free the Data will likely improve carry a BRCA mutation, have annual mammograms beginning at variant reporting for all laboratories. age 25 years, with a clinical breast exam by a breast specialist, a In order to pinpoint the mutation in a family, an affected in- yearly breast magnetic resonance imaging (MRI) with a clinical dividual most likely to carry the mutation should be tested first breast exam by a breast specialist, and a yearly clinical breast exam whenever possible. This is most often a person affected with the by a gynecologist.48,49 It is suggested that the mammogram and cancer in question at the earliest age. Test subjects should be MRI be spaced out around the calendar year so that some inter- selected with care, because it is possible for a person to develop vention is planned every 6 months. Recent data suggest that MRI sporadic cancer in a hereditary cancer family. For example, in an may be safer and more effective in BRCA carriers <40 years of early-onset breast cancer family, it would not be ideal to first test a age and may someday replace mammograms in this population.50 woman diagnosed with breast cancer at age 65 years because she BRCA carriers may take a selective estrogen-receptor modula- may represent a sporadic case. tor (SERM) or aromatase inhibitor in hopes of reducing their risks If a mutation is detected in an affected relative, other family of developing breast cancer. These medications have been proven members can be tested for the same mutation with a great degree effective in women at increased risk due to a positive family history

tahir99 - UnitedVRG Chapter 35 Genetic Counseling 393 of breast cancer.51–53 There are limited data on the effectiveness of cancer, the risk appears to be low and not elevated over that of the such medications in unaffected BRCA carriers54–56; however, there general population.75 Removing the uterus may make it possible are some data to suggest that BRCA carriers taking tamoxifen as for a BRCA carrier to take unopposed estrogen or tamoxifen in treatment for a breast cancer reduce their risk of a contralateral the future without the risk of uterine cancer, but this surgery is breast cancer.57 Additionally, the majority of BRCA2 carriers who associated with a longer recovery time and has more side effects develop breast cancer develop an estrogen-positive form of the dis- than does BSO alone. Each patient should be counseled about ease,58 and it is hoped that this population will respond especially the pros and cons of each procedure and the risks associated with well to chemoprevention. Further studies in this area are necessary premature menopause before having surgery.76 before drawing conclusions about the efficacy of chemopreven- A secondary, but important, reason for female BRCA carriers tion in this population. Prophylactic bilateral mastectomy reduces to consider prophylactic oophorectomy is that it also significantly the risk of breast cancer by >90% in women at high-risk for the reduces the risk of a subsequent breast cancer, particularly if they disease.59 Before genetic testing was available, it was not uncom- have this surgery before menopause.77,78 The reduction in breast mon for entire generations of cancer families to have at-risk tissues cancer risk remains even if a healthy premenopausal carrier elects removed without knowing if they were personally at increased risk to take low-dose hormone-replacement therapy (HRT) after this for their familial cancer. Fifty percent of unaffected individuals in surgery79. Early data suggest that tamoxifen, in addition to pre- hereditary cancer families will not carry the inherited predisposi- menopausal oophorectomy, in BRCA carriers may have little ad- tion gene and can be spared prophylactic surgery or invasive high- ditional benefit in terms of breast cancer risk reduction.80 Research risk surveillance regimens. Therefore, it is clearly not appropriate is needed in balancing quality of life issues secondary to estrogen to offer prophylactic surgery until a patient is referred for genetic deprivation with cancer risk reduction in these young female counseling and, if possible, testing.60 BRCA1/2 carriers. Women who carry BRCA1/2 mutations are also at increased New developments are also emerging in the treatment and, risk to develop second contralateral and ipsilateral primaries of the possibly, the prevention of BRCA-related cancers. Early data re- breast.61 These data bring into question the option of breast con- vealed that breast and ovarian cancers in BRCA carriers were par- serving surgery in women at high risk to develop a second primary ticularly sensitive to treatment with poly adenosine diphosphate within the same breast. For this reason, the BRCA1/2 carrier status (ADP)-ribose polymerases (PARP) inhibitors in combination with can have a profound impact on surgical decision making,62 and chemotherapy.81,82 New trials are focusing on which chemothera- many patients have genetic counseling and testing immediately peutic regimens are most effective in mutation carriers. More data after diagnosis and before surgery or radiation therapy. Those pa- are needed on larger cohorts of patients and are currently being tients who test positive and opt for prophylactic mastectomy can studies in multiple clinical trials. often be spared radiation and the resulting side effects that can Genetic counseling and testing is also available for dozens of complicate reconstruction. Approximately 30% to 60% of previ- cancer syndromes, including Lynch syndrome, von Hippel-Lindau ously irradiated patients who later opt for mastectomy with recon- syndrome, multiple endocrine neoplasias, and familial adenoma- struction report significant complications or unfavorable cosmetic tous polyposis. Surveillance and risk reduction for patients who results.62,63 are known mutation carriers for such conditions may decrease the Women who carry BRCA1/2 mutations are also at increased associated morbidity and mortality of these syndromes. AND SCREENING CANCER PREVENTION risk to develop ovarian, fallopian tube, and primary peritoneal can- cer, even if no one in their family has developed these cancers. Follow-up Surveillance for ovarian cancer includes transvaginal ultrasounds and CA-125 testing; however, the effectiveness of such surveil- A follow-up letter to the patient is a concrete means of document- lance in detecting ovarian cancers at early, more treatable stages ing the information conveyed in the sessions so that the patient and has not been proven in any population. Oral contraceptives reduce his/her family members can review it over time. This letter should the risk of ovarian cancer in all women, including BRCA carri- be sent to the patient and health-care professionals to whom the ers.64 Recent data indicate that the impact of this intervention on 56,65 patient has granted access to this information. A follow-up phone increasing breast cancer risk, if any, is low. Given the difficul- call and/or counseling session may also be helpful, particularly in ties in screening and in the treatment of ovarian cancer, the risk/ the case of a positive test result. Some programs provide patients benefit analysis likely favors the use of oral contraceptives in young 30 with an annual or biannual newsletter updating them on new in- carriers of BRCA1/2 mutations who are not yet ready to have their formation in the field of cancer genetics or patient support groups. ovaries removed. Prophylactic bilateral salpingo-oophorectomy It is now recommended that patients return for follow-up coun- (BSO) is currently the most effective means to reduce the risk of seling sessions months, or even years, after their initial consult to ovarian cancer and is recommended to BRCA1/2 carriers by the 66 discuss advances in genetic testing and changes in surveillance age of 35 to 40 or when childbearing is complete. Specific opera- and risk reduction options. This can be beneficial for individuals tive and pathologic protocols have been developed for this prophy- 67 who have been found to carry a hereditary predisposition, for those lactic surgery. In BRCA1/2 carriers whose pathologies come back in whom a syndrome/mutation is suspected but yet unidentified, normal, this surgery is highly effective at reducing the subsequent 68 and for those who are ready to move forward with genetic testing. risk of ovarian cancer. A decision analysis, comparing various sur- Follow-up counseling is also recommended for patients whose life veillance and risk-reducing options available to BRCA carriers, has 69 circumstances have changed (e.g., preconception, after childbear- shown an increase in life expectancy if BSO is pursued by age 40. ing is complete), who are preparing for prophylactic surgery, or Emerging data indicate that most ovarian cancers begin in the fal- who are ready to discuss the family genetics with their children. lopian tube, and that salpingectomy may someday be sufficient in reducing ovarian cancer risk in young women; however, more data are needed before this option is offered to patients outside of clini- 70 ISSUES IN CANCER GENETIC cal trials. A relatively small percentage of women who pursue COUNSELING BSO may develop primary peritoneal carcinoma.44,71 There has been some debate about whether BRCA1/2 carriers should also opt for total abdominal hysterectomy (TAH) due to the fact that Psychosocial Issues small stumps of the fallopian tubes remain after BSO alone. The question of whether BRCA carriers are at increased risk for uter- The psychosocial impact of cancer genetic counseling cannot ine serous papillary carcinoma (USPC) has also been raised.72–74 be underestimated. Just the process of scheduling a cancer risk If a relationship does exist between BRCA mutations and uterine counseling session may be quite difficult for some individuals 394 Cancer Prevention and Screening with a family history who are not only frightened about their own with direct-to-consumer tests and whole exome testing of chil- cancer risk, but also are reliving painful experiences associated dren.91 The risks of such testing to the child, and the child’s right with the cancer of their loved ones.13 Counselees may be faced not to be tested must be considered. Whenever childhood testing with an onslaught of emotions, including anger, fear of develop- is not medically indicated, it is preferable that testing decisions are ing cancer, fear of disfigurement and dying, grief, lack of control, postponed until the children are adults and can decide for them- negative body image, and a sense of isolation.24 Some counselees selves whether to be tested. wrestle with the fear that insurance companies, employers, family members, and even future partners will react negatively to their Confidentiality cancer risks. For many, it is a double-edged sword as they balance their fears and apprehensions about dredging up these issues with The level of confidentiality surrounding cancer genetic testing the possibility of obtaining reassuring news and much needed is paramount due to concerns of genetic discrimination. Careful information. consideration should be given to the confidentially of family his- A person’s perceived cancer risk is often dependent on many tory information, pedigrees, genetic test results, pathology reports, “nonmedical” variables. They may estimate that their risk is higher and the carrier status of other family members as most hospitals if they look like an affected individual, or share some of their per- and clinicians transition to electronic medical records systems. sonality traits.24 Their perceived risks will vary depending on if The goal of electronic records is to share information about the their relatives were cancer survivors or died painful deaths from patient with his/her entire health-care team. However, genetics is the disease. Many people wonder not if they are going to get can- a unique specialty that involves the whole family. Patient’s charts cer, but when. often contain Health Insurance Portability and Accountability Act The counseling session is an opportunity for individuals to ex- (HIPAA)–protected health information and genetic test results for press why they believe they have developed cancer, or why their many other family members. This information may not be appro- family members have cancer. Some explanations may revolve priate to enter into an electronic record. The unique issues of ge- around family folklore, and it is important to listen to and address netics services need to be considered when designing electronic these explanations rather than dismiss them.24 In doing this, the medical record standards. counselor will allow the clients to alleviate their greatest fears and Confidentiality of test results within a family can also be of to give more credibility to the medical theory. Understanding a issue, because genetic counseling and testing often reveals the patient’s perceived cancer risk is important, because that fear may risk statuses of family members other than the patient. Under decrease surveillance and preventive health-care behaviors.83 For confidentiality codes, the patient needs to grant permission be- patients and families who are moving forward with DNA testing, a fore at-risk family members can be contacted. For this reason, referral to a mental health-care professional is often very helpful. many programs have built in a “share information with family Genetic testing has an impact not only on the patient, but also on members” clause to their informed consent documents. It has his/her children, siblings, parents, and extended relatives. This can been questioned whether or not a family member could sue a be overwhelming for an individual and the family, and should be health-care professional for negligence if they were identified at discussed in detail prior to testing. high risk yet not informed.92 Most recommendations have stated To date, studies conducted in the setting of pre- and post- that the burden of confidentiality lies between the provider and genetic counseling have revealed that, at least in the short term, the patient. However, more recent recommendations state that most patients do not experience adverse psychological outcomes confidentiality should be violated if the potential harm of not no- after receiving their test results.84,85 In fact, preliminary data have tifying other family members outweighs the harm of breaking a revealed that individuals in families with known mutations who confidence to the patient.93 There is no patent solution for this seek testing seem to fare better psychologically at 6 months than difficult dilemma, and situations must be considered on a case- those who avoid testing.84 Among individuals who learn they are by-case basis with the assistance of the in-house legal department BRCA mutation carriers, anxiety and distress levels appear to in- and ethics committee. crease slightly after receiving their test results but returned to pre- test levels in several weeks.86 Although these data are reassuring, it is important to recognize that genetic testing is an individual deci- Insurance and Discrimination Issues sion and will not be right for every patient or every family. When genetic testing for cancer predisposition first became widely available, the fear of health insurance discrimination by both pa- Presymptomatic Testing in Children tients and providers was one of the most common concerns.94,95 It appears that the risks of health insurance discrimination were Presymptomatic testing in children has been widely discussed, overstated and that almost no discrimination by health insurers has and most concur that it is appropriate only when the onset of been reported.96 HIPAA banned the use of genetic information as a the condition regularly occurs in childhood or if there are useful preexisting condition.97,98 In May of 2008, Congress passed the Ge- interventions that can be applied.87 For example, genetic testing netic Information Nondiscrimination Act (GINA, HR 493), which for mutations in the BRCA genes and other adult-onset diseases provides broad protection of an individual’s genetic information is generally limited to individuals who are >18 years of age. The against health insurance and employment discrimination.99 In ad- American College of Medical Genetics states that if the “medical dition, the Heath Care and Education Reconciliation Act of 2010 or psychosocial benefits of a genetic test will not accrue until adult- (HR 4872) prohibits group health plans from denying insurance hood . . . genetic testing generally should be deferred.”88 In con- based on preexisting conditions and from increasing premiums trast, the DNA-based diagnosis of children and young adults at risk based on health status.100 Health-care providers can now more for hereditary medullary thyroid carcinoma (MTC) is appropriate confidently reassure their patients that genetic counseling and test- and has improved the management of these patients.89 DNA-based ing will not put them at risk of losing group or individual health testing for MTC is virtually 100% accurate and allows at-risk family insurance. members to make informed decisions about prophylactic thyroi- More and more patients are choosing to submit their genetic dectomy. FAP is a disorder that occurs in childhood and in which counseling and/or testing charges to their health insurance compa- mortality can be reduced if detection is presymptomatic.90 Testing nies. In the past few years, more insurance companies have agreed is clearly indicated in these instances. to pay for counseling and/or testing,101 perhaps in light of data that Questions have been raised about the parents’ right to demand show these services reduce errors related to ordering and inter- testing for adult-onset diseases, and this is now happening regularly preting genetic testing and that decision analyses have revealed

tahir99 - UnitedVRG Chapter 35 Genetic Counseling 395

subsequent prophylactic surgeries to be cost effective.102 The risk All laboratories that have entered the BRCA marketplace have of life or disability insurance discrimination, however, is more re- done so by including BRCA1 and BRCA2 in gene panels. These alistic. Patients should be counseled about such risks before they panels simultaneously analyze groups of genes that contribute to pursue genetic testing. increased risk for breast, colon, ovarian, uterine, and other cancers. The cost of this technology continues to decrease with some multi- gene panels costing just a few hundred dollars less than traditional Reproductive Issues BRCA testing (∼$4,000). Some panels include only well-known genes (e.g., p53, APC, MLH1), although many include lesser known Reproductive technology in the form of preimplantation genetic 103 genes (e.g., BRIP1, NBN, MRE11A) for which cancer risks are ill diagnosis, prenatal testing, or sperm sorting are options for men defined and medical management options are unknown. Because and women with a hereditary cancer syndrome, but are requested testing for these genes is new to the clinical setting, it is expected by few patients for adult-onset conditions in which there are vi- to take several years to compile accurate cancer risk estimates and able options for surveillance and risk reduction. Importantly, if a appropriate recommendations for surveillance and risk reduction. BRCA2 carrier is considering having a child, it is important to as- Furthermore, the rate of variants of uncertain significance will sess the spouse’s risk of also carrying a BRCA2 mutation. If the likely be more common in the lesser known genes. These changes spouse is of Jewish ancestry or has a personal or family history of have increased the complexity of genetic testing exponentially. In breast, ovarian, or pancreatic cancer, BRCA testing should be con- response, several state and one national insurance company have sidered and a discussion of the risk of Fanconi anemia in a child 104 mandated genetic counseling by certified providers before they will with two BRCA2 mutations should take place. cover cancer genetic testing. In a surprising response, the Ameri- can Society of Clinical Oncology (ASCO) opposed this insurer’s decision, despite more than a decade’s worth of data demonstrating RECENT ADVANCES AND FUTURE that the majority of physicians do not have the time or expertise to DIRECTIONS offer genetic counseling and testing(38,105‐108). The AMA will decide whether to back the ASCO resolution in June 2014. Cancer genetic counseling and testing were thrust into the na- Some companies are now offering direct-to-consumer (DTC) tional spotlight in the spring of 2013 when Hollywood icon Ange- genetic testing via websites. The accuracy of some of these lina Jolie publically disclosed that she was a BRCA1 carrier. One DTC genetic tests are in question, and the leading company, month later the Supreme Court unanimously ruled against gene 23andMe, has recently come under fire by the U.S. Food and patents. Referrals for genetic testing spiked across the country and Drug Administration.105 have not returned to baseline levels at most centers. Within hours Maintaining high standards for thorough genetic counseling, of the ruling, other labs began offering less expensive and more informed consent, and accurate result interpretation will be para- comprehensive BRCA testing, dramatically changing the market- mount in reducing potential risks and maximizing the benefits of place of genetic testing for hereditary breast cancer. genetic technology in the next century. CANCER PREVENTION AND SCREENING CANCER PREVENTION

SELECTED REFERENCES

The full reference list can be accessed at lwwhealthlibrary.com/oncology. 55. King M, Wieand S, Hale K. Tamoxifen and breast cancer incidence among women with inherited mutations in BRCA1 and BRCA2. JAMA 2001;286: 3. Brierley KL, Blouch E, Cogswell W, et al. Adverse events in cancer ge- 2251–2256. netic testing: medical, ethical, legal, and financial implications.Cancer J 57. Phillips KA, Milne RL, Rookus MA, et al. Tamoxifen and risk of contralat- 2012;18:303–309. eral breast cancer for BRCA1 and BRCA2 mutation carriers. J Clin Oncol 8. Doksum T, Bernhardt BA, Holtzman NA. Does knowledge about the genet- 2013;31:3091–3099. ics of breast cancer differ between nongeneticist physicians who do or do not 59. Hartmann L, Schaid D, Woods J. Efficacy of bilateral prophylactic mas- discuss or order BRCA testing? Genet Med 2003;5:99–105. tectomy in women with a family history of breast cancer. N Engl J Med 13. Claus E, Schildkraut J, Thompson W, et al. The genetic attributable risks of 1999;340:77–84. breast and ovarian cancer. Cancer 1996;77:2318–2324. 61. Turner B, Harold E, Matloff E, et al. BRCA1/BRCA2 germline mutations 14. Loman N, Johannsson O, Kristoffersson U. Family history of breast and ovar- in locally recurrent breast cancer patients after lumpectomy and radiation ian cancers and BRCA1 and BRCA2 mutations in a population-based series therapy: Implications for breast-conserving management in patients with of early-onset breast cancer. J Natl Cancer Inst 2001;93:1215. BRCA1/BRCA2 mutations. J Clin Oncol 1999;17:3017–3024. 21. Pilarski R. Cowden syndrome: a critical review of the clinical literature. 65. Milne R, Knight J, John E, et al. Oral contraceptive use and risk of early-onset J Genet Couns 2009 Feb;18:13–27. breast cancer in carriers and noncarriers of BRCA1 and BRCA2 mutations. 25. Love R, Evan A, Josten D. The accuracy of patient reports of a family history. Cancer Epidemiol Biomarkers Prev 2005;14:350–356. J Chronic Dis 1985;38(4):289–293. 66. Domchek S, Friebel T, Neuhausen S, et al. Mortality reduction after risk- 31. Stratton MR, Rahman N. The emerging landscape of breast cancer suscepti- reducing bilateral salpingo-oophorectomy in a prospective cohort of BRCA1 bility. Nat Genet 2008;40:17–22. and BRCA2 mutation carriers. Lancet Oncol 2006;7:223–229. 38. Plon SE, Cooper HP, Parks B, et al. Genetic testing and cancer risk management 67. Powel CB, Kenley E, Chen LM, et al. Risk-reducing salpingo-oophorectomy recommendations by physicians for at-risk relatives. Genet Med 2011;13:148–154. in BRCA mutation carriers: role of serial sectioning in the detection of occult 41. King MC, Marks JH, Mandell JB, et al. Breast and ovarian cancer risks due malignancy. J Clin Oncol 2005;23:127–132. to inherited mutations in BRCA1 and BRCA2. Science 2003;302:643–646. 77. Rebbeck T, Lynch H, Neuhausen S, et al. Prophylactic oophorectomy in 42. Antoniou A, Pharoah PD, Narod S, et al. Average risks of breast and ovarian carriers of BRCA1 or BRCA2 mutations. N Engl J Med 2002;346:1616–1622. cancer associated with BRCA1 or BRCA2 mutations detected in case Series 79. Rebbeck T, Friebel T, Wagner T, et al. Effect of short-term hormone re- unselected for family history: a combined analysis of 22 studies. Am J Hum placement therapy on breast cancer risk reduction after bilateral prophylactic Genet 2003;72:1117–1130. oophorectomy in BRCA1 and BRCA2 mutation carriers: the PROSE study 46. van Asperen C, Brohet R, Meijers-Heijboer, et al. Cancer risks in BRCA2 fami- group. J Clin Oncol 2005;23:7804–7810. lies: estimates for sites other than breast and ovary. J Med Genet 2005;42:711–719. 88. ASHG/ACMG. Points to consider: ethical, legal, and psychosocial implica- 47. Breast Cancer Linkage Consortium. Cancer risks in BRCA2 mutation carri- tions of genetic testing in children and adolescents. American Society of Hu- ers. J Natl Cancer Inst 1999;91:1310–1316. man Genetics Board of Directors, American College of Medical Genetics 48. Warner E, Plewes D, Hill K, et al. Surveillance of BRCA1 and BRCA2 muta- Board of Directors. Am J Hum Genet 1995;57:1233–1241. tion carriers with magnetic resonance imaging, ultrasound, mammography, 99. The Genetic Information Nondiscrimination Act of 2008 (H.R. 493). Li- and clinical breast examination. JAMA 2004;202:1317–1325. brary of Congress Web site. http://thomas.loc.gov/cgi-bin/bdquery/z?d110 49. Kriege M, Brekelmans CT, Boetes C, et al. Efficacy of MRI and mammogra- :h.r.00493. Accessed December 3, 2012. phy for breast-cancer screening in women with a familial or genetic predispo- 108. Bellcross C, Kolor K, Goddard K, et al. Awareness and utilization of BRCA1/2 sition. N Engl J Med 2004;29:351:427–437. testing among U.S. primary care physicians. Am J Prev Med 2011;40:61–66. tahir99 - UnitedVRG PART V

Practice of Oncology Design and Analysis of 36 Clinical Trials

Richard M. Simon

INTRODUCTION PHASE 1 CLINICAL TRIALS

Clinical trials are experiments to determine the value of a The main objectives of phase 1 trials have traditionally been to de- treatment. There are two key components to the experimen- termine a dose that is appropriate for use in phase 2 and 3 trials and tal approach. First, results rather than plausible reasoning are to determine information about the pharmacokinetics of distribu- required to support conclusions. Second, in an experiment the tion of the drug. Patients with advanced disease that is resistant to treatments are assigned so that one can conclude that differ- standard therapy but who have normal organ function are usually ences in outcome are due to differences in treatment effect. included in such trials. In observational studies, treatments are not assigned as part of Phase 1 trials are usually initiated at a low dose that is not the study, so differences in outcome between treatment groups expected to produce serious toxicity. A starting dose of one- may merely result from the fact that sicker patients received tenth the lethal dose (expressed as milligrams per square meter less intensive treatments. Experiments should be prospectively of body surface area) in the most sensitive species usually is planned and conducted under controlled conditions to provide used.3 The dose is increased for subsequent patients according definitive answers to well-defined questions. Using tumor reg- to a series of preplanned steps. Dose escalation for subsequent istry data to compare the survival rates of patients with prostate patients occurs only after sufficient time has passed to observe cancer treated with surgery to those of patients receiving ra- acute toxic effects for patients treated at lower doses. Cohorts diotherapy is an example of an observational study, not a clini- of three to six patients are treated at each dose level. Usually, if cal trial. In an observational study, the investigators are passive no dose-limiting toxicity (DLT) is seen at a given dose level, the observers. Treatment assignments, staging workup, and follow- dose is escalated for the next cohort. If the incidence of DLT is up procedures are out of the control of the investigators, and 33%, then three more patients are treated at the same level. If no are conducted with no considerations about the validity of the further cases of DLT are seen in the additional patients, then the subsequent attempt at comparison. The statistical associations dose level is escalated for the next cohort. Otherwise, dose esca- resulting from such studies are, consequently, a weak basis for lation stops. If the incidence of DLT is >33% at a given level, causal inferences about relationships between the treatments then dose escalation also stops. The phase 2 recommended dose administered and the outcomes observed. Surprisingly, this often is taken as the highest dose for which the incidence of does not seem to be realized by the politicians and health-care DLT is <33%. Usually, six or more patients are treated at the administrators allocating enormous sums of money to outcomes recommended dose. research based on electronic medical records from general The dose levels themselves are commonly based on a modified practice. In such observational studies, treatments are usually Fibonacci series. The second level is twice the starting dose, the third selected on the basis of subjective assessment of the progno- level is 67% greater than the second, the fourth level is 50% greater sis of the patient, specialties of the physician, and diagnostic than the third, the fifth is 40% greater than the fourth, and each sub- evaluations. Unknown patient selection factors generally are sequent step is 33% greater than that preceding it. Escalating doses more important determinants of patient outcome than are dif- for subsequent courses in the same patient are generally not done, ferences between treatments. For example, Subramanian and except at low doses before any DLT has been encountered. Simon1 found that in observational studies that developed gene expression prognostic signatures for patients with early stage nonsmall-cell lung cancer, those who received chemotherapy Accelerated Titration Designs had poorer survivals than those who did not even after adjusting for all recorded prognostic factors. There is no compelling scientific basis for the approach just out- Clinical trials require careful planning. The first result of lined, except that experience has shown it to be safe. Traditional the planning process is a written protocol. Typical subject head- phase 1 trials have three limitations: ings for the protocol are shown in Table 36.1, and the protocol 1. They sometimes expose too many patients to subtherapeutic development process is discussed in more detail by Green et al.2 doses of the new drug. The protocol should define treatment and evaluation policies for 2. The trials may take a long time to complete. a well-defined set of patients. It also should define the specific 3. They provide very limited information about interpatient vari- questions to be answered by the study and should directly jus- ability and cumulative toxicity. tify that the number of patients and the nature of the controls are adequate to answer these questions. Some clinical trials are New trial designs have been developed to address these prob- really only guidelines for clinical management supplemented lems.4 The accelerated titration designs5 permit within-patient dose by lofty objectives with no scientific meaning and no realistic escalation and use only one patient per dose level until grade 2 or chance of providing a reliable answer to a well-defined medical greater toxicity is seen. Doses are titrated within patients to achieve question. Such studies are a disservice to the patients who are grade 2 toxicity. The analysis consists of fitting a statistical model undergoing some inconvenience to contribute to the welfare of to the full set of data that includes all grades of toxicity for all future patients. courses of a patient’s treatment. The model includes parameters

398

tahir99 - UnitedVRG Chapter 36 Design and Analysis of Clinical Trials 399

TABLE 36.1 dose. A Bayesian prior distribution is established for the steepness of the dose-toxicity curve and the distribution is updated after each Subject Headings for a Protocol patient is treated. The model is based on using only first-course Introduction and scientific background treatment data and whether the patient experiences DLT. This ap- Objectives proach is called the continual reassessment method. For each new Selection of patients patient, the model is used to determine the dose predicted to cause Design of study (including schematic diagram) DLT to a specified percentage of the patients. That dose is assigned to the next patient. Many modifications of the original continual Treatment plan 9–11 Drug information reassessment method have been subsequently proposed. For some tumor vaccines and molecularly targeted drugs, toxic- Toxicities to be monitored and dosage modifications 12 Required clinical and laboratory data and study calendar ity may not be dose limiting, and the dose selected may be based Criteria for evaluating the effect of treatment and end point on preclinical findings or on practical considerations. For some definition molecularly targeted drugs, preclinical studies provide a target Statistical considerations serum concentration of the active moiety necessary to maximally Informed consent and regulatory considerations inhibit the target, and drug administration can be titrated for each Data forms patient to the targeted serum concentration. This approach can be References complex because it involves developing a population pharmacoki- Study chairperson, collaborating participants, addresses, and netic model relating dose to concentration as the study progresses. telephone numbers A simpler approach is to have separate cohorts of patients who are treated at each of several dose levels without intrapatient dose titra- tion. A population pharmacokinetic model relating dose to con- centration is fit to the data. Ideally, a trial design should provide the smallest dose that gives maximum biologic effect. For molecularly targeted therapeutics, that represent the steepness of the dose-toxicity curve, the degree the biologic effect might be a measure of the degree of inhibi- of interpatient variability in the location of the dose-toxicity curve, tion of the target. Because it can be very difficult to obtain tumor and the degree (if any) of cumulative toxicity. All these parameters samples before and after treatment, biologic effect is sometimes are estimated from the data. measured in an accessible surrogate tissue, such as peripheral Several variants of the accelerated titration design were stud- blood lymphocytes or skin, or by using functional imaging.13 For ied. Design A uses conventional 40% dose steps during the initial therapeutic vaccines, the biologic effect might be a measure of accelerated phase, whereas designs B and C use 100% dose steps stimulation of tumor reactive T cells. until one patient experiences DLT or two patients experience

Finding the dose that provides maximum biological effect is PRACTICE OF ONCOLOGY grade 2 toxicity. At that point, acceleration ceases and standard often not practical in a phase 1 trial, as it may require a large cohorts of three to six patients with 40% dose-step increments are number of patients. For example, to have 90% power for detect- used. These designs were compared to a control design using co- ing a one standard error difference in mean response between horts of three to six patients with 40% dose-step increments and no two dose levels at a one-sided 10% significance level requires 14 intrapatient dose escalation. patients per dose level. A more limited objective is to identify In the 20 phase 1 trials initially evaluated, only three showed a dose that is biologically active. Korn et al.14 developed a se- any evidence of cumulative toxicity. The average number of pa- quential procedure for finding such a dose when the measure of tients required was reduced from 39.9 for the control design to biologic response is binary. During an initial accelerated phase, 24.4, 20.7, and 21.2 for designs A, B, and C, respectively. The aver- they treat one patient per dose level until a biologic response is age number of patients who had grade 0 to 1 toxicity as their worst seen. Then, they treat cohorts of three to six patients per dose toxicity grade over three cycles of treatment was 23.3 for the con- level. With zero to one biologic responses among three patients trol but only 7.9, 3.9, and 4.8 for designs A, B, and C, respectively. at a dose level, they escalate to the next level. With two to three The average number of patients with a worst toxicity grade of 3 in- responses among three patients, they expand the cohort to six pa- creased from 5.5 for the control to 6.2, 6.8, and 6.2 for designs A, B, tients. With five to six biologic responses from the six patients, and C, respectively. The average number of patients with a worst they declare that dose to be the biologically active level and ter- toxicity grade of 4 increased from 1.9 for the control to 3.0, 4.3, minate the trial. With four or fewer biologic responses at a level, and 3.2 for designs A, B, and C, respectively. Accelerated titration they continue to escalate. designs appear to be effective in reducing the number of patients Designs have also been developed for phase zero proof of necessary for finding the maximum tolerated dose, for reducing concept trials.15,16 Patients are treated with single doses of a new the number who are undertreated, and for providing increased in- drug at very low concentrations not expected to cause toxicity. formation. They do not necessarily reduce the length of time nec- This enables the investigator to obtain an early assessment of essary for completion of the trial. They increase the information whether the molecular target of the drug is being inhibited by yield if investigators analyze the results of the trial using the model measuring a pharmacodynamic end point before and after drug developed by Simon et al.5 Software for fitting the model is avail- administration. These trials require prior development of an assay able at http://brb.nci.nih.gov. Software for determining dose assign- for measuring the pharmacodynamic end point and an adequate ments and for recording the data in a spreadsheet format are also database for estimating the variability of measurement for inde- available at that website. The model of Simon et al.5 uses actual pendent tissue samples of the same patient. This estimate should worst grade toxicity for each course of treatment of each patient, reflect variability of tissue sampling as well as technical variabil- and it enables one to determine whether there is cumulative toxic- ity of the assay. The approach developed depends on having a ity and to estimate the variability among patients in toxic effects. good estimate of assay variability and in having assay sufficiently The use of the accelerated titration design has been reviewed.6,7 reproducible to be able to reliably classify individual patients as responders or nonresponders based on the observed change in the Continual Reassessment Methods level of the pharmacodynamic end point. The designs described by Rubinstein et al.16 utilize a small numbers of patients for estab- O’Quigley, Pepe, and Fisher8 used a dose-toxicity model to guide lishing whether the drug causes target inhibition in a substantial the dose escalation, as well as to determine the maximum tolerated proportion of patients. 400 Practice of Oncology

PHASE 2 CLINICAL TRIALS a classifier of the tumors likely to respond to the drug. Dobbin, Zhao, and Simon22 have provided sample size guidelines for ge- nomewide expression profiling studies and generally recommend Patient Selection at least 20 responders for developing a classifier. Pusztai, Anderson, and Hess20 performed a computer simulation study to indicate that Phase 2 trials have traditionally been performed separately by HER-2 transcript overexpression would have been missed as a pre- tumor type in patients with the least amount of prior therapy for dictive biomarker for treatment of advanced breast cancer with whom no effective therapy is available. With cytotoxics, full-dose trastuzumab in whole genome expression profiling with only five chemotherapy is often impossible in patients debilitated by prior responders to analyze. They recommend analysis based on candi- treatment, and lack of chemotherapeutic activity in previously date genes if the number of responders are very limited. treated patients may not indicate lack of clinical usefulness in ear- lier disease. The development of molecularly targeted drugs has introduced new complexities with regard to selection and evalu- Single-Arm Phase 2 Trials ation of patients for phase 3 trials. When the target of the drug is clearly known, it may be more appropriate to select patients based Single Agents on target expression than based on primary site of disease. Even if target expression is not used as an eligibility criterion, the drug For most single-agent phase 2 trials, the objective is simply to de- should be evaluated in an adequate number of patients whose termine whether the drug has activity against the tumor type in tumors express the target. Consequently, it is important to have question. For this objective, response rate based on the response an adequate assay for the target available at the time that phase 2 evaluation criteria in solid tumors guidelines may provide a satis- development begins. factory approach.23 A variety of statistical accrual plans and sample In many cases, the drug will have multiple targets; there may be size methods have been developed for single-arm phase 2 trials. several candidate assays available for each target. Expression of the One of the most popular approaches is the optimal two-stage de- 24 target will often prove to be only part of the relevant genomic in- sign. n1 evaluable patients are entered into study in the first stage formation. For example, the effectiveness of antiepidermal growth of the trial. If no more than r1 responses are obtained among these factor receptor antibodies cetuximab and pannitumumab turned n1 patients, then accrual terminates and the drug is rejected as out to depend on whether the tumor had an activating K-RAS being of little interest. Otherwise, accrual continues to a total of mutation.17–19 n evaluable patients. At the end of the second stage, the drug is Whereas the major objective of phase 2 trials has traditionally rejected if the observed response rate is less than or equal to r/n, been to identify the primary tumor sites in which a new drug was where r and n are determined by the design used. active, a new important objective is to develop promising predic- Tables 36.2 and 36.3 illustrate some of these optimized de- tive biomarkers that identify the patients whose tumors are most signs, and a web-based interactive computer program is available (or least) likely to respond to the drug. The phase 2 development at http://linus.nci.nih.gov/brb. To select a design, the investigator stage is also the time to select the assay(s) that will be used in the specifies the target activity level of interest,p 1, and also a lower phase 3 trials of the new drug and to define the criteria that will activity level, p0, representing inadequate activity. The first row of be used to either select patients for such trials or to structure the each triplet of optimal designs provides designs with probability analysis, as will be described later in this chapter. 0.10 of accepting drugs worse than p0 and probability 0.10 of re- It is often undesirable to restrict entry to phase 2 trials based jecting drugs better than p1. Subject to these two constraints, the on what one thinks one knows about the drug target, at least in optimal designs minimize the average sample size. The average cases where this knowledge is uncertain. It is important, however, sample size is calculated at the lower activity level p0 to optimize to ensure that the activity of the drug is not missed because the protection of patients from exposure to inactive drugs. The tables phase 2 trials did not accrue enough of the right kinds of patients. show for each design the optimal values of r1, n1, r, and n; the aver- The decision of whether to restrict entry based on the presumed age sample size; and the probability of stopping after the first stage mechanism of action will depend in part on the adverse effects of for a drug with activity level p0. the drug. These tables also show the “minimax” designs, which provide If tumor specimens are archived for the patients entered on the smallest maximum sample size n that satisfies the two con- broad eligibility phase 2 trials, then one avoids the need to develop straints just described. Although minimax designs have somewhat assays in advance for all candidate targets, but it is not possible larger average sample sizes than do optimal designs, in some in- to ensure adequate accrual for subsets of patients whose tumors stances, they are preferable because the small increase in average are positive for the candidate markers. Pusztai, Anderson, and sample size is more than compensated for by a large reduction in Hess20 described a hybrid approach that begins with conducting maximum sample size. a standard single-arm two-stage design for evaluating whether the The designs shown in Tables 36.2 and 36.3 are two-stage de- overall response rate for unrestricted patients is sufficiently large. signs with the potential for early stopping for lack of activity. Op- If the overall response rate is sufficient in the first stage of the stan- timized three-stage designs have been described by Ensign et al.25 dard phase 2 trial, then the second stage is completed with accrual Others have extended the design to incorporate toxicity or tumor of additional unrestricted patients. If there are too few responses progression information.26–28 overall in the first stage, then one starts a two-stage phase 2 study Some authors have recommended use of progression-free restricting entry to patients who are marker positive. If there are survival instead of response29 for evaluating molecularly targeted multiple markers of interest, then one restricts entry to patients drugs that may be cytostatic. Single-arm phase 2 trials can be de- positive for one of the markers and ensures that each marker has signed using Tables 36.2 and 36.3 for testing whether the propor- sufficient number of positive patients for evaluation. LeBlanc tion of patients with stable disease at a specified landmark time like et al.21 have described how multiple primary sites can be incorpo- 12 months after the start of treatment is greater than a specified rated in a single phase 2 trial. value p0, but that is only meaningful if the value p0 is a stable, In some cases, the list of candidate targets can be narrowed robust, and well-characterized stable disease rate that results from using mRNA transcript expression profiling of the pretreatment multiple large studies with control regimens. Single-arm studies specimens. By comparing pretreatment expression levels of re- using stable disease are rarely planned or analyzed with that care sponders to nonresponders, one can potentially prioritize targets and hence conclusions of single-arm phase 2 trials claiming that for assay development. If one does not have a good list of candidate molecularly targeted agents cause disease stabilization are often targets, genomewide expression profiling can be used to develop dubious.30 Vidauurre et al.30 have questioned, however, whether

tahir99 - UnitedVRG Chapter 36 Design and Analysis of Clinical Trials 401

TABLE 36.2

a Simon Two-Stage Phase 2 Designs for p1 − p0 = 0.20

Optimal Design Minimax Design Reject Drug if Reject Drug if Response Rate Response Rate

P0 P1 r1/n1 r/n EN (p0)PET (p0) r1/n1 r/n EN (p0)PET (p0) 0.05 0.25 0/9 2/24 14.5 0.63 0/13 2/20 16.4 0.51 0/9 2/17 12.0 0.63 0/12 2/16 13.8 0.54 0/9 3/30 16.8 0.63 0/15 3/25 20.4 0.46 0.10 0.30 1/12 5/35 19.8 0.65 1/16 4/25 20.4 0.51 1/10 5/29 15.0 0.74 1/15 5/25 19.5 0.55 2/18 6/36 22.5 0.71 2/22 6/23 26.2 0.62 0.20 0.40 3/17 10/37 26.0 0.55 3/19 10/36 28.2 0.46 3/13 12/43 20.6 0.75 4/18 10/33 22.3 0.50 4/19 15/54 30.4 0.67 5/24 13/45 31.2 0.66 0.30 0.50 7/22 17/46 29.9 0.67 7/28 15/39 35.0 0.36 5/15 18/46 23.6 0.72 6/19 16/39 25.7 0.48 8/24 24/63 34.7 0.73 7/24 21/53 36.6 0.56 0.40 0.60 7/18 22/46 30.2 0.56 11/28 20/41 33.8 0.55 7/16 23/46 24.5 0.72 17/34 20/39 34.4 0.91 11/25 32/66 36.0 0.73 12/29 27/54 38.1 0.64 0.50 0.70 11/21 26/45 29.0 0.67 11/23 23/39 31.0 0.50 8/15 26/43 23.5 0.70 12/23 23/37 27.7 0.66 13/24 36/61 34.0 0.73 14/27 32/53 36.1 0.65 0.60 0.80 6/11 26/38 25.4 0.47 18/27 24/35 28.5 0.82 7/11 30/43 20.5 0.70 8/13 25/35 20.8 0.65 12/19 37/53 29.5 0.69 15/26 32/45 35.9 0.48 PRACTICE OF ONCOLOGY 0.70 0.90 6/9 22/28 17.8 0.54 11/16 20/25 20.1 0.55 4/6 22/27 14.8 0.58 19/23 21/26 23.2 0.95 11/15 29/36 21.2 0.70 13/18 26/32 22.7 0.67 a For each value of (p0, p1), designs are given for three sets of error probabilities (α, β). The first, second, and third rows correspond to error probability limits (0.10, 0.10), (0.05, 0.20), and (0.05, 0.10), respectively. α is the probability of accepting a drug with response probability p0. β is the probability of rejecting a drug with response probability p1. For each design, EN (p0) and PET (p0) denote the expected sample size and the probability of early termination when the true response probability is p0. molecularly targeted drugs are any more cytostatic than conven- For comparative trials of response rates using specific historic tional chemotherapy drugs. El-Maraghi and Eisenhauer31 have controls, the sample size should be planned using the formulas also recommended that objective response is a useful end point for appropriate for randomized clinical trials. By inserting the num- screening molecularly targeted agents. ber of historic controls to be used, one can compute the number of patients needed to treat on the new regimen in the single-arm 32 Combination Regimens phase 2 trial. For binary end point data, the results of these calculations are presented in Table 36.4 for 80% power with a Determination whether a new drug adds anticancer activity to an one-sided 10% significance level. The tabulated entries indicate active regimen is inherently comparative. In using Tables 36.2 that a 25 percentage-point difference can be detected with <40 and 36.3 to design a single-arm trial, p0 should represent the level new patients if there are at least 30 appropriate historic controls. of activity of existing standard regimens. If this response probability The table entries indicate that detecting a 15 percentage-point is not well determined, however, because it varies among studies difference is almost never feasible with this single-arm approach and varies based on patient prognostic factors, then a single-arm and that detecting a 20 percentage-point difference generally re- ≥ trial based on an assumed known p0 may not be appropriate. quires at least 50 appropriate historical controls and 60 new Several approaches to single-arm study design have been de- patients. veloped that attempt to either account for or control the variability Thall and colleagues33,34 have developed and used Bayesian in p0. One approach to controlling this variability is to base the methods for planning and conducting single-institution trials analysis of the single-arm trial on comparison to a specific set of comparing one or more new regimens to a specific set of historic control patients, matched for prognostic factors, and treated at the controls who received a control treatment at the same institution. same institution as those for the new study. This can be a better ap- The Bayesian designs provide for continual analysis of results with proach than just using an assumed known value of p0 as described either tumor response or time to event end points or for joint moni- previously, but it still assumes that adjustment for known prognos- toring of efficacy and toxicity. Their methods require a substantial tic factors is sufficient to ensure comparability. Although such his- number of patients who have been treated on protocol with an ap- toric control comparisons are not considered reliable enough to propriate control regimen and who have been staged comparably eliminate the need for phase 3 trials, if done carefully, they may to the patients to be treated with the new regimen. provide an adequate basis for decisions about which new regimens Korn et al.35 developed an approach for using historic control are worthy of phase 3 evaluation. data in phase 2 multicenter trials of metastatic melanoma. They 402 Practice of Oncology

TABLE 36.3

a Simon Two-Stage Phase 2 Designs for p1 − p0 = 0.15

Optimal Design Minimax Design Reject Drug if Reject Drug if Response Rate Response Rate

P0 P1 r1/n1 r/n EN (p0)PET (p0) r1/n1 r/n EN (p0)PET (p0) 0.05 0.20 0/12 3/37 23.5 0.54 0/18 3/32 26.4 0.40 0/10 3/29 17.6 0.60 0/13 3/27 19.8 0.51 1/21 4/41 26.7 0.72 1/29 4/38 32.9 0.57 0.10 0.25 2/21 7/50 31.2 0.65 2/27 6/40 33.7 0.48 2/18 7/43 24.7 0.73 2/22 7/40 28.8 0.62 2/21 10/66 36.8 0.65 3/31 9/55 40.0 0.62 0.20 0.35 5/27 16/63 43.6 0.54 6/33 15/58 45.5 0.50 5/22 19/72 35.4 0.73 6/31 15/53 40.4 0.57 8/37 22/83 51.4 0.69 8/42 21/77 58.4 0.53 0.30 0.45 9/30 29/82 51.4 0.59 16/50 25/69 56.0 0.68 9/27 30/81 41.7 0.73 16/46 25/65 49.6 0.81 13/40 40/110 60.8 0.70 27/77 33/88 78.5 0.86 0.40 0.55 16/38 40/88 54.5 0.67 18/45 34/73 57.2 0.56 11/26 40/84 44.9 0.67 28/59 34/70 60.1 0.90 19/45 49/104 64.0 0.68 24/62 45/94 78.9 0.47 0.50 0.65 18/35 47/84 53.0 0.63 19/40 41/72 58.0 0.44 15/28 48/83 43.7 0.71 39/66 40/68 66.1 0.95 22/42 60/105 62.3 0.68 28/57 54/93 75.0 0.50 0.60 0.75 21/34 47/71 47.1 0.65 25/43 43/64 54.4 0.46 17/27 46/67 39.4 0.69 18/30 43/62 43.8 0.57 21/34 64/95 55.6 0.65 48/72 57/84 73.2 0.90 0.70 0.85 14/20 45/59 36.2 0.58 15/22 40/52 36.8 0.51 14/19 46/59 30.3 0.72 16/23 39/49 34.4 0.56 18/25 61/79 43.4 0.66 33/44 53/68 48.5 0.81 0.80 0.95 5/7 27/31 20.8 0.42 5/7 27/31 20.8 0.42 7/9 26/29 17.7 0.56 7/9 26/29 17.7 0.56 16/19 37/42 24.4 0.76 31/35 35/40 35.3 0.94 a For each value of (p0, p1), designs are given for three sets of error probabilities (α, β). The first, second, and third rows correspond to error probability limits (0.10, 0.10), (0.05, 0.20), and (0.05, 0.10), respectively. α is the probability of accepting a drug with response probability p0. β is the probability of rejecting a drug with response probability p1. For each design, EN (p0) and PET (p0) denote the expected sample size and the probability of early termination when the true response probability is p0. reviewed 42 previous phase 2 trials in melanoma conducted by the ratio of progression times. As tumors grow larger, the doubling US cancer cooperative oncology groups. They found that after time may increase and hence in some cases the chance of false- adjustment for performance status, sex, presence of visceral positive findings may be inflated.37 disease, and presence of brain metastases, there was little inter- study variability in survival among the arms of the phase 2 tri- Randomized Phase 2 Trials als. Consequently, for any single-arm phase 2 trial of metastatic melanoma, one can use their results in conjunction with the Time to tumor progression or disease-free survival has been rec- prognostic makeup of the patients in the new study to synthesize ommended for evaluation of single-agent phase 2 trials of drugs a benchmark overall survival curve or a benchmark 1-year overall that may be cytostatic and for trials adding a new drug to an active survival rate for use in evaluating the new regimen. They provide regimen. Even single-agent phase 2 trials of cytotoxics have been an example of planning a phase 2 trial using this approach that criticized on the basis that they do not provide much evidence that required 72 patients to have 85% to 90% power for detecting a 15 the drug will be able to prolong survival when incorporated into percentage-point improvement in the 1-year overall survival rate a regimen with other active drugs. Demonstrating that the regi- with a one-sided type 1 error of 10%. They found that this ap- men incorporating the new drug prolongs progression-free survival proach was less satisfactory for use with progression-free survival compared to the control regimen may provide a stronger basis for because interstudy variability remained substantial after adjust- conducting a phase 3 trial of the new regimen. ment for prognostic factors. Simon et al.12 suggested two key design differences between such Mick, Crowley, and Carroll36 proposed that the time to pro- randomized phase 2 designs and phase 3 designs. A randomized gression of a patient on a phase 2 trial be compared to the time to phase 2 design may use an end point that is a sensitive indicator of progression of the same patient on his/her previous trial. The ratio antitumor effect, although it may not be an acceptable phase 3 end of these times was called a growth modulation index, and the agent point that directly reflects patient benefit. Such an endpoint does was considered active if the index was >1.3 on average. In prac- not need to be “validated.” It is not claimed to be a valid surrogate tice, however, follow-up intervals on various protocols are differ- for survival; no regulatory approval or practice standard decisions ent, and there may be substantial variability and bias in computing should be based on the phase 2 trials using such an intermediate

tahir99 - UnitedVRG Chapter 36 Design and Analysis of Clinical Trials 403

TABLE 36.4 How large should a randomized phase 2 design comparing a new treatment to a control regimen be? Consider, for example, a Number of Patients to Treat in Single-Arm randomized phase 3 trial comparing a new regimen to a control in Phase 2 Trial Using Historic Controls and a a patient population in which the median time to progression on Binary End Point the control is 6 months and the median survival is 2 years. A 25% reduction in the hazard of death amounts to a 4-month prolonga- Proportion of tion of median survival with exponential distributions. A phase 3 Success for Historic Controls Number of Historic Controls trial with 90% statistical power for detecting this effect at a two- sided 5% significance level would require about 510 deaths (see 30 40 50 75 100 Table 36.7). With an average follow-up time of 2 years, 50% of the 0.10 94b 69 59 50 46 patients would have events and so the number of patients required 36 32 30 28 27 for randomization would be just in excess of 1,000. A randomized 21 20 19 18 18 phase 2 trial with 90% power for detecting a 33% reduction in hazard of progression corresponding to a 2-month increase in me- 0.20 – 226 126 80 67 dian progression-free survival at a one-sided 10% significance level 68 49 43 36 33 would require observing 164 progression events (Table 36.5). With 29 25 24 21 21 an average follow-up time of 2 years, >90% would have progres- 0.30 c c 307 113 86 sion events and so a sample size of 180 total randomized patients 132 69 54 41 37 would suffice. Accrual to the randomized phase 2 study could po- 36 29 26 23 21 tentially be stopped early based on futility monitoring if results are 0.40 c c c 137 95 not promising for the new regimen. The results in Table 36.5 show 267 83 59 43 37 that if an imbalanced randomization is used in which two-thirds 39 29 25 22 20 of the patients are randomized to the new treatment, the number of progression events needed increases to 185 instead of 164. So c c c 0.50 136 91 although a larger total sample size would be required, somewhat 370 80 54 38 33 fewer patients would receive the control regimen. 34 25 22 18 17 The randomized phase 2 design with control regimen has also 0.60 c c 910 104 72 been discussed by Korn et al.38 and by Rubinstein et al.39 Random- 178 56 39 28 25 ized phase 2 trials can require fewer patients than phase 3 trials, 22 17 14 12 12 but they generally require more patients than single-arm phase 2 trials. Nevertheless, they are generally necessary for evaluating a One-sided significance level of 10% and power of 80%. time to event end points or for evaluating combination regimens. PRACTICE OF ONCOLOGY b First entry is number of new patients required to detect a 15 percentage- point difference. Second and third entries are for detecting 20 percentage- Table 36.6 shows number of patients required for randomized point and 25 percentage-point differences, respectively. phase 2 trials where the primary end point is either response rate c Number of required new patients exceeds 1,000. or the proportion of patients without progression by a specified landmark time. Randomized Screening Designs end point. The purpose of the phase 2 trial is merely to determine whether to conduct a phase 3 trial that will evaluate the new regi- Phase 2 trials are generally viewed as a means of determining men with an accepted phase 3 end point The phase 2 trial may also whether a particular regimen is worthy of phase 3 evaluation. serve to optimize the regimen that might be carried forward to phase They can, however, be viewed as way to screen a wide range of 3 and to provide information about the best target population. The new regimens in order to select the most promising for phase 3 second key difference noted by Simon et al.12 is that the type I error evaluation. Traditional single-arm phase 2 designs are problematic “alpha level” for planning and analyzing the phase 2 trial can be for screening when there is substantial interstudy variation in pa- increased from the two-sided 5% level used for phase 3 trials. By let- tient selection and outcome evaluation. Simon, Wittes, and Ellen- ting this alpha level increase to a one-sided 10%, meaningful savings berg40 proposed the randomized phase 2 design in which multiple in number of patients required can be achieved. new regimens are randomized against each other as one way of

TABLE 36.5 Number of Total Events to Observe in Two-Arm Randomized Phase 2 Trial Based on Progression-Free Survival

Equal Randomization 2:1 Randomizationa Reduction Ratio of in Hazard Medians 훂 = 0.05b 훂 = 0.10b 훂 = 0.05b 훂 = 0.10b Power = 0.8 Power = 0.9 Power = 0.8 Power = 0.9 Power = 0.8 Power = 0.9 Power = 0.8 Power = 0.9 25% 1.33 301 417 219 319 339 469 246 358 30% 1.43 195 270 141 206 219 303 159 232 33% 1.5 155 215 113 164 175 242 127 185 40% 1.67 96 132 70 101 108 149 78 114 50% 2.0 52 72 38 55 59 81 43 62 a Two-thirds of patients are randomized to the new treatment group. b One-sided significance level. 404 Practice of Oncology

TABLE 36.6 Number of Patients in Each Arm of Randomized Phase 2 Trial Without Progresstion at T in Control Arm to New Treatment Arma

T-mo DFS for 5% One-Sided Significance Level 10% One-Sided Significance Level Control Group Increase in T-mo DFS Increase in T-mo DFS 0.10 0.15 0.20 0.25 0.10 0.15 0.20 0.25 0.05 129 72 48 35 99 56 38 28 0.10 176 91 58 41 133 70 45 32 0.15 216 108 66 46 163 82 51 36 0.20 250 121 73 50 188 92 56 39 0.25 278 132 79 53 208 100 60 41 0.30 300 141 83 55 224 106 63 42 0.35 315 146 85 56 235 110 65 43 0.40 324 149 86 56 243 112 65 43

DFS, disease-free survival a Eighty percent statistical power.

avoiding such interstudy variablility in prioritizing the candidate because all patients start on the new regimen, accrual rate may be regimens.40 This randomized design can provide more interpre- better with the randomized discontinuation design. table results if it also incorporates a control arm. This design is more efficient than separate randomized phase 2 trials because the Seamless Phase 2/3 Designs control arm does not have to be replicated in all of the random- Hunsberger, Zhao, and Simon47 developed a design for a seam- ized phase 2 trials. Using the example described previously, if it less phase 2/3 design. Patients are randomized between a new regi- takes 90 patients per arm to conduct a randomized phase 2 trial, men and control. An interim analysis is performed using a phase instead of 180 × 5 = 900 patients to conduct randomized phase 2 2 end point such as response rate or time to progression to decide trials of five new regimens, one would require only 90× 6 = 540 whether the results with the new treatment as sufficiently promising patients, a savings of 40%. The savings in number of patients can to continue to a phase 3 sample size. If accrual continues, then the be even more dramatic if one takes the position that the objective final analysis is performed using an acceptable phase 3 end point. is not to evaluate all five new regimens, but rather to select the best A similar approach was described by Goldmanm LeBlanc, and one and determine whether it is worthy of phase 3 evaluation. For Crowley.48 Phase 2/3 designs using Bayesian methods have been this selection objective, one does not require 90 patients per arm.40 reviewed by Thall.49 Sher and Heller50 proposed conducting phase These designs have been discussed and extended by others.41–44 3 trials with multiple experimental regimens, a control arm, and Simon et al.12 showed that one can take advantage of the non- early termination of all experimental arms that are not promising. toxic nature of some molecularly targeted drugs to efficiently eval- They used the statistical design of Schaid, Wieand, and Therneau51 uate multiple regimens in the same study. They propose using a for time to event data. Thall, Simon, and Ellenberg52 had studied factorial design in which concurrent randomizations are made for such designs when the end point was binary. A similar approach each drugs. For example, if there are three drugs (A, B, C) being was recommended by Parmar et al.53 Freidlin et al.54 have discussed evaluated, then some patients will receive all three, some will re- statistical and practical aspects of conducting clinical trials with a ceive pairs (AB, AC, or BC), some will receive single drugs (A, B, control arm and multiple new treatment arms. Freidlin, McShane, C), and one group will receive none of the drugs. In evaluating and Polley55 have also introduced a design for a randomized phase each drug, the time to progression for all patients receiving that 2 design of a new drug with a candidate predictive biomarker for drug are compared to the times for all patients not receiving that determining whether the drug is entirely inactive, active only in the drug. The trial can be sized as if it were a single two-arm trial. marker positive group, or active regardless of the biomarker status. The design is effective as long as there are not negative interac- This design enables investigators to appropriately plan whether to tions among drugs. Negative interactions would result from the continue biomarker development into phase 3 development. toxicity of one drug interfering with the full-dose administration of other drugs, which may not be a problem for many molecularly targeted drugs. The design is also useful for attempting to identify DESIGN OF PHASE 3 CLINICAL TRIALS combinations that are therapeutically synergistic, a circumstance of particular importance with molecularly targeted drugs. Good therapeutic research requires asking important questions Rosner, Stadler, and Ratain45 describe a “randomized discon- and getting reliable answers. The most important clinical tri- tinuation design” for phase 2 studies of therapeutically targeted als are often the most difficult to conduct.56 They may involve drugs. All eligible patients are started on the drug and given two to withholding a treatment established by tradition, transferring four courses of treatment. Patients are then evaluated: Those with patient management responsibility across specialties, standardiz- progression are removed from study, those with objective tumor ing procedures among physicians, and sharing recognition with a response are continued on treatment, and the remaining patients large group of collaborators. are randomized to either continue or discontinue the drug. The continued and discontinued groups of randomized patients are compared with regard to time to progression. Freidlin and Simon46 End Points evaluated and further developed this design. It may require as large a number of patients started on treatment as a straightforward Phase 3 trials attempt to provide guidance to practicing physi- randomized phase 2 design. The advantage of the design is that cians to help them make treatment decisions with their patients.

tahir99 - UnitedVRG Chapter 36 Design and Analysis of Clinical Trials 405

Consequently, the trials should provide reliable information con- Clinical trials can be conducted with fewer patients if patients cerning end points of relevance to the patients. The major end are selected based on assays that identify the tumors likely to be points for evaluating the effectiveness of a treatment should be di- sensitive to the drug in question. Simon and Maitournam67,68 and rect measures of patient welfare. Survival and symptom control are others69,70 have evaluated the efficiency of such targeted designs. two such end points. The latter is not routinely used because of the When fewer than half of the patients “test positive” and when the difficulty of measuring it reliably and because it may be influenced new treatment has little benefit for patients who test negative, the by concomitant treatments. required sample size can be dramatically reduced by restricting el- Although durable complete regression of metastatic disease igibility to patients who test positive. Simon and Zhao have made is usually a good surrogate for prolonged survival, partial tumor available a web-based computer program to enable investigators to shrinkage usually is not an appropriate end point for phase 3 trials. compare such designs to standard broad eligibility designs (http:// Torri et al.57 performed a meta-analysis of the relationship between linus.nci.nih.gov/brb). difference in response rates and difference in median survivals for This targeted approach was effectively used for the development randomized clinical trials of advanced ovarian carcinoma. They of trastuzumab in patients with metastatic breast cancer. In that case, found that large improvements in response rates corresponded to about 450 patients whose tumors overexpressed HER-2 participated very small improvements in median survival. Hence, use of re- in a randomized clinical trial that provided convincing evidence sponse rate as an end point may result in giving patients increas- that trastuzumab prolonged survival. Had the study been conducted ingly intensive and toxic therapy with little or no net benefit to without evaluating HER-2 expression, >8,000 patients would have them. Proper validation of an end point as a surrogate for clini- been needed for similar statistical power.67 Even had a huge study cal benefit requires a series of randomized clinical trials in which of unselected patients been conducted and given a statistically sig- treatment differences with regard to the candidate surrogate are nificant result, the size of the benefit would have been very small as related to treatment differences with regard to clinical benefit.58–60 the benefit in the 25% of patients with HER-2 overexpression would It is not sufficient to show that clinical outcome is related to the have been diluted by lack of benefit from the remaining 75%. It is candidate surrogate measured on the same treatment arm as this questionable whether such a small benefit overall would have justi- may just reflect the known responder versus nonresponder bias. fied approval or use of a drug with clear and serious toxicities. Disease-free survival is often accepted as an important measure In many cases where the biologic credentials of a predictive of clinical benefit to be used as an end point for adjuvant treatment biomarker are less compelling, one will want to include patients trials. There is more controversy, however, about the use of time who are both marker positive and marker negative but to require to progression in metastatic disease trials. The controversy relates that all patients have the marker evaluated, to size the trial so that to whether prolonged time to progression provides clinical ben- there is adequate statistical power for evaluating treatment effect efit and whether it can be measured without bias. With unblinded separately in the patients who are marker positive and to use a mul- evaluation of time to progression, there could be a reduced thresh- tiple testing method that ensures that the study-wise type I error

old for declaring progression for control patients so that they can level does not exceed 5%. Simon and others have described “all PRACTICE OF ONCOLOGY cross over to the new treatment.61 Central party blinded review comers” designs of this type.71–74 Zhao and Simon provide web- of progression is often used to avoid such potential bias. Because based computer programs at http://linus.nci.nih.gov/brb to facili- the review is not performed in real-time, however, it can intro- tate use of such designs in clinical trials. Freidlin, McShane, and duce additional biases of “informative censoring.” Freidlin et al.62 Korn75 have described the use of biomarker designs in cancer clini- proposed an approach to adjusting for increased surveillance of cal trials. These enrichment and all-comers designs presume that a the control group. Dodd et al.63 proposed that central review be single predictive biomarker with an analytically validated test and performed only for a subset of patients to evaluate whether local a threshold of positivity has been developed prior to the start of the assessments were biased, not to replace local assessments. phase 3 clinical trial. Because of the complexity of cancer biol- ogy, this is not always possible. Several adaptive designs have been Patient Eligibility developed to enable some aspects of biomarker specification to be included in the phase 3 trial while also rigorously evaluating the statistical significance of the treatment effect in the “biomarker To ensure that the results of phase 3 trials are applicable to patients positive” population. The “adaptive threshold design” avoids the seen in the community outside of clinical research settings, the tri- requirement that the threshold of positivity be prespecified,76 als often involve numerous centers and extensive community par- and the “adaptive signature design” enables multiple candidate ticipation. In order to ensure broad generalization of conclusions, biomarkers to be evaluated.77,78 Hong and Simon79 developed a most multicenter phase 3 trials have employed broad eligibility cri- run-in design that permits a pharmacodynamic, immunologic, or teria. In the United Kingdom, many trials have been designed using intermediate response end point measured after a short run-in pe- the uncertainty principle, an approach that leaves much of the deci- riod on the new treatment to be used as the predictive biomarker. sion making about eligibility to the treating physician. There may Simon, Paik, and Hayes80 described a prospective-retrospective ap- be guidelines for eligibility, but the ultimate decision is made by the proach to using archived tumor specimens for a focused re-analysis treating physician; if he or she is uncertain about which treatment is of a randomized phase 3 trial with regard to a predictive biomarker. more appropriate for the patient, the patient is eligible. The approach requires that archived specimens be available on There is a growing recognition, however, that the one of the most patients, and that an analysis plan focused on a single marker key hallmarks of cancer is intertumor heterogeneity. Tumors that be developed prior to performing the blinded assays. This ap- arise in the same primary site are often quite different with re- proach was used in establishing that a K-RAS mutation was a nega- gard to their oncogenesis, pathophysiology, and drug sensitivity. tive predictive biomarker for response of patients with colorectal Consequently, conducting broad eligibility clinical trials with cancer to antiepidermal growth factor receptor antibodies. drugs only expected to be effective for an identifiable subset of pa- tients is often no longer an appropriate research strategy.64–66 Par- ticularly with molecularly targeted drugs, effectiveness is likely to Randomization be limited to a sensitive subset of tumors that may be characterized based on whether the molecular target of the drug is deregulated To determine whether a new treatment cures any patients with in the tumor. Even with cytotoxics, many patients are generally a disease that is uniformly and rapidly fatal, history is a satisfac- treated for each patient who benefits. The high costs of many mo- tory control. Once we leave this setting of complete determinism, lecularly targeted drugs make the traditional broad eligibility trial however, the definition of an adequate nonrandomized control approach increasingly unsustainable. group becomes problematic. In comparing outcomes for patients 406 Practice of Oncology receiving two different treatments when the treatment assignment It is generally best to limit stratification to those factors definitely was not randomized, often diagnostic and staging procedures and known to have important independent effects on outcome. If two other factors used explicitly or implicitly for selecting patients for factors are closely correlated, only one needs to be included. Many each treatment are different. Some of these factors may be difficult clinical trialists believe that stratification is an unnecessary compli- to quantify but strongly prognostic. Some of the factors may be cation because adjustment for imbalances of known factors can be previously unsuspected as being of prognostic importance. Also, made in the analysis and has negligible effect on statistical power. supportive care, secondary treatments, and methods of evaluation Stratification may help to ensure balance for interim analyses when and follow-up may be different for the two treatment groups. In the sample sizes may be limited and provides the medical audience comparisons of a new treatment to a historic control, patients re- with confidence in the results, which often is not available when ceiving the new treatment are often much more highly selected depending on complex adjustment methods to deal with prognostic than the control group. Often, there is inadequate information to imbalances. Stratification also is a convenient way of specifying a determine whether prognostic differences are present, and current priori what are considered the important prognostic factors. known prognostic factors may not have been measured for the con- Many clinical trials use dynamic stratification methods. The trols. It generally is difficult or impossible to determine whether most popular such method is that conceived by Pocock and the controls would have been eligible for the current study and in Simon,81 which permits effective balancing with regard to many what way they represent a selection of all eligible patients. prognostic factors. There has been some concern about the ef- Formation of the control group by random treatment assign- fect of adaptive stratification on analysis of treatment differences. ment as an integral part of the planned study can avoid most of the Multiple studies, including those by Kalish and Begg,82 have systematic biases just mentioned. Randomization does not ensure demonstrated that if the stratification factors are included in the that the study will include a representative sample of all patients model used for final analysis, the effect of adaptive stratification with the disease, but it does help to ensure an unbiased evaluation is to make the true type I error less than the nominal rate, hence of the relative merits of the two treatments for the types of patients the analyses are slightly conservative. Simon and Simon83 showed entered. that model-based analyses are not necessary to use with adaptive It is sometimes said that randomization is unnecessary because stratification methods. One can define a linear test statistic that matched historic or concurrent controls can be selected. However, reflects the treatment effect difference on the outcome adjusted matching can be done only with regard to known prognostic fac- for the stratification variables, and generate the null distribution of tors, and those factors often do not account for enough of the varia- the test statistic by reapplying the adaptive stratification method. tion in patient outcome to assure that an unbiased historic control The Pocock-Simon81 method of adaptively stratified treatment as- group can be constructed. It also is sometimes said that randomiza- signment is not deterministic. Consequently, one can replicate the tion is not effective in ensuring that the treatment groups are simi- stratified treatment assignments, holding fixed the order of patient lar with regard to unknown prognostic factors unless the number registrations and the stratification variables of the patients, recom- of patients is large. This is true but reflects a misunderstanding pute the value of the test statistic for the rerandomized treatment of the purpose of randomization. Randomization does not ensure assignment, repeat this process a thousand times, and thereby that the groups are medically equivalent, but it distributes the un- generate the null distribution of the test statistic. Consequently, known biasing factors according to a known random distribution although the use of adaptive stratification methods, like the use of so that their effects can be rigorously allowed for in significance all stratification methods, are not essential, the criticisms of their tests and confidence intervals. This is true regardless of the study effect on final analyses are unjustified. size. A significance level represents the probability that differences in outcome can be the result of random fluctuations. Without a Sample Size randomized treatment allocation, a “statistically significant differ- ence” may be the result of a nonrandom difference in the distribu- The protocol for a phase 3 trial should specify the number of pa- tion of unknown prognostic factors. tients to be accrued and the duration of follow-up after the close of In many cases, there is a role for both randomized and nonran- accrual when the final analysis will be performed. Methods of sam- domized trials in drug development. The nonrandomized format ple size planning are usually based on the assumption that at the can in some cases be used for determining which regimens are conclusion of the follow-up period, a statistical significance test will sufficiently promising for randomized phase 3 evaluation and in be performed comparing the experimental treatment to the control clinical settings in which outcome is uniformly poor. For major treatment with regard to a single primary end point. A statistical questions of public health importance, unless the expected treat- significance level of 0.05 means that if there is no true difference ment effect on outcome is very large, the need for reliable answers in treatment effectiveness, the probability of obtaining a difference dictates the use of randomized phase 3 trials. in outcomes as extreme as that observed in the data is 0.05. The Randomization of a patient should be performed after the pa- significance level does not represent the probability that the null tient has been found eligible and has consented to participate in the hypothesis is true; it represents a probability of an observed differ- trial and to accept either of the randomized options. A truly random ence, assuming that the null hypothesis is true. Conventional sta- and nondecipherable randomization procedure should be used tistical theory ascribes no probabilities to hypotheses, only to data. and implemented by calling a central randomization office staffed A one-sided significance level represents the probability, by by individuals who are independent of participating physicians. chance alone, of obtaining a difference as large as and in the same direction as that actually observed. A two-sided significance level Stratification represents the probability of obtaining by chance a difference in either direction as large in absolute magnitude as that actually When important prognostic factors are known for patients in a observed. The two-sided significance level is usually twice the one- randomized trial, it is often advisable to stratify the randomiza- sided significance level. Controversy exists over theappropriateness tion to ensure equal distribution of these factors. This is usually of one-sided or two-sided significance levels. Although this is a accomplished by preparing a separate randomization list for each somewhat trivial issue, a two-sided significance level of 0.05 has stratum of patients. Each list must be balanced so that after each become widely accepted as a standard level of evidence. block of 4 to 10 patients within the stratum, the treatment groups The probability of obtaining a statistically significant result contain equal numbers of patients. Within the blocks, the se- when the treatments differ in effectiveness is called the power of quence of treatment assignments is random. The stratification fac- the trial. As the sample size and extent of follow-up increases, the tors must be known for each patient at the time of randomization. power increases. The power depends critically, however, on the

tahir99 - UnitedVRG Chapter 36 Design and Analysis of Clinical Trials 407

TABLE 36.7 treatment. For exponential distributions, the percentage reduction in hazard of death can be expressed as a ratio of median surviv- Number of Events Needed for Comparing als, which is displayed in the second column of Table 36.7. When Survival Curves the primary end point is overall survival, the events are deaths; for disease-free survival curves, events are deaths or recurrences. The Ratio of Median Percentage Survival for Number of translation of the number of deaths or events required to the num- Reduction in Exponential Total Deaths to ber of patients required depends on the actual shape of the survival Hazard of Death Distributions Observea distributions, the rate of accrual, and the duration of follow-up after close of accrual. Generally, however, it is best to specify the time of 25 1.33 508 the final analysis as the time when the specified number of deaths 30 1.43 330 or events is obtained—not in terms of absolute calendar time. In some cases, it may be convenient to think in terms of the 33 1.50 257 proportion of patients without progression or death beyond some 40 1.67 162 landmark time, such as 5 years. Tables 36.8 and 36.9 provide re- 50 2.0 88 quired numbers of patients for clinical trials planned on this basis. This approach is less flexible for studies in which survival or disease- a Total number of deaths in both groups to have power = 0.90 for detecting free survival is the end point, as it presumes that all patients will be ratio of median survival. Type I error α = 0.05 (two-sided). followed for the landmark time as a minimum. These tables can, however, be used generally for detecting differences in a binary end point, denoted success rate in the tables. For comparing treatments in phase 3 trials, differences of >15 to 20 percentage-points usually size of the true difference in effectiveness of the two treatments. are considered unrealistic. Establishing a sample size that provides Generally, one sizes the trial so that the power is either 0.80 or 0.90 good statistical power for detecting realistically expected treatment when the true difference in effectiveness is the smallest size that is improvements is important. Many published “negative” results are considered medically important to detect. actually uninterpretable because the sample sizes are too small.85 Statisticians have developed useful methods for planning sam- ple size to compare survival curves or disease-free survival curves in phase 3 trials. Table 36.7 shows the number of total events needed FACTORIAL DESIGNS assuming that the hazard ratio—the ratio of forces of mortality for the two treatment groups—is constant over time.84 Table 36.7 The 2K factorial design was described in the section on random- shows the total number of events that must occur in a given co- ized phase 2 trials, but it can also be used for phase 3 trials. The

hort to provide 90% power for detecting a specified reduction in two-by-two factorial design is the version most often used. There PRACTICE OF ONCOLOGY the hazard for the experimental treatment relative to the control are four treatment groups: one receiving neither of the two drugs

TABLE 36.8 Number of Patients in Each of Two Treatment Groups to Compare Proportions (One-Sided Test)

Larger Minus Smaller Success Rate Smaller Success Rate 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.05 512a 172 94 62 45 35 28 23 19 16 381b 129 72 48 35 27 22 18 15 13 0.10 786 236 121 76 54 40 31 25 21 17 579 176 91 58 41 31 24 20 16 14 0.15 1,026 292 144 88 60 44 34 27 22 18 752 216 108 66 46 34 26 21 17 14 0.20 1,231 339 163 98 66 48 36 29 23 19 900 250 121 73 50 37 28 22 18 15 0.25 1,402 377 178 105 70 50 38 29 23 19 1,024 278 132 79 53 38 29 23 18 15 0.30 1,539 407 189 111 73 52 38 30 23 19 1,122 300 141 83 55 39 30 23 18 15 0.35 1,642 429 197 114 74 52 38 29 23 18 1,196 315 146 85 56 40 30 23 18 14 0.40 1,711 441 201 115 74 52 38 29 22 17 1,246 324 149 86 56 39 29 22 17 14 0.45 1,745 446 201 114 73 50 36 27 21 16 1,271 327 149 85 55 38 28 21 16 13 0.50 1,745 441 197 111 70 48 34 25 19 15 1,271 324 146 83 53 37 26 20 15 12 a Upper figure: significance level = 0.05, power = 0.90. b Lower figure: significance level = 0.05, power = 0.80. 408 Practice of Oncology

TABLE 36.9 Number of Patients in Each of Two Treatment Groups to Compare Proportions (Two-Sided Test)

Larger Minus Smaller Success Rate Smaller Success Rate 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.05 620a 206 113 74 54 42 33 27 23 19 473b 159 88 58 43 33 27 22 18 16 0.10 956 285 146 92 64 48 38 30 25 21 724 218 112 71 50 38 30 24 20 17 0.15 1,250 354 174 106 73 53 41 33 26 22 944 269 133 82 57 42 32 26 21 18 0.20 1,502 411 197 118 79 57 44 34 27 22 1,132 313 151 91 62 45 34 27 22 18 0.25 1,712 459 216 127 84 60 45 35 28 23 1,289 348 165 98 65 47 36 28 22 18 0.30 1,880 495 230 134 88 62 46 36 28 22 1,414 375 175 103 68 48 36 28 22 18 0.35 2,006 522 239 138 89 63 46 35 27 22 1,509 395 182 106 69 49 36 28 22 18 0.40 2,090 537 244 139 89 62 45 34 26 21 1,571 407 186 107 69 48 36 27 21 17 0.45 2,132 543 244 138 88 60 44 33 25 19 1,603 411 186 106 68 47 34 26 20 16 0.50 2,132 537 239 134 84 57 41 30 23 17 1,603 407 182 103 65 45 32 24 18 14 a Upper figure: significance level = 0.05, power = 0.90. b Lower figure: significance level= 0.05, power = 0.80.

A or B, one receiving A, one receiving B, and one receiving both. For such trials, the secondary benefits of the new regimen, while Although there are four treatment groups, the average effect of important, is not worth reductions in effectiveness in the primary each treatment factor can be evaluated using all of the patients. end point. Unfortunately, it is not possible to establish that the two To evaluate the effect of A, you compare outcomes for patients treatments are completely equivalent with regard to the primary receiving A to outcomes for those not receiving A, ignoring B. Usu- end point. The usual approach is to plan the trial to have high ally, the sample size for a two-by-two factorial trial is computed statistical power for detecting small reductions in effectiveness, assuming that there is no interaction between the effects of the two and this requires a large sample size. Because failure to reject the drugs. The sample size is approximately the same as for a simple standard null hypothesis of no treatment difference results in adop- two-arm trial. The factorial design offers the possibility of answer- tion of a new, and potentially inferior, regimen, misinterpretation ing two questions for the cost of one, but there is a risk of ambiguity of the results of noninferiority trials can result in serious problems. in the interpretation of results.86 For situations in which negative For the analysis of such trials, confidence intervals rather than sta- interactions are unlikely or in which it is unlikely that both factors tistical significance tests should be emphasized.88 The confidence will have substantial effects, the factorial design can provide a sub- interval for the true difference of effectiveness gives a much clearer stantial improvement in the efficiency of clinical trials. picture of which differences are consistent with the data. Makuch Simon and Freedman87 developed a Bayesian method for the and Simon89 and Durrleman and Simon90 discuss this approach design and analysis of factorial trials. Their approach avoids the for planning and monitoring therapeutic equivalence trials. need to dichotomize one’s assumptions that interactions either do Noninferiority trials are generally planned to distinguish the or do not exist, and provides a flexible approach to the design and null hypothesis that the treatments are equivalent from the alter- analysis of such clinical trials. The Bayesian approach also avoids a native that the new treatment is inferior by an amount δ. One of preliminary test of interaction; such tests have poor power and bas- the key problems in designing a noninferiority trial is specification ing the analysis on such tests is problematic. The Bayesian model of δ. A small value of δ leads to a large trial. A large value of δ can suggests that in planning a factorial trial in which interactions are lead to a small but meaningless trial. The reduction in effective- unlikely but cannot be excluded, the sample size should be in- ness that the trial will be able to detect should be some fraction of creased by approximately 30%, as compared to a simple two-arm the effectiveness of the standard treatment. For example, suppose clinical trial for detecting the same size of treatment effect. The the standard treatment is 12 months of a chemotherapy regimen 30% figure allows for a 5% prior probability of a medically impor- that increases 5-year survival by 10 percentage-points relative to no tant, qualitative interaction between the treatment effects. chemotherapy and the new regimen of interest is use of the same regimen for only 6 months. If we want to have high power for de- tecting a reduction in effectiveness by half, then δ should represent Noninferiority Trials a difference of 5 percentage-points in 5-year survival. If we want high power for detecting a reduction in effectiveness by one-quar- Noninferiority trials often compare a standard treatment to a less ter, then δ should represent a difference of 2.5 percentage-points invasive or more convenient therapy that is not expected to be supe- in 5-year survival. An appropriate value of δ can only be deter- rior to the standard treatment with regard to the primary end point. mined based on a careful review of the studies that established the

tahir99 - UnitedVRG Chapter 36 Design and Analysis of Clinical Trials 409

effectiveness of the standard treatment. If those studies do not exist previous experience. Spiegelhalter, Freedman, and Parmar92 have or are not adequate, a noninferiority trial may not be appropriate. suggested analysis of a clinical trial with regard to both an “enthu- Another problem in the design of noninferiority trials is the lack siastic” prior and a “skeptical” prior. The former might be held by of internal validation of the assumption that the control treatment a developer of the treatment and the later by a regulator. Robust is actually effective for the patient population at hand. If the effec- conclusions are obtained when the data is so extensive and strong tiveness of the standard treatment is highly variable among studies, that the posterior distributions are little changed regardless of there is the risk that a new regimen will be found noninferior to the whether you use an enthusiastic or skeptical prior. Unfortunately, standard because the standard is not effective in the current study. such robustness generally requires a very large sample size, much Consequently, noninferiority trials are only appropriate when the larger than indicated by use of standard frequentist methods. For standard regimen is highly and reproducibly effective. some parameters, there may be a consensus prior distribution. None of the conventional frequentist approaches to the design For example, for evaluating cytotoxics there was generally broad and analysis of therapeutic equivalence trials satisfactorily account consensus that large treatment-effect by patient-subset interactions for the uncertainty in estimation of the effectiveness of the standard were unlikely, and Simon93 used this in a Bayesian approach to treatment. Simon91 developed a Bayesian approach that addresses subset analysis.Generally, however, there is no meaningful prior this problem. The effectiveness of the control treatment C relative consensus about the effect of treatment. Randomized clinical tri- to the previous standard (P) is represented by a parameter β. Infor- als are done because the opinions of experts are often wrong. The mation about the effectiveness of C relative to P is summarized by subjective nature of the prior distribution is problematic for the a prior distribution, which is normal with mean μβ and standard interpretation of phase 3 clinical trials. deviation σβ. These values are obtained from a random-effects There are several important misconceptions about the use of meta-analysis of the previously conducted randomized trials com- Bayesian methods for clinical trials. First, some people believe that paring C to P. The result of the noninferiority trial is summarized Bayesian methods provide an adequate alternative to randomized by an estimate δˆ with standard error of the effectiveness of test regi- treatment assignment. In fact, however, randomization is just as men E relative to C. With some simplifying assumptions, the pos- important for the validity of Bayesian methods as for frequentist terior distribution of the effectiveness of E relative to P is a normal methods.94 Second, some people mistakenly believe that Bayes- 2 2 90 distribution with mean μβ + δˆ and variance σβ + σ . Simon also ian clinical trials require fewer patients than frequentist trials. shows how the sample size of the therapeutic equivalence trial may Bayesian sample size calculations depend on the prior distribution be planned and how the size depends critically on the strength and used. Using skeptical priors, the sample size needed with Bayes- consistency of the evidence that the active control C is superior to ian methods may be much larger than the conventional sample P and on the size of that difference in effectiveness. size. Third, some statisticians believe that the main impediment to use of Bayesian methods in clinical trials has been the difficulty of computing posterior distributions. The main limitation has, how-

Bayesian Methods ever, been the fact that subjectivity of analysis is problematic for OF ONCOLOGY PRACTICE phase 3 clinical trials. Conventional statistical methods (i.e., frequentist method) regard Bayesian methods can be very useful for phase 1 and phase 2 the data collected in an experiment as being random; they test hy- trials. For such trials, the prior distribution need only be appro- potheses about parameters that represent fixed but unknown treat- priate for the investigator or sponsor. For phase 3 trials, the situa- ment effects. For example, frequentist methods derive probability tion is more complex. Bayesian methods are applicable to phase 3 statements about differences in observed response rates under an problems in which a concensus prior is appropriate. Such priors assumed null hypothesis that the true response probabilities are are possible for parameters representing interaction effects,93 for equal. Bayesian statistical methods consider the parameters, as well the effectiveness of active controls in noninferiority trials,91 and for as the data, as being random and selected from prior distributions. unexpected findings with multiple safety end points.95 As indicated What does the assumption that the true treatment effect is a ran- previously, however, subjective opinion of the investigator or spon- dom draw from a prior distribution mean? One interpretation is sor should have no role in testing the primary hypothesis of whether that we regard the prior distribution as expressing our subjective the new treatment is better than the control for the prespecified beliefs about the value of the treatment effect based on previous target population. Once the basic effectiveness of the treatment is experience with this treatment and other similar treatments. Such established, however, there are many other analyses that can help subjective prior distributions would vary among individuals based physicians decide how to use the new treatment. Those analyses on their experience, biases, circumstances, and perhaps economic generally cannot be answered as precisely or with as little chance interests. Bayesian methods use Bayes’ theorem to update the prior of error as the testing of the primary null hypothesis. Different phy- distributions of the parameters based on data from the study to sicians may have varying prior beliefs about the treatment and how produce the posterior distributions of the parameters. Using the its effectiveness might vary among patients; Bayesian methods may posterior distributions, hypotheses about whether the treatments be useful for physicians in determining how to implement the re- are equivalent can be tested. Consequently, Bayesian methods can sults in the context of the patients they see. One must recognize, derive direct probability statements about the parameters, such as however, that Bayesian models can be overfit to data like any other “the probability that the treatment effect is 0.04.” The probability models and can produce poor predictions. statements about the parameters seem to tell us what we want to know, but the results may depend as much on our prior distribu- tions as on the data. ANALYSIS OF PHASE 3 CLINICAL TRIALS Many Bayesian statisticians use “noninformative” prior distri- butions. For example, a noninformative prior distribution for the difference in response probabilities might be constant for all differ- Intention-to-Treat Analysis ences between −1 and +1. That noninformative prior represents the belief that huge differences are just as likely as small differ- The intention-to-treat principle indicates that all randomized ences, positive differences as likely as negative differences. Conse- patients should be included in the primary analysis of the trial. quently, methods based on apparently innocuous noninformative For cancer trials, this has often been interpreted to mean all “eli- prior distributions may not be appropriate for real-world studies. gible” randomized patients. Excluding patients from analysis be- In some cases, the treatment being evaluated can be considered cause of treatment deviations, early death, or patient withdrawal “exchangeable” with other treatments that have been previously can severely distort the results.96–98 Often, excluded patients have evaluated and a prior distribution can be defined based on that poorer outcomes than do those who are not excluded. Investigators 410 Practice of Oncology

frequently rationalize that the poor outcome experienced by a pa- TABLE 36.10 tient was due to lack of compliance to treatment, but the direction of causality may be the reverse. For example, in the Coronary Drug Nominal Two-Sided Significance Levels for Project, the 5-year mortality for poor adherents to the placebo regi- Early Stopping in Interim Monitoring Methods men was 28.3%, significantly greater than the 15.1% experienced That Maintain an Overall Type I Error Level by good adherents to the placebo regimen.99 In randomized trials, of 0.05 there may be poorer compliance in one treatment group than the other, or the reasons for poor compliance may differ. Exclud- Analysis O’Brien and Fleming Number Pocock125 Haybittle103 Fleming104 et al.105 ing patients, or analyzing them separately (which is equivalent to excluding them), for reasons other than eligibility is generally 1 0.016 0.0027 0.00001 0.0051 considered unacceptable. The intention-to-treat analysis with all 2 0.016 0.0027 0.0013 0.0061 eligible randomized patients should be the primary analysis. If the conclusions of a study depend on exclusions, these conclusions 3 0.016 0.0027 0.008 0.0073 are suspect. The treatment plan should be viewed as a policy to be 4 0.016 0.0027 0.023 0.0089 evaluated. The treatment intended cannot be delivered uniformly Final 0.016 0.049 0.041 0.0402 to all patients, but all eligible patients should generally be evalu- able in phase 3 trials.

Interim Analyses cancer cooperative groups with interim analysis of phase 3 clinical trials was reviewed by Korn, Freidlin, and Mooney.106 If statistical significance tests are performed repeatedly, the Extreme treatment differences at an interim analysis are less probability that the difference in outcomes will be found to be usual in cancer clinical trials than finding that interim results do statistically significant (at the 0.05 level) at some point may be not support the hypothesis that the experimental treatment is sub- considerably >5%. This probability is called the type I error of the stantially better than the control. Futility analyses are important in 100 analysis plan. Fleming, Green, and Harrington have shown that order to avoid exposing patients to a more toxic and debilitating the type I error can be as great as 26% if a statistical significance new treatment E once the essential outcome of the trial is well as- test is performed every 3 months of a 3-year trial that compares sured.107 Data-monitoring committees are charged with helping to two identical treatments. Some trials are published without stating make these difficult judgments. A variety of statistical approaches the target sample size, without indicating whether a target sample to “futility monitoring” have been developed.108 Goldman, size was stated in the protocol, and without describing whether the LeBlanc, and Crowley48 have shown that futility analyses based on published analysis represented a planned final analysis or was one intermediate end points like disease-free survival can be particu- of multiple analyses performed during the course of the trial. In larly effective even in trials where the primary endpoint is survival. such cases, one must suspect that the investigators were not aware The method of stochastic curtailment109 is widely used for of good statistical practices and of the dangers of informal multiple “ futility analyses.” At any interim analysis, the probability of reject- analyses. ing the null hypothesis at the end of the trial is computed. This Interim analyses can be very misleading and interfere with probability is calculated as being conditional on the data already a physician’s attempt to state honestly to the patient that there obtained and on the assumption that the alternative hypothesis of is no reliable evidence indicating that one treatment option or superiority of the experimental treatment used initially in plan- the other is preferable. Consequently, it has become standard ning the sample size for the trial is true. If this conditional power in phase 3 multicenter clinical trials to have a data-monitoring is less than approximately 0.20, then the trial may be terminated committee review interim results, rather than having the moni- with acceptance of the null hypothesis. The 0.20 cutoff can be toring done by participating physicians. This approach helps to raised substantially to at least 0.40 if this type of interim analy- protect patients by having interim results carefully evaluated by sis is performed only a few times during the course of the trial. an experienced group of individuals and helps to protect the study With stochastic curtailment, interim analyses need not be equally from damage that ensues from misinterpretation of interim re- spaced, and the number of interim analyses need not be specified 101,102 sults. Generally, interim outcome information is available in advance. to only the data-monitoring committee. The study leaders are not part of the data-monitoring committee, because they may have a perceived conflict of interest in continuing the trial. The data- Significance Levels, Hypothesis Tests, and monitoring committee determines when results are mature and Confidence Intervals should be released. These procedures are used only for phase 3 trials. Medical decision making is complicated, and clinicians frequently A number of useful statistical designs have been developed for misinterpret statistical significance tests in search of clear-cut monitoring interim results. The simplest is due to Haybittle.103 answers from ambiguous data. A statistical significance level for Interim differences are discounted unless the difference is statisti- comparing outcomes represents the probability of obtaining a dif- cally significant at the two-sidedp <0.0025 level. If the interim ference as large as that actually observed if the treatments were differences are not significant at that level, the trial continues actually of equal efficacy and differences occur merely by chance. until its originally intended size. The final analysis is performed After significance tests had been used for many years, Neyman and without regard to the interim analyses, and the type I error is al- Pearson formalized a mathematical theory of hypothesis testing. In most unaffected by the monitoring. Many others have developed this theory, a study must prespecify a null hypothesis, an alterna- group-sequential methods for interim monitoring based on a pre- tive hypothesis, and a decision rule for accepting one hypothesis specified number of planned interim analyses. One of the most and rejecting the other based on the data obtained. The theory has commonly used methods is that of O’Brien and Fleming.104 The appealed to clinicians because it simplifies complex medical deci- critical p value for determining whether an interim difference sion making by providing yes or no answers: either the difference should be judged statistically significant depends on the number is statistically significant or it is not. The distinction between one- of analyses that will be performed during the trial. For a five-stage and two-sided decision rules becomes crucial because a one-sided trial—four interim analyses and one final analysis—the critical p = 0.05 is simply nonsignificant if a type I error of 0.05 based on p values are shown in Table 36.10.105 The experience of the US a two-sided decision rule is prespecified.

tahir99 - UnitedVRG Chapter 36 Design and Analysis of Clinical Trials 411

The concept of prespecification of hypotheses is important for more effort than the life table method when the number of pa- medical experimentation. However, the accept–reject nomen- tients is large. clature of the Neyman-Pearson theory provides an oversimplified The first step in the application of either method is the calcula- and sometimes misleading interpretation of the data. Significance tion of survival time for all patients. Survival is the duration from levels can serve as useful aids to interpretation of results, but quib- the chosen baseline (e.g., date of randomization) until death or date bling about whether a one-sided p = 0.04 is significant makes little last known to be alive for patients who are not known to have died. sense. Significance levels are influenced by sample sizes, and fail- To use the life table method, intervals for the grouping of survival ure to reject the null hypothesis does not mean that the treatments times are determined. The life table, shown in Table 36.11, is then are equivalent. There is no simple index of truth for interpreting filled out. This sample life table is prepared with yearly intervals in results. Some attempt to use the notion of statistical significance the first column. The number of patients alive at the beginning of in this way, but thorough presentation, skeptical evaluation, and the interval is entered in column 2. The number who died in the cautious interpretation of results always are required. interval is entered in column 4. Patients dying exactly at a time that Confidence intervals are generally much more informative represents a boundary between two intervals (e.g., 365 days) are than are significance levels. A confidence interval for the size of considered to have died in the preceding interval (e.g., 0 to 1 year). the treatment difference provides a range of effects consistent with Column 3 contains the number of patients who are lost to follow- the data. The significance level tells nothing about the size of the up during the interval or who are alive with maximum follow-up treatment effect because it depends on the sample size. However, duration included in the interval. These latter patients are referred it is the size of the treatment effect, as communicated by a con- to as withdrawn alive in the conventional life table terminology. fidence interval, that should be used in weighing the costs and The life table method assumes that patients lost to follow-up or benefits of clinical decision making. Many so-called negative re- withdrawn alive during the interval are at risk of death for one-half sults are actually noninformative, and confidence intervals help of the interval. Hence, column 5—the number alive at the start of to determine when this is the case. Simon88 has presented a non- the interval minus half the number lost or withdrawn during the technical discussion of how to calculate confidence intervals for interval—represents an approximate number of patients at risk of treatment differences with the types of end points commonly used death during the interval. Column 6 gives the ratio of the number in cancer clinical trials. of patients who died during the interval to the number at risk during the interval. Column 7 gives the estimated probability of surviving the interval for patients alive at the start of the interval. Calculation of Survival Curves Column 8 should be studied carefully, because it provides the life table estimate of the survival distribution and indicates the Most cancer clinical trials display results by showing survival curves logic behind the method. The probability of surviving >3 years or disease-free survival curves. Survival curves display the probabil- after randomization, for example, equals the entry in the third row

ity of surviving beyond any specified time, with time shown on the of column 8 (0.50). The logic is as follows: To survive 3 full years, PRACTICE OF ONCOLOGY horizontal axis. In disease-free survival curves, it is the time until the patients must survive through the first year; and given that they recurrence or death that is shown. Other time-to-event distribu- have survived the first year, they must survive the second year; and tions can be similarly represented using the same methods. The given that they have survived the second year, they must survive usual statistical methods are not appropriate for analyzing survival the third year. Consequently, the probability of surviving for at because they ignore the fact that surviving patients have a limited least 3 years is estimated by the product p1 × p2 × p3 of factors follow-up period after which their survivals are “censored.” in column 7. By using this product, the life table method takes The most satisfactory way of representing such data is to esti- maximal advantage of the mortality experience of patients with mate the survival function S(t). This function represents the proba- limited follow-up. The entry Sx in column 8, row x, represents the bility of surviving more than t time units. Time t is measured from life table estimate of the probability of surviving more than x years diagnosis, start of treatment, or some other meaningful time point. from randomization. Computational shortcuts to observe are those For randomized studies, it is best to measure time from the date for column 8 (Sx = px = Sx–1) and for column 2 (lx+1 = lx−wx−dx). of randomization. There are basically two satisfactory methods for The product limit method of Kaplan and Meier112 is similar estimating S(t). The first is the life table or actuarial method110,111 in concept to the life table method. With the Kaplan-Meier ap- and is appropriate when the number of patients is large. The other proach, however, the intervals are defined by the actual survival method is the product limit method of Kaplan and Meier.112 This times of patients who have died. Suppose, for example, that the method is appropriate for any number of patients, but it involves survivals are 3, 3, 3+, 5, 6, 8+, 8+, 10, 10, and 12+ months,

TABLE 36.11 Life-Table Method for Estimating a Survival Distribution

No. Lost to No. Alive at Follow-Up or No. Died Proportion Proportion Cumulative Beginning Withdrawn Alive During At Risk During Dying Surviving Proportion Years After of Interval During Interval Interval Interval (Col 4/Col 5) (1 − Col 6) Surviving (Sx) Randomization lx wx dx (Col 2 − ½Col 3) qx px (p1 × p2 × . . . × px) 0–1 252 38 94 233 0.40 0.60 0.60 1–2 120 34 10 103 0.10 0.90 0.54 2–3 76 30 4 61 0.07 0.93 0.50 3–4 42 18 4 33 0.12 0.88 0.44 4–5 20 12 0 14 0.00 1.00 0.44 5–6 8 8 0 4 0.00 1.00 0.44

Col, column. 412 Practice of Oncology

TABLE 36.12 Kaplan-Meier Method for Estimating a Survival Distribution

Effective No. No. Lost to Exposed to Risk Cumulative No. Alive at Follow-Up or No. Died of Dying Just Proportion Proportion Proportion Beginning Withdrawn Alive During Before End of Dying Surviving Surviving Months After of Interval During Interval Interval Interval (Col 4/Col 5) (1 − Col 6) (p1 × p2 × . . . × px) Randomization lx wx dx (Col 2 − Col 3) qx px Sx 0–3 10 0 2 10 0.2 0.8 0.8 3–5 8 1 1 7 0.14 0.86 0.68 5–6 6 0 1 6 0.17 0.83 0.57 6–10 5 2 2 3 0.67 0.33 0.19

Col, column. where a plus sign follows survivals for patients still alive. Then the the standard error of the estimate can be conservatively estimated44 intervals are 0 to 3, 3 to 5, 5 to 6, and 6 to 10 months, as shown in as the Kaplan-Meier estimate of a survival distribution is based on Table 36.12. With the Kaplan-Meier method, deaths occur only at the assumption that censoring is noninformative, which means the ends of intervals. The entry in column 5 equals lx – wx rather that the censoring time is independent of the prognosis of the pa- than lx – 2wx for the life table method. This is because deaths occur tient. Most censoring in a randomized clinical trial results from the only at the ends of intervals here, and the number of patients at risk fact that some patients are alive and still being followed at the time of death just before the interval end is lx – wx. In the entry wx in of analysis. This is noninformative censoring. However, if patients column 3 for the Kaplan-Meier method, patients who are lost to are lost to follow-up—if they fail to return to clinic when they are follow- up or withdrawn alive at the end of an interval are consid- too sick to travel—then the censoring is informative and all the ered not lost or withdrawn until the following interval. These dif- usual methods of survival analysis are invalidated. Consequently, it ferences between the Kaplan-Meier and life table methods render is essential to obtain follow-up information actively on all patients the former more appropriate for studies with fewer patients. before analysis. If some patients have not been contacted for many Once the values Sx have been calculated for the Kaplan-Meier months and their status is unknown, that information should be method, they may be graphed with time on the horizontal axis. obtained before any analysis is performed. Examining the distribu- The graph is a step function that starts at time zero and ordinate tion of time since the last contact for patients not known to have 1.0. It drops to value Sx at time x, where x is the time at the right died is a good way to examine the adequacy of follow-up. end of an interval. The survival curve corresponding to Table 36.12 The issue of informative censoring also arises in considering end is shown in Fig. 36.1. The tic marks are placed on the curve at 3, 8, points other than death. For example, one may be attempting to es- and 12 months to represent the follow-up times of living patients. timate the distribution of time until tumor recurrence in the central The step function can be extended horizontally out to 12 months nervous system (CNS) in a pediatric leukemia trial. How should to represent follow-up of the last patient, but the right-hand end one handle patients whose disease recurs in the marrow without of the curve usually is very imprecisely estimated, and conclud- evidence of CNS recurrence? One may be tempted to censor the ing that a plateau exists at the level shown on the curve is often time to CNS recurrence of such patients at their time of marrow erroneous. recurrence, but that implicitly assumes that the censoring is nonin- For any time t, the Kaplan-Meier curve is an estimator of the true formative. Because CNS and marrow recurrence may be biologi- unknown value of S(t). The estimator is approximately normally cally linked, the assumption of noninformative censoring may not distributed in large samples. If m patients remain alive at time x, be valid. Other issues of informative censoring can be similarly problematic. Clearly, one should never censor patients because of lack of compliance with therapy, as this can severely bias results. More extensive discussions of statistical methods for the analysis of 1.00 clinical trial data are given by Marubini and Valsecchi.113

0.80 Multiple Comparisons Table 36.13 shows the probability of obtaining one statistically sig- nificant p( <0.05) difference by chance alone as a function of the 0.60 number of independent comparisons of two equivalent treatments. With only five comparisons, the chance of at least one false- positive conclusion is 22.6%. When the number of end points, 0.40 interim analyses, and patient subsets are considered in the analysis of clinical trials, these results are disturbing.114 The comparisons

Probability of survival performed in clinical trials are not entirely independent, but this 0.20 does not have much effect on ameliorating the problem. Fleming and Watelet115 performed a computer simulation to determine the chance of obtaining a statistically significant treatment difference when two equivalent treatments in six subsets determined by three 24681012dichotomous variables are compared. The chance of a statistically significant difference between treatments in at least one subset Month was 20% at the final analysis and 39% in the final or one of the Figure 36.1 Example of estimated survival distribution. three interim analyses. Subset analysis, comparison of treatments

tahir99 - UnitedVRG Chapter 36 Design and Analysis of Clinical Trials 413

TABLE 36.13 TABLE 36.14 Probability of Obtaining at Least One Summary of Guidelines for Reporting Clinical Statistically Significant (P <0.05) Difference Trials116 by Chance Along in Multiple Comparisons of Two Equivalent Treatments ■ Quality control of data and response evaluations should be discussed. Probability of at Least One ■ All patients registered on study should be accounted for. Comparisons “Significant” Difference (%) ■ Inevaluability rate for major end points should not exceed 15%. ■ 15 No exclusions of eligible patients in comparing outcomes by treatment group. 2 9.7 ■ The sample size should be large enough to establish or 3 14.3 conclusively rule out effects of clinically important magnitude. Confidence limits for size of treatment versus control 4 18.5 effectiveness should be given. 5 22.6 ■ Publication should provide protocol-specified sample size and 10 40 interim analysis plan as well as actual timing of analyses. ■ Claims of therapeutic effectiveness should not be based on 20 64.1 phase 2 trials. ■ Generalizability of conclusions should be carefully discussed. Subset-specific claims should be justified based on prospective with regard to multiple end points, and multiple interim analy- planning and statistical control of study-wise type I error. ses are common sources of erroneous conclusions. The primary end point should be defined in the protocol. Subset analyses and analyses with regard to secondary end points should be specified conducted with 90% power and a 5% statistical significance level, in advance, and statistical significance should be declared only for the false positive discovery rate is about 9%. significance levels much defined in advance to limit the study-wise An additional factor to consider is that of publication bias,125 type 1 error to 5%. which denotes the preference of journals to publish positive rather Generally, it is not valid to adjust the analysis by characteristics than negative results. A negative result may not be published at all, measured after the start of treatment (e.g., compliance, dose de- particularly from a small trial. If it is published, it is likely to appear livered, toxicity). New approaches to subset analysis and multiple in a less widely read journal than it would if the result were positive. end point analysis using Bayesian methods have been described by These observations emphasize that results in the medical lit- PRACTICE OF ONCOLOGY 93,116 Dixon and Simon. Qualitative interaction tests are described erature often cannot be accepted at face value. It is important to 117 by Gail and Simon. recognize that “positive” results need confirmation, particularly positive results of small studies, before they can be believed and applied to the general population. REPORTING RESULTS OF CLINICAL TRIALS

Effective reporting of results is an integral part of good research. META-ANALYSIS Unfortunately, numerous reviews have indicated that the quality 98,118,119 120 of reporting of clinical trial results is poor. Pocock et al. A meta-analysis is a quantitative summary of research on a topic. concluded that “overall, the reporting of clinical trials appears to It is distinguished from the traditional literature review by its em- be biased toward an exaggeration of treatment differences.” Tan- phasis on quantifying results of individual studies and combining 98,114 nock and Murphy have given clear illustration of how this results across studies. Key components of this approach for thera- is easily done. The guidelines summarized in Table 36.14 are peutics are to include only randomized clinical trials, to include 121 adapted from those proposed by Simon and Wittes. all relevant randomized clinical trials that have been initiated (regardless of whether they have been published), to exclude no randomized patients from the analysis, and to assess therapeutic FALSE POSITIVE REPORTS IN THE effectiveness based on the average results pooled across trials.126 LITERATURE Attention is restricted to randomized trials, because the bias from nonrandomized comparisons may be larger than the small Many of the positive results reported in the literature for small to moderate therapeutic effects likely to be present. Including all clinical trials are probably false positive results.122–124 In 100 trials, relevant randomized trials that have been initiated in a geographic suppose that there are 10 in which the experimental treatment is area (e.g., the world, or the Americas and Europe) represents an sufficiently better than the control such that there is a 90% chance attempt to avoid publication bias. Avoiding exclusion of any ran- of the difference being detected in a small or moderate-sized domized patients also functions to avoid bias. Assessing therapeu- clinical trial. Of these 10 trials, obtaining a statistically significant tic effectiveness based on average pooled results is an attempt to difference is expected in 9. Of the remaining 90 trials, we assume make the evaluation on the totality of evidence rather than on that the treatments are approximately equivalent to the control. A extreme isolated reports. In calculating average treatment effects, statistically significant difference could be expected in 5% (4.5) a measure of difference in outcome between treatments is cal- of these. Hence, of the 13.5 (9 + 4.5) trials that yield statistically culated separately for each trial. For example, an estimate of the significant results, the finding is false positive in 4.5 or 33% of the logarithm of the hazard ratio can be computed for each trial. A cases. The 33% false discovery rate is striking but it depends on the weighted average of these study-specific differences is then com- assumption that only 10% of the trials study new treatments with puted, and the statistical significance of this average is evaluated. large treatment effects. For large clinical trials, the size of treat- This approach to meta-analysis requires access to individual pa- ment effect that can be detected with high statistical power is likely tient data for all randomized patients in each trial. It also requires to be larger. The Eastern Cooperative Oncology Group reported collaboration of the leaders of all the relevant trials and is very that about one-third of their phase 3 clinical trials resulted in statis- labor-intensive. Nevertheless, it represents the gold standard for tically significant results.123 Assuming that most of these trials are meta-analysis methodology. 414 Practice of Oncology

A major issue of concern in meta-analyses is whether the indi- evaluate survival. Similarly, subset analysis can usually be mean- vidual trials are sufficiently similar to make calculation of average ingfully evaluated only in the context of a meta-analysis, because effects medically meaningful. If the therapeutic interventions or individual trials are not sized for this objective. control treatments differ too greatly or if the patient populations Meta-analysis is not an alternative to properly designed and are too different, the results may not be medically meaningful as a sized randomized clinical trials. Some have suggested that one need basis for making treatment decisions for individual patients. Often not be concerned about computing sample size in the traditional in cancer therapeutics, the studies will not be identical in their ways, as small, randomized trials can be pooled for meta-analysis. treatment regimens or their patient populations, but they will not Because most investigators would prefer to “do their own thing,” be so different as to make the results meaningless. In this case, this would lead to a proliferation of diverse trials of inconsequential the meta-analysis may be useful for answering important ques- individual size that may be too heterogeneous to permit a mean- tions about a class of treatments that the individual trials cannot ingful meta-analysis. Given that sufficient large, randomized clini- address reliably. For example, trials evaluating adjuvant treatment cal trials of very similar treatment regimens have been conducted, of primary breast cancer often are designed to detect differences meta-analysis can provide supplemental information about a given in disease-free survival, and a meta-analysis is often required to class of treatments that is not available from the individual trials.

SELECTED REFERENCES

The full reference list can be accessed at lwwhealthlibrary.com/oncology. 46. Freidlin B, Simon R. An evaluation of the randomized discontinuation design. J Clin Oncol 2005;23:1–5. 1. Subramanian J, Simon R. Gene expression-based prognostic signatures in 48. Goldman B, LeBlanc M, Crowley J. Interim futility analysis with intermediate lung cancer: ready for clinical use? J Natl Cancer Inst 2010;102:464–474. endpoints. Clin Trials 2008;5:14–22. 2. Green S, Benedetti J, Crowley J. Clinical Trials in Oncology. 2nd ed. 51. Schaid DJ, Wieand S, Therneau TM. Optimal two-stage screening designs London: Chapman & Hall/CRC; 2003. for survival comparisons. Biometrics 1990;77:507–513. 3. Leventhal BG, Wittes RE. Research Methods in Clinical Oncology. New 54. Freidlin B, Korn EL, Gray R, et al. Multi-arm clinical trials of new agents: York: Raven Press; 1988. some design considerations. Clin Cancer Res 2008;14:4368–4371. 4. Eisenhauer EA, O’Dwyer PJ, Christian M, et al. Phase I clinical trial design 55. Freidlin B, McShane LM, Polley MYC. Randomized phase II trial designs in cancer drug development. J Clin Oncol 2000;18:684–692. with biomarkers. J Clin Oncol 2012;30:3304–3309. 5. Simon R, Freidlin B, Rubinstein L. Accelerated titration designs for phase I 56. Simon R. Randomized clinical trials. Principles and obstacles. Cancer clinical trials in oncology. J Natl Cancer Inst 1997;89:1138–1147. 1994;74:2614–2619. 9. Babb J, Rogatko A, Zacks S. Cancer phase I clinical trials: efficient dose esca- 57. Torri V, Simon R, Russek-Cohen E, et al. Relationship of response and lation with overdose control. Stat Med 1998;17:1103–1120. survival in advanced ovarian cancer patients treated with chemotherapy. 10. Goodman SN, Zahurak ML, Piantadosi S. Some practical improvements in J Natl Cancer Inst 1992;84:407–414. the continual reassessment method for phase I studies. Stat Med 1995;14: 58. Buyse M, Molensberghs G, Burzykowski T, et al. The validation of surrogate 1149–1161. endpoints in meta-analyses of randomized experiments. Biostatistics 2000; 12. Simon RM, Steinberg SM, Hamilton M, et al. Clinical trial designs for the 1:49–67. early clinical development of therapeutic cancer vaccines. J Clin Oncol 2001; 60. Korn EL, Albert PS, McShane LM. Assessing surrogates as trial endpoints 19:1848–1854. using mixed models. Stat Med 2004;24:163–182. 15. Kummar S, Kinders R, Rubinstein L, et al. Compressing drug development 61. Fleming TR, Rothmann MD, Lu HL. Issues in using progression-free timelines in oncology using phase ‘0’ trials. Nat Rev Cancer 2007;7:131–139. survival when evaluating oncology products. J Clin Oncol 2009;27: 17. Karapetis CS, Khambata-Ford S, Jonker DJ, et al. K-ras mutations and 2874–2880. benefit from cetuximab in advanced colorectal cancer. N Engl J Med 2008; 62. Freidlin B, Korn EL, Hunsberger S, et al. Proposal for the use of progres- 359:1757–1765. sion-free survival in unblinded randomized trials. J Clin Oncol 2007;25: 20. Pusztai L, Anderson K, Hess KR. Pharmacogenomic predictor discovery in 2122–2126. phase II clinical trials for breast cancer. Clin Cancer Res 2007;13:6080–6086. 63. Dodd LE, Korn EL, Freidlin B, et al. Blinded independent central review of 21. LeBlanc M, Rankin C, Crowley J. Multiple Histology Phase II Trials. Clin progression-free survival in phase III clinical trials: important design element Cancer Res 2009;15:4256–4262. or unnecessary expense? J Clin Oncol 2008;26:3791–3796. 22. Dobbin KK, Zhao Y, Simon RM. How large a training set is needed to 64. Simon R. An agenda for clinical trials: clinical trials in the genomic era. Clin develop a classifier for microarray data? Clin Cancer Res 2008;14:108–114. Trials 2004;1:468–470. 23. Eisenhauer EA, Therasse P, Bogaerts J, et al. New response evaluation cri- 65. Simon R. A roadmap for developing and validating therapeutically relevant teria in solid tumors: revised RECIST guideline (version 1.1). Eur J Cancer genomic classifiers. J Clin Oncol 2005;23:7332–7341. 2009;45:228–247. 66. Simon R. New challenges for 21st century clinical trials. Clin Trials 2007; 24. Simon R. Optimal two-stage designs for phase II clinical trials. Control Clin 4:167–169. Trials 1989;10:1–10. 67. Simon R, Maitournam A. Evaluating the efficiency of targeted designs for 29. Seymour L, Ivy SP, Sargent D, et al. The design of phase II clinical trials randomized clinical trials. Clin Cancer Res 2005;10:6759–6763. testing cancer therapeutics: consensus recommendations from the clinical 68. Simon R, Maitournam A. Evaluating the efficiency of targeted designs for trial design task force of the National Cancer Institute investigational drug randomized clinical trials. Erratum. Clin Cancer Res 2006;12:3229. steering committee. Clin Cancer Res 2010;16:1764–1769. 69. Hoering A, LeBlanc M, Crowley J. Randomized phase III clinical trial 30. Vidauurre T, Wilkerson J, Simon R, et al. Stable disease is not preferentially designs for targeted agents. Clin Cancer Res 2008;14:4358–4367. observed with targeted therapies and as currently defined has limited value 70. Mandrekar SJ, Sargent DJ. Clinical trial designs for predictive biomarker in drug development. Cancer J 2009;15:366–373. validation: theoretical considerations and practical challenges. J Clin Oncol 31. El-Maraghi RH, Eisenhauer EA. Review of phase II trial designs used in stud- 2009;27:4027–4034. ies of molecular targeted agents: outcomes and predictors of success in phase 71. Simon R. Using genomics in clinical trial design. Clin Cancer Res III. J Clin Oncol 2008;26:1346–1354. 2008;14:5984–5993. 34. Estey EH, Thall PF. New designs for phase 2 clinical trials. Blood 2003; 72. Simon RM. Genomic Clinical Trials and Predictive Medicine. Cambridge, 102:442–448. UK: Cambridge University Press; 2013. 35. Korn EL, Liu PY, Lee SJ, et al. Meta-analysis of phase II cooperative 73. Karuri SW, Simon R. A two-stage Bayesian design for co-development of new group trials in metastatic stage IV melanoma to determine progression- drugs and companion diagnostics. Stat Med 2012;31:901–914. free and overall survival benchmarks for future phase II trials. J Clin Oncol 74. Freidlin B, Sun Z, Gray R, et al. Phase III clinical trials that integrate treat- 2008;26:527–534. ment and biomarker evaluation. J Clin Oncol 2013;31:3158–3161. 38. Korn EL, Arbuck SG, Pluda JM, et al. Clinical trial designs for cytostatic 75. Freidlin B, McShane LM, Korn EL. Randomized clinical trials with agents: are new approaches needed? J Clin Oncol 2001;19:265–272. biomarkers: design issues. J Natl Cancer Inst 2010;102:152–160. 39. Rubinstein LV, Korn EL, Freidlin B, et al. Design issues of randomized 80. Simon RM, Paik S, Hayes DF. Use of archived specimens in evaluation of phase 2 trials and a proposal for phase 2 screening trials. J Clin Oncol 2005; prognostic and predictive biomarkers. J Natl Cancer Inst 2009;101:1–7. 23:7199–7206. 81. Pocock S, Simon R. Sequential treatment assignment with balancing for 40. Simon R, Wittes RE, Ellenberg SS. Randomized phase II clinical trials. prognostic factors in the controlled clinical trial. Biometrics 1975;31:103–115. Cancer Treat Rep 1985;69:1375–1381. 83. Simon R, Simon N. Using randomization tests to preserve type I error with 45. Rosner G, Stadler W, Ratain M. Randomized discontinuation design: Appli- response-adaptive and covariate-adaptive randomization. Stat Probab Lett cation to cytostatic antineoplastic agents. J Clin Oncol 2002;20:4478–4484. 2011;81:767–772.

tahir99 - UnitedVRG Chapter 36 Design and Analysis of Clinical Trials 415

84. Rubinstein L, Gail M, Santner T. Planning the duration of a comparative 103. Haybittle JL. Repeated assessment of results in clinical trials of cancer clinical trial with loss to follow-up and a period of continued observation. treatment. J Radiol 1971;44:793–797. J Chronic Dis 1981;34:469–479. 104. O’Brien PC, Fleming TR. A multiple testing procedure for clinical trials. 88. Simon R. Confidence intervals for reporting results from clinical trials. Ann Biometrics 1979;35:549–556. Intern Med 1986;105:429–435. 107. Freidlin B, Korn EL. Monitoring for lack of benefit: a critical component of 91. Simon R. Bayesian design and analysis of active control clinical trials. a randomized clinical trial. J Clin Oncol 2009;27:629–633. Biometrics 1999;55:484–487. 109. Lan KKG, Simon R, Halperin M. Stochastically curtailed test in long-term 92. Spiegelhalter DJ, Freedman LS, Parmar MK. Bayesian approaches to clinical trials. Commun Stat Seqen Anal 1982;1:207–219. randomized trials. J R Stat Soc Series A General 1994;157:357–387. 112. Kaplan EI, Meier P. Nonparametric estimation from incomplete observa- 93. Simon R. Bayesian subset analysis: application to studying treatment-by- tions. J Am Stat Assoc 1958;53:457–481. gender interactions. Stat Med 2002;21:2909–2916. 114. Tannock IF. False-positive results in clinical trials: multiple significance 95. Simon R. Discovering the truth about tamoxifen: problems of multiplicity in tests and the problem of unreported comparisons. J Natl Cancer Inst 1996; the evaluation of biomedical data. J Natl Cancer Inst 1995;87:627–629. 88:206–207. 96. Peto R, Pike MC, Armitage P. Design and analysis of randomized clini- 115. Fleming TR, Watelet L. Approaches to monitoring clinical trials. J Natl cal trials requiring prolonged observation of each patient. II. Analysis and Cancer Inst 1989;81:188–193. examples. Br J Cancer 1977;35:1–39. 116. Dixon DO, Simon R. Bayesian subset analysis. Biometrics 1991;47:871–881. 97. Barr J, Tannock I. Analyzing the same data two ways: a demonstration model illus- 117. Gail M, Simon R. Testing for qualitative interactions between treatment trate the reporting and misreporting of clinical trials. J Clin Oncol 1989;7:969–978. effects and patient subsets. Biometrics 1985;41:361–372. 98. Tannock I, Murphy K. Reflections on medical oncology: an appeal for better 123. Simon R. Commentary on “Clinical trials and sample size considerations: clinical trials and improved reporting of their results. J Clin Oncol 1983;1:66–70. Another perspective.” Stat Sci 2000;15:95–110. 101. Ellenberg S, Fleming TR, DeMets D. Data Monitoring Committees in 125. Begg CB, Berlin JA. Publication bias and dissemination of clinical research. Clinical Trials: A Practical Perspective. Hoboken, NJ: Wiley; 2002. J Natl Cancer Inst 1989;81:107–115. 102. Smith M, Ungerleider R, Korn E, et al. The role of independent data 126. Collins R, Gray R, Godwin J, et al. Avoidance of large biases and large monitoring committees in randomized clinical trials sponsored by the random errors in the assessment of moderate treatment effects: the need for National Cancer Institute. J Clin Oncol 1997;15:2736–2743. systematic overviews. Stat Med 1987;6:245–254. PRACTICE OF ONCOLOGY PRACTICE Section 1 Cancer of the Head and Neck

Molecular Biology of Head 37 and Neck Cancers

Thomas E. Carey and Mark E. P. Prince

INTRODUCTION those of the oral cavity and larynx, while at the same time there has been a steady increase in tonsillar cancers.12 Trends for head and neck cancer incidence by site in the United States and Can- Incidence, Risk Factors, and Etiology ada indicate that oropharynx cancers have increased continuously over the past 25 years.9,13 The oropharyngeal cancer increase Head and neck squamous cell carcinoma (HNSCC) accounts began in the late 1970s and is attributed to increased rates of in- for 90% of all malignant disease in the head and neck region of fection with high-risk human papillomaviruses (hrHPV) second- the body. With 550,000 new cases diagnosed worldwide each ary to changing sexual mores that are traceable to widespread use year, HNSCC is a major public health problem. In the United of oral contraceptives, reduced use of condoms, and freedom to States, there are 40,000 new cases and roughly 14,000 deaths from have more sexual partners, without the fear of an unwanted preg- HNSCC each year. Historically, HNSCC has been a disease of nancy. Studies in the United States, Canada, and Western Europe older males with heavy lifelong tobacco use, high alcohol con- clearly show a trend for increasing incidence of hrHPV, notably, sumption, poor diet, and bad dentition. As smoking increased HPV16 in oropharyngeal cancers by year of diagnosis.14–22 At our among women, the male to female ratio of 5:1 observed in the institution, there was a steady increase in the proportion of HPV- 1960s declined to 3:1 in the 1990s. Most mucosal squamous can- positive oropharynx cancer in three clinical trials done over the cers of the head and neck, particularly those of the oral cavity, past decade (Table 37.1). Surveillance, Epidemiology, and End larynx, and hypopharynx are still associated with these etiologic Results (SEER) data illustrate the decline in smoking-related can- factors as well as other cultural habits, such as oral tobacco use, 1,2 cers of the larynx, the increase in oropharyngeal cancer, and the and in other countries, betel and areca nut chewing. A small but remarkable decline in cervical cancer that is attributable to an mysterious segment of head and neck tumors arise in individuals early diagnosis with Pap smears, hrHPV testing, and colposcopy under 40 years of age who have no known etiologic factors. In the of cervical lesions (Fig. 37.1). In 2013, the incidence of new oro- past 10 years, we noted that at our institution, HNSCC patients pharyngeal cancer cases was expected to surpass that of cervical under age 35 included more women than men, and during this cancer. Unfortunately, no early detection and treatment for HPV- period, four cases of tongue cancer in pregnant women were also 3 induced oropharynx cancer is available, but this is a rich area observed. The etiology of the cancers in the men and women in for investigation. the young age group is unknown. A small percentage of HNSCC HPV is the most common sexually transmitted disease with have a familial origin; however, when corrected for tobacco and 4 an overall genital HPV prevalence of 42.5% in females 14 to 59 alcohol use, most associations lose statistical significance. Nev- years of age as determined in the National Health and Nutrition ertheless, familial head and neck cancers associated with inher- Examination Survey (NHANES) (2003–2006).23 In the oral cavity ited defects of the CDKN2A locus causing loss of function of the 5 and oropharynx, the lymphoid-rich tissue of the tonsils is the likely cyclin-dependent kinase inhibitor p16INK4A and the hMDM2 reservoir of HPV. To replicate, HPV must infect the basal epithe- regulator p14ARF occur. Two members of a family were treated lial cell. The thin epithelium of the tonsillar crypts serves as an recently for HNSCC in our department. An analysis revealed that ideal location for the virus to infect (Fig. 37.2). HPV infection may each had the same somatic CDKN2A mutation. Families with in- be facilitated by concurrent inflammation within the tonsil due herited mutations of cancer-related genes tend to have multiple to other microorganisms. D’Souza et al.24 conducted a hospital- types of malignant tumors, and second primary tumors of different based case control study of 100 patients with newly diagnosed oro- histologic types.6 In fact, CDKN2A was originally called multiple 7,8 pharynx cancer to 200 noncancer patients to evaluate associations tumor suppressor gene 1 (MTS-1) because it was so frequently between HPV infection and oropharyngeal cancer. A high lifetime mutated in multiple tumor types, including melanoma, pancreatic number of vaginal sex partners (≥26) and oral sex partners (≥6) cancers, breast cancers, and head and neck cancers. The genes as- was significantly associated with oropharyngeal cancer as was se- sociated with an inherited predisposition to head and neck cancer 9 ropositivity for antibodies to the viral L1 protein or oral infection include the master regulator, TP53 (Li-Fraumeni syndrome), and with any hrHPV type, most frequently HPV16. However, other Fanconi anemia genes, the FANC family of DNA repair genes that 10,11 high-risk HPV types are also found in roughly 10% of oropharynx are also associated with the development of HNSCC. Head tumors and in other head and neck cancer locations, notably the and neck cancers arising in young patients and others that are in- nasopharynx and the oral cavity; the proportion of other high-risk dependent of tobacco exposure provide a rich area for an in-depth HPV types is much higher.25,26 Oral and vaginal HPV infections molecular assessment to determine the underlying genetic factors. in a series of HIV-positive and HIV-negative women revealed a significant but lower incidence of oral HPV (15%) compared to High-Risk Human Papillomaviruses vaginal HPV (51%).27 Oral HPV infection incidence in the gen- eral U.S. population was evaluated in the NHANES study.28 The As smoking has declined in the United States, so has the inci- prevalence of oral HPV infection shows a biphasic distribution dence of smoking-related head and neck cancer, particularly with one peak in the 30- to 34-year-old age group and a second 416

tahir99 - UnitedVRG Chapter 37 Molecular Biology of Head and Neck Cancers 417

TABLE 37.1 the excellent prognosis of HPV-positive oropharynx cancers, in HPV-positive oral cancers,37 and HPV-positive nasopharyngeal Changing HPV Involvement in Oropharynx Cancers: cancers,38 the outcome may be worse than HPV-negative cancers Over 10 Years of Study Accruals: University of arising in these sites. Michigan Head and Neck Oncology Program At the University of Michigan, 80% to 90% of oropharyngeal tu- mors (see Table 37.1) that present are HPV positive,21,25 and >80% HPV Positivity by Clinical Trial and Time Period of patients with HPV-positive oropharynx cancer are responders to UMCC 9921 UMCC 0221 UMCC 0221–like concurrent carboplatin and taxol with intensity-modulated radia- HPV Status 1999–2001 2001–2005 2005–2011 tion treatment (chemoRT).39 Critical questions for the future are HPV negative 14 (34%) 8 (10%) 6 (8%) to determine which patients can benefit from less intensive treat- ment, which patients need the intensive concurrent chemoRT HPV positive 27 (66%) 75 (90%) 71 (92%) treatment, and which patients need a different treatment because 40,41 UMCC, University of Michigan Cancer Center. they fail to be cured by chemoRT. Spector et al. showed that nodal status and, in particular, matted nodes (defined as three nodes abutting one another with loss of an intervening fat plane replaced with evidence of extracapsular spread) defines a group of peak in the 60- to 64-year-old age group. Overall, oral HPV infec- patients with a greatly increased risk of distant metastasis and poor tion was (6.9%), with high-risk HPV types in 3.7%. Men had a overall and disease-free survival. The molecular correlates of these higher prevalence than women (10.1% versus 3.6%), and blacks highly malignant HPV-positive tumors are yet to be defined fully, had a higher prevalence of oral HPV (10.5%) than whites (6.5%), but in work submitted for publication, Walline et al. have shown although this difference did not reach statistical significance. It is that integration of hrHPV into cancer-related genes is common not yet certain how oral HPV infection relates to the risk of devel- in tumors of patients who suffer from recurrent disease and poor oping HPV-induced head and neck cancer. survival, whereas integration of the viral genome into intragenic Fortunately, HPV-positive oropharyngeal cancers show im- sites was more commonly associated with good response to therapy proved response to therapy when compared to HPV-negative can- and prolonged progression-free survival. Immune responses to cers of the same site.29–32 Patients presenting with HPV-positive HPV antigens have also been suggested to be a cause of the good oropharyngeal cancers also tend to be younger, healthier, and response to therapy among HPV-positive oropharyngeal cancers. have a much lower frequency of smoking and alcohol abuse.33 However, failure of a potent immune response in some cases may However, some studies have shown that a history of smoking in occur because of frequent alterations affecting the tumor necrosis patients with HPV-positive oropharynx cancer may increase the factor (TNF) receptor TRAF3, which participates in the activa- likelihood of recurrence or metastasis after treatment.21,34 This tion of the immune response and nuclear factor kappa B (NF-κB) 42 has prompted many to consider reducing the intensity of treat- signaling (TCGA Head and Neck Cancer Consortium). These OF ONCOLOGY PRACTICE ment to decrease treatment-related morbidity.35,36 In contrast to changes may provide a mechanism for HPV-positive tumors to

Age-Adjusted SEER Incidence Rates 16

Cervix uteri–females

14 Oropharynx and tonsil–males Larynx–both sexes

12

10

8

Rate per 100,000 6

4

2

0 197519761977197819791980198119821983198419851986198719881989199019911992199319941995199619971998199920002001200220032004200520062007200820092010

Year of Diagnosis Figure 37.1 Changing incidence rates of larynx, cervix, and oropharynx cancers. Adapted by R. Meza, from Surveillance, Epidemiology, and End Results (SEER) Program Populations (1969–2011) (www.seer.cancer.gov/popdata), National Cancer Institute, Division of Cancer Control and Population Sciences (DCCPS), Surveillance Research Program, Surveillance Systems Branch, released January 2013. 418 Practice of Oncology / Cancer of the Head and Neck

and neck cancers and are associated with poor prognosis.46,47 It is risk Tonsil puzzling that only about 10% of HNSCC cases respond to treat-   ment that targets EGFR.48 Alternate driver pathways are suspected as an explanation for this observation. Early molecular studies of HNSCC also revealed that mutations, loss of heterozygosity, and methylation of the CDNK2A locus is very common and occurs Oral rinse Tonsil 49–51 &     early in HNSCC development. Mutation and overexpression yp 52 &  xpr  of p53 was also identified in early studies of HNSCC. CyclinD1 amplification and overexpression has also been identified as an im-    r portant biomarker that is associated with outcomes and responses   #  to treatment.53 In the landmark VA larynx trial, a biomarker analysis Normal Tonsil     revealed that overexpressed p53 predicted the response to induction &     chemotherapy and radiation (RT).54 Because p53 overexpression is &xpr  nearly always associated with mutant p53, this suggested that tu- xpr &     mors with wild-type p53 were less responsive to chemotherapy and  "    RT, even though normal cells containing wild-type p53 are very  s  ! "! sensitive to chemotherapy or RT. Thus, it was unclear why tumors with wild-type p53 were more resistant to chemotherapy and RT. Subsequent studies revealed that the combination of wild-type p53 and overexpression of Bcl-xL resulted in cisplatin resistance.55 This suggested that Bcl-xL blocks p53-induced apoptosis, and that wild- type p53 mediates cell cycle arrest and repair mechanisms, allow-      "  ing tumor cells to escape from treatment-induced damage. In spite of our awareness of EGFR overexpression and its po- Malignant transformation tential value as a target for therapy, it is still impossible to predict which head and neck cancers will respond and which will not. The same is true for many other molecular markers when only one molecular target is addressed. It is now becoming clear that it will be necessary to identify the driver pathways and target more than one cancer driver. It was postulated many years ago that a certain number of “hits” is necessary to drive a tumor. Knudsen correctly explained the tumor suppressor nature of the retinoblas- toma gene by comparing the kinetics of tumor development in children with an inherited predisposition and in children who developed sporadic retinoblastoma. When time to tumor develop- In"sion  ment in affected children was plotted against months from birth,     there was a straight line for children who inherited a mutant al- lele from an affected parent. However for children with sporadic Figure 37.2 Human papillomavirus (HPV) infection and replication in the retinoblastoma the line had a plateau for several months and then oropharynx, with the tonsil as the primary target for HPV infection and fell linearly. From this he reasoned that two hits were necessary for carcinogenesis. HPV detection can be carried out in oral rinses, normal tumor development. For children who inherited one normal and tonsils, and tumors and lymph nodes. one mutant gene he observed first-order kinetics, and for children who did not inherit a mutant gene he observed second-order kinet- ics. Thus only one additional event was necessary for tumor devel- opment in the inherited disease, but two events were necessary in 42 evade antiviral immune responses. As a group, HPV-positive head the sporadic cases, i.e. acquisition of a mutation in a cell and then and neck cancers have fewer gross genetic changes than HPV- a second event in the same cell, resulting in loss of the healthy negative tumors and share specific similarities, such as wild-type gene. For tumors that arise in adult epithelial tissues, the situation TP53 and intact p16ink4a (CDKN2A), both of which are rare in is more complex. It was estimated many years ago using a similar HPV-negative HNSCC. Nearly all HPV-positive HNSCC express mathematical approach in cancers that develop with aging that it is the HPV E6 and E7 viral oncogenes and exhibit low or absent likely that four to five events are necessary for a tumor to develop. expression of the HPV E2 gene that serves as a negative regulator To attempt to identify cancer drivers, we characterized chromo- of E6 and E7 expression (Walline et al. submitted). Viral integra- somal abnormalities in two cell lines derived from a patient with 43 tion appears to be a requirement for carcinogenesis. Integration recurrent laryngeal cancer. One portion of the tumor was within typically disrupts E2, which results in the unregulated expression the endolarynx. From this, the UM-SCC-17A cell line was estab- of the E6 and E7 oncoproteins and includes the expression of the lished, and from a second portion of the tumor that had invaded 44 alternate HPV E6 transcripts E6*I and E6*II. It appears that viral through the thyroid cartilage into the soft tissues of the neck, UM- integration also causes a significant disruption of the host genome SCC-17B was established. There were a total of 22 chromosome 45 at sites of viral integration. When significant disruption occurs rearrangements in the two cell lines, of which 8 were unique to the leading to other genomic damage, this could also be a factor in the UM-SCC-17A cell line from the endolarynx and 9 were unique to subset of nonresponsive HPV-positive tumors. the UM-SCC-17B cell line from the invasive lobe of the tumor. Of interest, however, were the five rearrangements that were common to both cell lines, which could represent the loci of genetic events MOLECULAR MECHANISMS IN HEAD AND associated with tumor initiation, whereas the others represent ei- NECK SQUAMOUS CELL CARCINOMA ther random events or events associated with progression.56 It will be of interest to use modern sequencing methods to identify the ge- Early molecular studies of head and neck cancers revealed that am- netic aberrations linked to these five early chromosome rearrange- plification of the epidermal growth factor receptor (EGFR) gene ments to test this hypothesis. Hahn and Weinberg tested what steps with associated EGFR overexpression are common among head were necessary to transform normal human bronchial epithelial

tahir99 - UnitedVRG Chapter 37 Molecular Biology of Head and Neck Cancers 419 cells into tumor cells. They targeted a small but consistent set of tumors lacked EGFR amplification but contained amplification specific pathways based on the observed abnormalities of many ep- of fibroblast growth factor receptor (FGFR)1, 2, or 3. A few tumors ithelial cancers. The necessary pathways include: ectopic expres- had more than one RTK amplification, but mostly subsets could sion of the catalytic unit of human telomerase, human telomerase be categorized by amplification of one RTK. CyclinD1 (CCND1) reverse transcriptase (hTERT), to stabilize telomeres over many amplification was very common, occurring in 30% of all tumors. cell divisions; abrogation of the Rb pathway with SV40 large T on- CCND1 amplification was observed in some tumors with RTK coprotein to induce continuous cell cycle entry; inhibition of TP53 amplification and in a subset of tumors withCMYC amplification, with large T oncoprotein to facilitate cell cycle progression and to although there was also a subset that had either CCDN1 or CMYC inhibit TP53-mediated induction of apoptosis; and introduction of amplification, but not both.HRAS mutations were found in 5%, the activated (oncogenic) allele of HRAS, to provide continuous PIK3CA mutations were found in 18%, PTEN mutations in 12%, signaling through the RAS, RAF, mitogen-activated protein kinase whereas TP53 mutations were present in 83%. (MEK), extracellular signal-related kinase (ERK) kinase cascade to The tumor subsets defined by each of these distinct categories trigger the expression of transcription factors associated with pro- offers promise for better response when the correct amplified or liferation.57,58 This demonstration is critical to our thinking of how mutated gene is targeted in those categories. Inhibitors for EGFR, we will develop precision medicine for cancer because, similar to FGFR, and PIK3CA are all now available and suggest that the ap- the chromosome rearrangements observed in the UM-SCC-17A propriate matching of the tumor characteristics with the correct and UM-SCC-17B cell lines mentioned previously, each epithe- inhibitor is a fruitful area for study. Although CCND1 inhibitors lial cancer is likely to have multiple abnormalities, of which some are not yet available, several companies have cyclin-dependent ki- are primary driver events and others are passenger events that nase inhibitors (CDK4/6), and use of these agents is being tested have developed incidentally due to genomic destabilization. The in tumors with a loss of CDKN2A and might also be applicable in knowledge that a limited number of events are sufficient for the tumors with CCND1 amplification.NOTCH1 mutations found conversion of a normal human epithelial cell to an immortalized, in 18% of HNSCC are a recently appreciated feature of a subset of invasive tumor cell population suggests that a series of principles HNSCC and are typically loss-of-function mutations, suggesting can be developed that control the conversion of normal to malig- that NOTCH1 is a tumor suppressor in this cancer type. In normal nant behavior and that we should be able to develop strategies that squamous epithelium, NOTCH1 acts to drive squamous differen- will target the two or three critical pathways of an individual tumor tiation by feeding back on the Wnt/β-catenin signaling pathway that serve as the primary drivers of neoplastic behavior. active in the proliferating basal cell. Loss of the NOTCH1 tumor suppressor function allows unopposed activity of the Wnt pathway, allowing this to become a target for therapy. Liu et al.60 recently THE CANCER GENOME ATLAS PROJECT reported the development of a novel inhibitor of the Wnt pathway that targets Wnt secretion by inhibiting porcupine, the enzyme re-

Most studies have looked at individual target genes in isolation or in quired for Wnt secretion. This inhibitor, LGK974, was more effec- OF ONCOLOGY PRACTICE small combinations of genes to identify tumor driver events; how- tive in tumor cell lines shown to have NOTCH1 loss of function ever, these have largely been thwarted by the range of mechanisms mutations, suggesting that this may be a useful agent in HNSCC that a tumor cell might acquire stochastically to become malignant. tumors with NOTCH pathway abnormalities. This limited our ability to assess the drivers of individual cancers and The next horizon for targeted agents is to find the correct com- has complicated the assessment of the efficacy of pathway-targeted binations of targeted compounds for individual tumors. For this agents. The development of modern molecular analysis techniques effort, it will be necessary to identify targets in different pathways has made it possible to analyze nearly all of the abnormalities in an for which effective agents exist. Such a combination might include individual tumor. Furthermore, The Cancer Genome Atlas (TCGA) inhibitors of RTKs, together with a CDK4/6 inhibitor or a Wnt project has provided us with an impressive collection of data from pathway inhibitor and PIK3CA inhibitor. Many combinations multiple tumors of the same types that can be used to bring us closer are possible, and the literature will be replete with combinations to developing knowledge of the types of oncogenic drivers and tumor of multiple targeted agents within the next few years as precision suppressors that characterize each tumor type. By examining indi- medicine approaches evolve. Rapid and comparatively inexpen- vidual tumors in similar detail to identify the aberrant pathways, we sive targeted or exome sequencing is now becoming available at can deduce the primary drivers and propose effective combinations the lab bench, with bioinformatics software and cloud computing of targeted therapies that can work against that individual tumor. to provide results within days. This will be an exciting era. The Head and Neck Cancer Consortium assessed 279 previ- ously untreated HNSCC tumors (35 of which contained HPV) for genomic alterations.42 Sequence (exome, RNA, miRNA, some CANCER STEM CELLS whole genome sequence), structural, epigenetic (DNA methyla- tion), and copy number analyses as well as protein expression were The isolation of highly tumorigenic subpopulations of cancer carried out. The genomic gains and losses of the smoking-related cells from solid cancers, commonly referred to as cancer stem HNSCC strongly resembled those in the squamous cancers of the cells (CSC), has generated a great deal of interest in the relevance lung59; however, those HNSCC tumors containing HPV had less of these cells to cancer progression and treatment failure. CSCs complex genomic signatures and lacked TP53 mutations, HRAS- have been shown to be critical in the growth and development of activating mutations, and CDKN2A deletions and mutations, but primary tumors and for the development of regional and distant had more frequent PIK3CA mutations. Also of interest was the iden- metastatic disease in HNSCC. CSCs are also believed to be highly tification of a subset of HPV-negative HNSCC that had relatively resistant to conventional therapy and, therefore, likely responsible fewer TP53 mutations and more frequent PIK3CA and HRAS mu- for disease recurrence and treatment failures when they occur. tations. Interestingly the tumors with HRAS mutations tended to CSCs were first isolated from lymphomas. A subsequent in- correlate with wild-type TP53–containing tumors. This set was also vestigation of solid tumors has revealed the presence of highly tu- notable for a less complex pattern of chromosome rearrangements morigenic cancer cells in essentially every solid cancer type. These and better survival than the other HPV-negative HNSCC tumors. highly tumorigenic cells fit the current criteria for being classified as In the entire HNSCC set, the most commonly mutated genes were CSCs: they are tumorigenic; they are able to reproduce the original TP53, FAT1, CDKN2A, PIK3CA, NOTCH1, and MLL2, with tumor heterogeneity, including the tumorigenic and nontumori- mutation frequencies ranging from 72% to 18%. Amplifications genic cell subpopulations; and they are self-renewing. CSCs gen- of receptor tyrosine kinases (RTK) were common. EGFR ampli- erally represent a very small subpopulation of the cancer cells. In fication was identified in 16% of tumors. Other smaller subsets of the majority of solid tumors, they typically represent less than 10% 420 Practice of Oncology / Cancer of the Head and Neck of the entire cancer cell population. A variety of surface markers resulted in a significant reduction in HNSCC CSC renewal and in and biologic markers have been used to isolate the CSC population chemotherapy and radiation resistance.63 Conversely, overexpres- from the other cancer cells in HNSCC, including CD 44, CD 133, sion of BMI1 in the non-CSC population of HNSCC resulted in ESA, and aldehyde dehydrogenase activity. So far, no single marker the acquisition of self-renewal and stem cell–like properties. or combination of markers has proven useful for isolating CSCs Epithelial to mesenchymal transition is required to allow for from every tumor site. Additionally, none of these stem cell markers cancer cell migration to occur and for the subsequent develop- has been found to be a useful or effective target for cancer therapy. ment of metastasis. Snail and Twist have been identified as criti- The term cancer stem cell refers to the biologic behavior of the cal transcription factors regulating epithelial to mesenchymal cells and not their cell of origin. Although CSCs exhibit many transition in stem cells and cancer cells. Increased expression of of the properties of normal stem cells, their cell of origin has yet Twist has been found in both CD44-positive and ALDH-positive to be determined. Likely, CSCs can originate from normal stem HNSCC, and increased Snail expression has been confirmed in cells, early progenitor cells, and possibly more differentiated cells ALDH HNSCC CSCs.64 Snail expression in HNSCC CSCs has if they undergo the correct combination of mutations and epige- been correlated with metastasis, local recurrence, and prognosis. netic changes required to achieve the CSC phenotype. Increased Twist expression increases HNSCC CSC motility and Many normal stem cell genetic pathways are expressed in loss of E-cadherin–mediated cell-to-cell contact. CSCs and are important regulators of their behavior. These genes Normal cells and CSCs express high levels of adenosine triphos- include the transcription factors responsible for maintaining plu- phate (ATP)-binding cassette (ABC) transporter genes, including ripotency, epithelial to mesenchymal transition, self-renewal, xeno- ABCB1, which encodes P-glycoprotein, and ABCG2. The drug-efflux- biotic efflux, and quiescence. Embryonic stem cells are believed to ing property of stem cells conferred by ABC transporters is the basis for be dependent on at least three critical pathways that regulate their the side-population phenotype that arises from the exclusion of the activity: NANOG, Oct-4, and Sox-2. NANOG has been shown to fluorescent dye Hoechst 33342 and has been used to isolate CSCs inhibit differentiation, and Oct-4 is critical to self-renewal. Oct-4 from HNSCC.65 CSCs are likely to share many of the properties of and NANOG have also been recognized as important transcrip- normal stem cells that provide for a long lifespan, including resistance tion factors in the reprogramming of pluripotency in adult cells. to drugs and toxins through the expression of ABC transporters. There- NANOG, Oct-3/4, and Sox-2 have been found to be upregulated fore, tumors might have a built-in population of drug-resistant pluripo- in CSC-enriched cells derived from HNSCC grown as spheroids. tent cells—the CSCs—due to their expression of the ABC-transported Oct-4, which is known to be an important stem cell gene in epithe- genes that can survive chemotherapy and repopulate the tumor. lial stem cells, has been found to be unregulated in aldehyde dehy- The Wnt/β-catenin pathway is known to be important to organo- drogenase (ALDH)-positive HNSCC cells.61 The overexpression genesis and embryonic development. The major function of this of NANOG and Oct-4 has been correlated with chemotherapy pathway to is to regulate β-catenin function by its phosphorylation resistance and the stage of cancer, suggesting that these genes play and proteasomal degradation. Abnormalities in the pathway can lead a role in treatment outcomes and in the prognosis in HNSCC. to β-catenin accumulation, which, in turn, results in the activation BMI1 is a gene that is believed to be an essential stem cell–re- of several pathways such as cyclin D1 and CMYC, which control the lated gene that maintains the self-renewal of stem cells. To date, G1 to S phase transition in the cell cycle. Wnt signaling is also in- although BMI1 has not been shown to be a reliable marker for volved in the critical cell migration process known as the epithelial– CSCs in HNSCC, its potential to induce tumorigenic changes has mesenchymal transition. These effects contribute to the importance made it an attractive target for study in CSCs. BMI1 is expressed in of the Wnt/β-catenin in determining the CSC phenotype. CSCs from a variety of tumor types, including HNSCC, and is an The expression of CSC genes in primary HNSCC has been important component of the polycomb complex 1, an epigenetic found in some studies to be associated with the prognosis and re- regulator of stem cell self-renewal and carcinogenesis.62 Through sponse to therapy. The role of CSCs in the treatment resistance and repression of the CDKN2A locus and inhibition of p16Ink4A ex- recurrence is still being elucidated. Gaining a better understanding pression, BMI1 is believed to promote cellular proliferation by of the molecular pathways that regulate CSC behavior will be essen- blocking p16Ink4A-induced cellular senescence. BMI1 knockdown tial to developing therapies that target this critical population of cells.

SELECTED REFERENCES

The full reference list can be accessed at lwwhealthlibrary.com/oncology. 21. Maxwell JH, Kumar B, Feng FY, et al. Tobacco use in human papillomavirus- positive advanced oropharynx cancer patients related to increased risk of dis- 3. Eliassen AM, Hauff SJ, Tang AL, et al. Head and neck squamous cell carci- tant metastases and tumor recurrence. Clin Cancer Res 2010;16:1226–1235. noma in pregnant women. Head Neck 2013;35:335–342. 22. Worden FP, Kumar B, Lee JS, et al. Chemoselection as a strategy for organ 10. Wreesmann VB, Estilo C, Eisele DW, et al. Downregulation of Fanconi preservation in advanced oropharynx cancer: response and survival positively anemia genes in sporadic head and neck squamous cell carcinoma. ORL J associated with HPV16 copy number. J Clin Oncol 2008;26:3138–3146. Otorhinolaryngol Relat Spec 2007;69:218–225. 23. Hariri S, Unger ER, Sternberg M, et al. Prevalence of genital human pap- 11. van Zeeburg HJ, Snijders PJ, Wu T, et al. Clinical and molecular charac- illomavirus among females in the United States, the National Health And teristics of squamous cell carcinomas from Fanconi anemia patients. J Natl Nutrition Examination Survey, 2003-2006. J Infect Dis 2011;204:566–573. Cancer Inst 2008;100:1649–1653. 24. D’Souza G, Kreimer AR, Viscidi R, et al. Case-control study of human papil- 13. Chaturvedi AK, Engels EA, Anderson WF, et al. Incidence trends for human lomavirus and oropharyngeal cancer. N Engl J Med 2007;356:1944–1956. papillomavirus-related and -unrelated oral squamous cell carcinomas in the 25. Walline HM, Komarck C, McHugh JB, et al. High-risk human papillomavi- United States. J Clin Oncol 2008;26:612–619. rus detection in oropharyngeal, nasopharyngeal, and oral cavity cancers: com- 14. Kreimer AR, Clifford GM, Boyle P, et al. Human papillomavirus types in parison of multiple methods. JAMA Otolaryngol Head Neck Surg 2013;139: head and neck squamous cell carcinomas worldwide: a systematic review. 1320–1327. Cancer Epidemiol Biomarkers Prev 2005;14:467–475. 26. Maxwell JH, Kumar B, Feng FY, et al. HPV-positive/p16-positive/EBV- 15. Gillison ML, Alemany L, Snijders PJ, et al. Human papillomavirus and dis- negative nasopharyngeal carcinoma in white North Americans. Head Neck eases of the upper airway: head and neck cancer and respiratory papillomato- 2010;32:562–567. sis. Vaccine 2012;30 Suppl 5:F34–F54. 28. Gillison ML, Broutian T, Pickard RK, et al. Prevalence of oral HPV infection 16. Chaturvedi AK, Engels EA, Pfeiffer RM, et al. Human papillomavirus and in the United States, 2009–2010. JAMA 2012;307:693–703. rising oropharyngeal cancer incidence in the United States. J Clin Oncol 29. Fakhry C, Gillison ML. Clinical implications of human papillomavirus in 2011;29:4294–4301. head and neck cancers. J Clin Oncol 2006;24:2606–2611. 17. Gillison ML. Human papillomavirus-associated head and neck cancer is a 30. Fakhry C, Westra WH, Li S, et al. Improved survival of patients with human distinct epidemiologic, clinical, and molecular entity. Semin Oncol 2004;31: papillomavirus-positive head and neck squamous cell carcinoma in a pro- 744–754. spective clinical trial. J Natl Cancer Inst 2008;100:261–269.

tahir99 - UnitedVRG Chapter 37 Molecular Biology of Head and Neck Cancers 421

32. Worden FP, Kumar B, Lee JS, et al. Chemoselection as a strategy for organ 48. Bonner JA, Harari PM, Giralt J, et al. Radiotherapy plus cetuximab for lo- preservation in advanced oropharynx cancer: response and survival positively coregionally advanced head and neck cancer: 5-year survival data from a associated with HPV16 copy number. J Clin Oncol 2008;26:3138–3146. phase 3 randomised trial, and relation between cetuximab-induced rash and 34. Ang KK, Harris J, Wheeler R, et al. Human papillomavirus and survival of survival. Lancet Oncol 2010;11:21–28. patients with oropharyngeal cancer. N Engl J Med 2010;363:24–35. 53. Bradford CR, Kumar B, Belillile E, et al. Biomarkers in advanced larynx 35. Adelstein DJ, Ridge JA, Brizel DM, et al. Transoral resection of pharyngeal cancer. Laryngoscope 2014;124:179–187. cancer: summary of a National Cancer Institute Head and Neck Cancer 55. Kumar B, Cordell KG, D’Silva N, et al. Expression of p53 and Bcl-xL as Steering Committee Clinical Trials Planning Meeting, November 6-7, 2011, predictive markers for larynx preservation in advanced laryngeal cancer. Arch Arlington, Virginia. Head Neck 2012;34:1681–1703. Otolaryngol Head Neck Surg 2008;134:363–369. 36. Kimple RJ, Harari PM. Is radiation dose reduction the right answer for HPV- 57. Hahn WC, Counter CM, Lundberg AS, et al. Creation of human tumour positive head and neck cancer? Oral Oncol 2013 Oct 14. cells with defined genetic elements. Nature 1999;400:464–468. 37. Duray A, Descamps G, Decaestecker C, et al. Human papillomavirus DNA 60. Liu J, Pan S, Hsieh MH, et al. Targeting Wnt-driven cancer through the in- strongly correlates with a poorer prognosis in oral cavity carcinoma. Laryngo- hibition of Porcupine by LGK974. Proc Natl Acad Sci U S A 2013;110: scope 2012;122:1558–1565. 20224–20229. 38. Stenmark MH, McHugh JB, Schipper M, et al. Nonendemic HPV-positive 63. Chen YC, Chang CJ, Hsu HS, et al. Inhibition of tumorigenicity and en- nasopharyngeal carcinoma: association with poor prognosis. Int J Radiation hancement of radiochemosensitivity in head and neck squamous cell cancer- Oncol Biol Phys 2014;88:580–588. derived ALDH1-positive cells by knockdown of Bmi-1. Oral Oncol 2010;46: 41. Spector ME, Gallagher KK, Light E, et al. Matted nodes: poor prognostic 158–165. marker in oropharyngeal squamous cell carcinoma independent of HPV and 64. Yu CC, Lo WL, Chen YW, et al. Bmi-1 regulates snail expression and pro- EGFR status. Head Neck 2012;34:1727–1733. motes metastasis ability in head and neck squamous cancer-derived ALDH1 42. Hayes N, consortium T. Comprehensive genomic characterization of head positive cells. J Oncol 2011;2011. and neck squamous cell carcinomas. Nature 2013. 65. Clay MR, Tabor M, Owen JH, et al. Single-marker identification of head and 47. Rubin Grandis J, Melhem MF, Gooding WE, et al. Levels of TGF-alpha neck squamous cell carcinoma cancer stem cells with aldehyde dehydroge- and EGFR protein in head and neck squamous cell carcinoma and patient nase. Head Neck 2010;32:1195–1201. survival. J Natl Cancer Inst 1998;90:824–832. PRACTICE OF ONCOLOGY PRACTICE Cancer of the Head 38 and Neck

William M. Mendenhall, John W. Werning, and David G. Pfister

INCIDENCE AND ETIOLOGY or perichondrium. There are no capillary lymphatics in the eye, and few in the orbit. The estimated number of new head and neck cancer cases (ex- Most head and neck malignant neoplasms arise from the sur- cluding skin cancer) in the United States in 2013 is 53,640; this face epithelium and are squamous cell carcinoma (SCC) or one of represents 3.2% of the total new cancer cases.1 Approximately 27% its variants, including lymphoepithelioma, spindle cell carcinoma, of these patients are women.1 African Americans have a higher verrucous carcinoma, and undifferentiated carcinoma. Lympho- age-adjusted incidence than other ethnic groups. The usual time mas and a wide variety of other malignant and benign neoplasms of diagnosis is after the age of 40, except for salivary gland and make up the remaining cases.14–16 nasopharyngeal cancers (NPCs), which may occur in younger age Lymphoepithelioma is an SCC with a lymphoid stroma and groups. For many primary sites, tobacco use is associated with an occurs in the nasopharynx, tonsillar fossa, and base of tongue; it increased risk. Alcohol has also been implicated as a causative fac- may also occur in the salivary glands. In the spindle cell variant, tor; the effects of alcohol and tobacco may be synergistic.2 Head there is a spindle cell component that resembles sarcoma inter- and neck cancer patients have an increased risk for developing mixed with SCC. It is generally managed like other high-grade a second primary tumor (SPT), both within the head and neck SCCs. Verrucous carcinoma is a low-grade SCC found most often and elsewhere (e.g., esophageal and lung cancers),3 attributed to in the oral cavity, particularly on the gingiva and buccal mucosa. It the field effect associated with tobacco and alcohol use.4 Human usually has an indolent growth pattern and is often associated with papillomavirus infection (HPV; most commonly HPV-16) plays a the chronic use of snuff or chewing tobacco. role in the development of certain head and neck cancers, particu- Small cell neuroendocrine carcinoma occurs rarely through- larly those in the oropharynx.5,6 Patients with high-risk HPV (HR- out the head and neck. Upper aerodigestive tract lymphomas al- HPV)–positive head and neck cancers tend to be younger and less most always show a diffuse non-Hodgkin histologic pattern. likely to have a strong history of tobacco and ethanol use, have a history of multiple sex partners (particularly oral-genital sex), have a better prognosis, and appear to have a lower rate of SPTs.6–8 Prior NATURAL HISTORY tobacco exposure adversely affects the prognosis of HPV-related oropharynx cancers.8,9 An increasing incidence of oral tongue squamous cell carcinoma in nonsmoking Caucasian women has Patterns of Spread been reported that does not appear to be driven by prior HPV infection, whereas the incidence of other oral cavity cancers is Primary Lesion 10 declining. There is a long-standing association between Epstein- SCCs usually begin as surface lesions. Superficial tumors arising Barr virus (EBV) and NPC.11 Occupational exposures are associ- 12 in the Waldeyer ring may be difficult to distinguish from normal ated with the development of sinonasal tract tumors. lymphoid tissue. Very early surface lesions may show only ery- thema and a slightly elevated mucosa. Spread is dictated by local anatomy, and thus varies by each ANATOMY AND PATHOLOGY site. Muscle invasion is common, and the tumor may spread along muscle or fascial planes a surprising distance from the palpable The anatomy pertaining to a particular primary site is described or visible lesion. The tumor may attach early to the periosteum in subsequent sections. To facilitate communication, lymph or perichondrium, but bone or cartilage invasion is usually a late nodes are organized into levels. Level I includes the submental event. Bone and cartilage usually act as a barrier to spread; the and submandibular areas; levels II through IV include the inter- tumor that encounters these structures will often be diverted and nal jugular vein lymph nodes; level V includes the posterior tri- spread along a path of less resistance. Slow-growing gingival neo- angle (Fig. 38.1).13 Furthermore, which lymph node levels are plasms may produce a smooth pressure defect of the underlying involved is predictive of the primary site. For example, lip, oral bone without bone invasion. cavity, and facial skin tumors typically spread to levels I and II ini- Tumor extension into the parapharyngeal space allows superior tially; larynx and pharynx cancers have a predilection for spread or inferior spread from the skull base to the low neck. to levels II and III. Spread inside the lumen of the sublingual, submandibular, and There are no capillary lymphatics in the epithelium. The parotid gland ducts is uncommon. The nasolacrimal duct, how- tumor must penetrate the lamina propria before lymphatic in- ever, is often invaded in ethmoid sinus and nasal carcinomas. vasion can occur. One can predict the richness of the capillary Perineural invasion (PNI) is observed in SCCs as well as sali- network in a given head and neck site by the relative incidence vary gland tumors, especially adenoid cystic carcinomas. When ad- of lymph node metastases at presentation. The nasopharynx and vanced, PNI may produce neurologic symptoms and is associated pyriform sinus have the most profuse capillary lymphatic net- with a poorer rate of local control.17 Tumors may track along a nerve works. The paranasal sinuses, middle ear, and vocal cords have to the skull base and central nervous system (CNS) or peripherally. few or no capillary lymphatics. Muscle and fat contain few capil- Vascular space invasion is associated with an increased risk for lary lymphatics, as do bone and cartilage within the periosteum regional and distant metastases. 422

tahir99 - UnitedVRG Chapter 38 Cancer of the Head and Neck 423

the primary tumor and any clinically positive lymph nodes is documented. Almost all patients undergo contrast enhanced computed tomography (CT) and/or magnetic resonance imag- ing (MRI) to further define the extent of local–regional disease. The scan(s) should be obtained prior to biopsy so that biopsy changes are not confused with the tumor. A chest radiograph is obtained to determine the presence of distant metastases and/ or a synchronous primary lung cancer. Patients with N3 neck disease, as well as those with N2 disease with nodes below the level of the thyroid notch, have a 20% to 30% risk of developing distant metastases and are considered for a chest CT or positron emission tomography (PET). 2 Tumors amenable to transoral biopsy may be biopsied using local anesthetics in the clinic. Otherwise, direct laryngoscopy 1 under anesthesia is performed to determine the extent of the 3 tumor and to obtain a tissue diagnosis. Given the risk of syn- 5 chronous cancers, some advocate routine triple endoscopy (i.e., laryngoscopy/pharyngoscopy, bronchoscopy, and esopha- goscopy). The additional yield is low, unless diffuse mucosal abnormalities or a malignant lymph node without an identified 4 primary site, particularly in the low neck, are present. Patients presenting with a metastatic node from an unknown primary site undergo fine-needle aspiration (FNA) of the node. An ex- Figure 38.1 Head and neck lymph node levels. cisional biopsy is not routinely performed unless lymphoma is suspected or FNA results are equivocal. If SCC is a consider- ation, the excision should be done in a manner to facilitate subsequent management, including neck dissection. Occasion- Lymphatic Spread ally, the diagnosis may be made by clinical and radiographic evaluation, and a biopsy should be avoided in situations where The differentiation of the tumor, the size of the primary lesion, the the treatment is definitive for RT and where obtaining a tissue presence of vascular space invasion, and the density of capillary sample is risky (e.g., paragangliomas, juvenile nasopharyngeal lymphatics predict the risk of lymph node metastasis. Recurrent le- 14,25 angiofibromas). OF ONCOLOGY PRACTICE sions have an increased risk. Histology also impacts the likelihood Before the initial treatment, the patient should be evaluated by of lymphatic spread. Low-grade minor salivary gland tumors and members of the team who may be involved in the initial manage- sarcomas have a lower risk of lymph node metastases than SCCs ment as well as possible salvage therapy. Head and neck surgeons, arising in similar mucosal sites. radiation oncologists, medical oncologists, diagnostic radiologists, A patient may present with SCC in a cervical lymph node, and plastic surgeons, pathologists, dentists, speech and swallowing despite an extensive work-up, the site of origin may remain unde- therapists, and social workers may all play a role. termined in approximately 50% of patients.18 If only the neck is treated, a primary lesion may appear later, but sometimes is never 19 found. STAGING The relative incidence of clinically positive lymph nodes on ad- mission is determined by primary site and T stage.20 Well-lateralized The staging system of the American Joint Committee on Can- lesions spread to ipsilateral neck nodes.21 Lesions on or near the cer (AJCC) is used.26 In general, excisional biopsy (TX) indicates midline, tongue base, and nasopharyngeal lesions (even when lat- that the primary tumor cannot be assessed; T0 indicates no evi- eralized), may spread to both sides of the neck, although the risk is dence of primary tumor; and Tis indicates carcinoma in situ. For higher to the side occupied by the bulk of the lesion. Patients with tumors of the oral cavity and oropharynx, further staging of the clinically positive ipsilateral neck nodes are at risk for contralateral primary lesion is based primarily on size criteria: 2 cm or less for disease, especially if the nodes are large or multiple; obstruction of T1; greater than 2 cm but no more than 4 cm for T2; greater than the lymphatic pathways by surgery or radiotherapy (RT) also shunts 4 cm for T3; and T4 tumors involve major invasion or encase- the lymphatic flow to the opposite neck.21 When lymph node me- ment of surrounding structures (e.g., bone, carotid artery, deep tastases appear at an unusual site, a careful search must be made for musculature). For the other primary sites, further staging is less a second primary. The likelihood of retropharyngeal adenopathy is easily generalized because the anatomic extent of spread and/or related to the presence of clinically involved lymph nodes and the functional criteria (e.g., vocal cord mobility) are used and, for primary site, and is particularly high for NPCs.22 certain sites, are combined with tumor size (e.g., hypopharynx, Distant Spread major salivary glands), and so will be given in the discussion of each respective primary site.26 Neck staging is common to all The risk of distant metastasis is related more to N stage and the head and neck sites, except the nasopharynx (Table 38.1).26 Le- location of involved nodes in the low neck, rather than to T stage.23 sions may be clinically or pathologically staged. Clinical staging The risk is less than 10% for N0 or N1 disease and rises to approxi- is more commonly used for treatment planning and the report- mately 30% for N3 disease as well as N1 or N2 nodes with disease ing of results. The format for combining T and N stages into an below the level of the thyroid notch. Distant metastases are found overall stage is depicted in Table 38.2 and is common to all sites most often in the lung.24 except the nasopharynx.26 Stage IV represents a wide spectrum of disease. One patient may have a T1, T2, or T3 lesion with low-volume N2 neck disease DIAGNOSIS and a high probability of cure (stage IVA), whereas another may have a T4b primary cancer and/or N3 neck disease and a relatively A general medical evaluation is performed, including a thor- poor prognosis (stage IVB)27; distant metastases indicate stage IVC ough head and neck examination. The location and extent of disease, and the treatment intent is typically palliative. 424 Practice of Oncology / Cancer of the Head and Neck

TABLE 38.1 of immediate and late RT sequelae is avoided, and (4) RT is reserved for a head and neck SPT, which may not be as suitable for surgery. 2010 American Joint Committee on Cancer Stages of The advantages of RT may include (1) the risk of a major post- Regional Lymph Node Involvement operative complication is avoided, (2) no tissues are removed so that the probability of a functional or cosmetic defect may be re- NX Regional lymph nodes cannot be assessed duced, (3) elective neck RT can be included with little added mor- bidity, and (4) the surgical salvage of RT failure is probably more N0 No regional lymph node metastasis likely than the salvage of a surgical failure. N1 Metastasis in a single ipsilateral lymph node, 3 cm or less Salvage of a surgical failure may be attempted by operation, in greatest dimension RT, or both. Surgical recurrences usually develop at the resection N2 Metastasis in single ipsilateral lymph node, more than 3 margins, in or near the suture line. It is difficult to distinguish the cm but no more than 6 cm in greatest dimension; or in normal surgical scarring from recurrent disease, and the diagnosis multiple ipsilateral lymph nodes, none more than 6 cm in of recurrence is often delayed. Tumor response to RT under these greatest dimension; or in bilateral or contralateral lymph circumstances is poor. Surgery, RT, or both, however, may salvage nodes, no more than 6 cm in greatest dimension small mucosal recurrences and some neck recurrences. For bulk- ier recurrences treated with RT, concurrent chemotherapy is often N2a Metastasis in single ipsilateral lymph node, more than 3 incorporated. cm but no more than 6 cm in greatest dimension N2b Metastasis in multiple ipsilateral lymph nodes, none more than 6 cm in greatest dimension MANAGEMENT N2c Metastasis in bilateral or contralateral lymph nodes, none more than 6 cm in greatest dimension Primary Site N3 Metastasis in a lymph node more than 6 cm in greatest dimension The management of the primary cancer will be considered separately for each anatomical site. Patients who are in poor nu- Used with the permission of the American Joint Committee on Cancer (AJCC), Chicago, Illinois. The original source for this material is the AJCC tritional condition may require a nasogastric (NG) tube or a per- Cancer Staging Handbook, Seventh Edition (2010) published by Springer cutaneous gastrostomy (PEG) before initiating RT, particularly Science and Business Media LLC, www.springerlink.com. if concomitant chemotherapy is used. Opinions vary regarding the role of prophylactic NG or PEG placement in anticipation of RT-based local toxicity in patients without significant baseline dysphagia or weight loss; a reactive strategy is preferred by many PRINCIPLES OF TREATMENT FOR 28 SQUAMOUS CELL CARCINOMA and may facilitate swallowing recovery. If external-beam radio- therapy (EBRT) is selected, it may be given with either conven- tional once-daily fractionation, 66 to 70 Gy at 2 Gy per fraction, General Principles for Selection of Treatment 5 days a week in a continuous course, or with an altered frac- tionation schedule. Whether an altered fractionation schedule Surgery and RT are the only curative treatments for head and neck is better than conventional fractionation depends on the altered carcinomas. Although chemotherapy alone is not curative, it en- fractionation technique that is selected. Two altered fraction- hances the effects of RT, and thus is routinely used as part of com- ation schedules shown to result in improved local–regional con- bined modality treatment in patients with stage III or IV disease. trol rates are the University of Florida hyperfractionation and the The advantages of surgery compared with RT, assuming similar M.D. Anderson concomitant boost techniques.29 The results of cure rates, may include the following: (1) a limited amount of tissue a prospective randomized Radiation Therapy Oncology Group is exposed to treatment, (2) the treatment time is shorter, (3) the risk (RTOG) trial comparing these schedules with conventional frac- tionation and the Massachusetts General Hospital accelerated split-course schedule are shown in Table 38.3. Acute toxicity is TABLE 38.2 increased with altered fractionation; late toxicity is comparable with conventional fractionation.30 2010 American Joint Committee on Cancer Overall Conventional EBRT techniques and/or brachytherapy will be Stage Grouping discussed in the subsequent site-specific sections. EBRT may also be delivered with intensity-modulated radiation therapy (IMRT) to produce a more conformal dose distribution and to reduce the Stage 0 Tis N0 M0 dose to the normal tissues.31–33 The disadvantages of IMRT are that Stage I T1 N0 M0 it is more time consuming to plan and treat the patient, the dose Stage II T2 N0 M0 distribution is often less homogenous so that “hot spots” may in- crease the risk of late complications, the risk of a marginal miss Stage III T3 N0 M0 may be increased because the fields are more conformal, the total T1–T3 N1 M0 body RT dose is higher because of increased “beam on” time and Stage IVA T4a N0–N1 M0 scatter irradiation, and it is more costly. Therefore, a clear reason for using IMRT versus conventional RT must be identified. The T1–T4a N2 M0 usual indication for IMRT is to reduce the dose to the contralateral 34 Stage IVB Any T N3 M0 parotid gland, and thus limit long-term xerostomia. Another indi- cation is to reduce the CNS dose in patients with NPC. Finally, it T4b Any N M0 may be used to avoid a difficult low neck match in patients with la- Stage IVC Any T Any N M1 ryngeal or hypopharyngeal cancers and a low-lying larynx. Proton therapy, which offers potential targeting and dosing advantages for Used with the permission of the American Joint Committee on Cancer 35 (AJCC), Chicago, Illinois. The original source for this material is the AJCC selected tumors, is useful for reducing the dose to the brain and Cancer Staging Handbook, Seventh Edition (2010) published by Springer the visual apparatus for patients with nasal cavity and paranasal Science and Business Media LLC, www.springerlink.com. sinus malignancies.

tahir99 - UnitedVRG Chapter 38 Cancer of the Head and Neck 425

TABLE 38.3 Altered Fractionation: 5-Year Outcomes from the Radiation Therapy Oncology Group 90-03 Trial

Fractionation Schedule Accelerated Split Accelerated Conventional Hyperfractionation Course Concomitant Boost Parameter (70 Gy/35 Fx/7 wk) (81.6 Gy/68 Fx/7 wk) (67.2 Gy/42 Fx/6 wk) (72 Gy/42 Fx/6 wk) Number of patients 268 263 274 268 Local–regional failure 59.1% 51.2% (p = 0.037) 57.8% (p = 0.042) 51.7% Disease-free survival 21.2% 30.7% (p = 0.013) 26.6% (p = 0.042) 28.9% Overall survival 29.5% 37.1% (p = 0.063) 30.8% 33.5% Cause-specific survival 42.9% 45.5% 40.9% 43.4% Grade 3 late toxicity 25.2% 27.4% 26.8% 33.3% (p = 0.066)

Note: P values reflect comparison of the experimental arms with standard fractionation. Fx, fractions. From Trotti A, Fu KK, Pajak TF, et al. Long term outcomes of RTOG 90-03: a comparison of hyperfractionation and two variants of accelerated fractionation to standard fractionation radiotherapy for head and neck squamous cell carcinoma. Int J Radiat Oncol Biol Phys 2005;63:S70–S71.

NECK The more extensive the neck dissection, the higher the risk of complications. Complications after neck dissection include hema- toma, seroma, lymphedema, wound infections and dehiscence, In a classic radical neck dissection, the superficial and deep cervi- damage to the 7th, 10th, 11th, and 12th cranial nerves, carotid cal fascia with its enclosed lymph nodes (levels I to V) is removed exposure, and carotid rupture. The last-mentioned complication in continuity with the sternocleidomastoid muscle, the omohyoid can be minimized by covering the carotid artery with a dermal muscle, the internal and external jugular veins, cranial nerve XI, and graft at the time of surgery.36 Pain and dysfunction in the neck or the submandibular gland. The incisions used by the surgeon will be

shoulder may occur. Rehabilitation and anti-inflammatory medi- OF ONCOLOGY PRACTICE governed largely by the primary lesion. The radical neck dissection cation are commonly utilized with varying benefits; acupuncture can be modified to spare certain structures with the intent of decreas- had demonstrated a benefit compared to the usual care in one ing morbidity and improving functional outcome without compro- randomized study.38 mising disease control. There are three main types of modified radical neck dissections: type I, CN XI is spared; type II, CN XI and the internal jugular vein are spared; and type III (functional), CN Clinically Negative Neck XI, the internal jugular vein, and the sternocleidomastoid muscle are spared. Selective neck dissections are more limited and include The estimated incidence of subclinical disease in the regional lym- the resection of lymph node levels that are at greatest risk for nodal phatics when the neck is cN0 is presented in Table 38.4.39 Both RT metastatic spread. Examples include the lateral, posterolateral, and and neck dissection are approximately 90% efficient at eradicating supraomohyoid, which include resections of lymph node levels II subclinical regional disease.36 Alternatively, a policy of close ob- through IV, II through V, and I through III, respectively. servation may be adopted for the cN0 neck to avoid unnecessary A modified or selective neck dissection is recommended for the treatment, and the neck is managed by surgery and/or RT if cervi- cN0 neck, for selected clinically positive necks (mobile, 1 to 3 cm cal metastases develop. The salvage rate for patients developing lymph nodes), and for removing residual disease after RT when clinically positive lymph nodes with the primary lesion controlled there has been excellent regression of N2 or N3 disease.36,37 is 50% to 60%.39

TABLE 38.4 Definition of Risk Groups for the Clinically N0 Neck

Estimated Risk of Group Subclinical Neck Disease T Stage Site I: Low risk <20% T1 Floor of mouth, oral tongue, retromolar trigone, gingiva, hard palate, buccal mucosa II: Intermediate risk 20%–30% T1 Soft palate, pharyngeal wall, supraglottic larynx, tonsil T2 Floor of mouth, oral tongue, retromolar trigone, gingiva, hard palate, buccal mucosa III: High risk >30% T1–T4 Nasopharynx, pyriform sinus, base of tongue T2–T4 Soft palate, pharyngeal wall, supraglottic larynx, tonsil T3–T4 Floor of mouth, oral tongue, retromolar trigone, gingiva, hard palate, buccal mucosa

Reprinted with permission from Mendenhall WM, Million RR. Elective neck irradiation for squamous cell carcinoma of the head and neck: analysis of time-dose factors and causes of failure. Int J Radiat Oncol Biol Phys 1986;12:741–746. 426 Practice of Oncology / Cancer of the Head and Neck

TABLE 38.5 Failure of Initial Neck Treatment (596 Patients with Carcinoma of the Tonsillar Fossa, Base of Tongue, Supraglottic Larynx, or Hypopharynx at M.D. Anderson Hospital 1948–1967)

Stage N0 No Treatment Treatment Partial Complete N1 N2A N2B N3A N3B Radiation 15% 2% 15% 27% 27% 38% 34% Surgery 55% (16/29) 35% 7% 11% 8% 23% 42% 41% Combined 1/5 0/6 0 0 0 23% 25%

Adapted from Barkley HT Jr, Fletcher GH, Jesse RH, et al. Management of cervical lymph node metastases in squamous cell carcinoma of the tonsillar fossa, base of tongue supraglottic larynx, and hypopharynx. Am J Surg 1972;124:462–467.

Elective neck irradiation (ENI) and elective neck dissection are Modified neck dissection is sufficient treatment for the ipsilat- equally effective in the management of the N0 neck, with control eral neck for patients with N1 or N2A disease without ECE. RT, rates exceeding 90%.39,40 Treatment of the entire neck is advised often combined with concurrent chemotherapy, is added for those for primary lesions with a high rate of subclinical disease, such with more advanced neck disease.42 as the base of tongue, soft palate, supraglottis, and hypopharynx. When the primary lesion is to be managed by RT or chemora- Patients with lateralized T1 to T2 tonsillar cancers do not require diotherapy (chemoRT), then RT-based therapy alone is sufficient elective treatment for the contralateral N0 neck41; T3 or T4 can- for patients in whom the node(s) regress completely as docu- cers or those with significant extension into the tongue and/or soft mented on CT obtained 4 weeks post-RT.37,48 RT is followed by palate should receive bilateral neck treatment to the entire neck. a neck dissection for patients with residual nodes that are 1.5 cm When the primary tumor is to be treated surgically, an elec- or larger, as well as those that demonstrate focal defects, enhance- tive neck dissection should be performed when the risk of regional ment, and/or calcification.48 A PET scan done 3 months after RT lymph node metastasis is 10% to 15% or greater. Modified neck is completed is an alternative to CT to assess whether there is per- dissection has a good rate of disease control; patients who are sistent disease.49 found to have multiple positive nodes or extracapsular extension McGuirt and McCabe50 compared results of definitive surgery (ECE) are then referred for postoperative RT,42 and concurrent with and without a prior open neck biopsy and concluded the risks chemotherapy is recommended in the latter circumstance.43–46 If of neck failure, distant metastases, and complications were all in- the primary lesion is to be treated with EBRT, ENI adds relatively creased. Ellis et al.51 studied the results of therapy following open little cost and modest morbidity. biopsy of a lymph node before treatment. Patients received defini- tive RT to the primary site and neck; a subset of patients underwent a neck dissection after RT. Open biopsy had no adverse impact Clinically Positive Neck Lymph Nodes on these patients compared with those who did not undergo an open biopsy.51 Therefore, after open biopsy of the neck, RT-based The rates of neck failure by N stage and treatment group reported therapy is recommended as the initial treatment, particularly if the from the M.D. Anderson Cancer Center and the University of primary tumor is to be managed by RT or chemoRT. Under these Florida are shown in Tables 38.5 and 38.6, respectively.40,47 In gen- circumstances, no further neck treatment is needed if the neck eral, RT precedes surgery if the primary site is to be treated by RT node had been removed; if there was residual gross tumor in the or if the node was fixed. The operation precedes RT if the primary neck after open biopsy, a planned neck dissection should be added site is to be treated surgically. depending on the results of radiologic reassessment.48

TABLE 38.6 5-Year Rate of Neck Control According to the 1983 American Joint Committee on Cancer Stage and Treatment (459 Patients, 593 Heminecksa)

RT Alone RT + Neck Dissection Stage No. of Heminecks Control No. of Heminecks Control Significance N1 215 86% 38 93% p = 0.28 N2A 29 79% 24 68% p = 0.60 N2B 138 70% 80 91% p < 0.01 N3A 29 33% 40 69% p < 0.01

Note: The University of Florida data, patients were treated October 1964 to October 1985; analysis occurred December 1988 by Eric R. Ellis, MD. a Excludes 67 heminecks that received incisional or excisional biopsies before treatment. From Mendenhall WM, Parson JT, Mancuso AA, et al. Head and neck: management of the neck. In: Perez CA, Brady LW, eds. Principles and Practice of Radiation Oncology, 2nd ed. Philadelphia: J.B. Lippincott Company; 1992: 790–805; and American Joint Committee on Cancer. Manual for Staging of Cancer. Philadelphia: J. B. Lippincott Company; 1983.

tahir99 - UnitedVRG Chapter 38 Cancer of the Head and Neck 427

CHEMOTHERAPY selected patients may derive apparent significant prolongations in survival, average survival improvements appear small at best. Drug therapy may be administered to prevent the development Morton et al.65 reported a 2-month improvement in median sur- of SPTs (chemoprevention), to palliate symptoms in patients with vival after treatment with cisplatin, with or without bleomycin, incurable disease, to improve the odds of cure or organ preserva- compared to no treatment. The duration of responses is typically tion when combined with definitive local-regional therapy, or to measured in weeks to months, not years; survival beyond 2 years decrease treatment toxicity. The first two indications are discussed is infrequent; and cures are anecdotal. Thus, the primary intent of here; the last two are discussed in a subsequent section. chemotherapy in this setting is to achieve tumor regression with the hope that the potential palliative benefit and possible modest survival improvement will outweigh the side effects of treatment. Chemoprevention A number of drugs have been demonstrated in clinical trials to have activity in head and neck SCCs, and the list is well sum- Chemoprevention is the administration of natural or synthetic marized in prior reviews.63,64 The most commonly used include agents to reduce the risk of developing SPTs. Patients who have a methotrexate, cisplatin, carboplatin, 5-fluorouracil, paclitaxel, and head and neck SCC have an increased risk of developing an upper docetaxel, with reported major response rates ranging from 15% aerodigestive tract SPT because of exposure to carcinogens and/or to 42%. Among other drugs with reported major response rates of genetic predisposition.3 The risk of developing an SPT is approxi- 52 15% or greater are bleomycin, cyclophosphamide, doxorubicin, mately 2.7% to 4% per year and may impact survival. Current data hydroxyurea, ifosfamide, irinotecan, oral uracil, ftorafur (with leu- indicate the risk of SPTs is lower among patients with HR-HPV–re- 6,8 covorin), , vinblastine, and vinorelbine. Some of these lated head and neck cancers. Analogs of vitamin A, particularly of agents (e.g., cyclophosphamide, doxorubicin, hydroxyurea) have the retinoids, have been a particular focus of clinical investigations. their activity based on reported assessments in a limited number At present, there is no standard role for the use of chemopreventive of patients from over 2 decades ago, an era when methods and agents in the management of head and neck cancer. criteria for response assessments may have differed from current Retinoids and beta-carotene both may cause regression of oral 53 standards. Anticipated response rates and toxicity profiles may vary leukoplakia; the former appear more efficacious. Lesions com- based on patient selection and drug schedule. A poor performance monly recur after cessation of drug therapy. Chemoprevention status is associated with both lower response rates and greater po- agents do not reduce the risk of recurrence of the index cancer. 2 tential for toxicity. The larger the amount of prior treatment also High-dose13-cis-retinoic acid (100 mg/m daily for 12 months) adversely affects response rates.64 has been shown in a randomized, placebo-controlled trial to re- Methotrexate is a historic standard drug used in the recurrent duce the risk of SPTs in patients previously treated for stage I to 54 or metastatic disease setting. The typical standard dosing is 40 mg/ IV, M0, head and neck cancer. However, a large, placebo-con- m2 intravenously weekly, with dose attenuation or increase (up trolled, randomized trial in 1,190 survivors of stage I and II head 2 to 60 mg/m ) based on toxicity, with mucositis being a frequent OF ONCOLOGY PRACTICE and neck SCC found no difference in the rate of SPTs or survival 55 reason for dose adjustment. The favorable side effect profile and after 3 years of a low-dose schedule of this agent (30 mg per day). convenience of administration of methotrexate make it well-suited Outcomes were better among nonsmokers and those who quit for use in this patient population in which medical comorbidity is compared to smokers, emphasizing the important role of tobacco common, as is more advanced age. In randomized trials, higher cessation as part of head and neck cancer management and strate- doses increase response rates and toxicity without a significant im- gies to decrease SPTs. Similarly, etretinate was not shown to be 66,67 56 provement in overall survival. Newer analogs of methotrexate efficacious in decreasing SPTs. (e.g., edatrexate) have not been shown to offer a therapeutic advan- With regard to other agents, vitamin A, N-acetylcysteine, both, tage in phase III trials.68 or neither were evaluated using a factorial design in the EU- Cisplatin is a cornerstone drug in the modern management of ROSCAN study. No significant improvement in survival or SPTs 57 58 head and neck cancer. Cisplatin is customarily dosed at 75 to 100 was observed. Bairati et al. randomized head and neck cancer mg/m2 intravenously every 3 to 4 weeks. The potential for renal survivors to 3 years of therapy with alpha-tocopherol and beta- (i.e., increase in creatinine, electrolyte abnormalities), otologic carotene versus placebo; the rate of SPTs was actually higher dur- (i.e., high frequency hearing loss, tinnitus), neurologic (i.e., pe- ing the period of treatment, a difference that did not persist with ripheral neuropathy), and gastrointestinal (i.e., nausea and vomit- longer follow-up. A randomized phase II, placebo-controlled trial ing) toxicity are widely appreciated, but these risks are manageable demonstrated no significant benefit of celecoxib at 100 mg or 200 59 if patients are appropriately screened for therapy, monitored mg, both twice daily, on the control of oral premalignant lesions. closely, and state of the art supportive care measures are applied. Targeting the epidermal growth factor receptor (EGFR) path- Further dose escalation of cisplatin has not been established to way is receiving attention, because there is an association between improve outcome. A randomized trial comparing 60 mg/m2 versus progressive EGFR dysregulation and the transition from normal 2 60 120 mg/m of cisplatin failed to demonstrate a significant improve- mucosa to dysplasia to SCC. The concept of bioadjuvant ther- 69 55 ment in response or survival. Carboplatin is the best studied and apy, whereby drug combinations intended to reduce both the risk most commonly used platinum analog in head and neck can- of SPTs and relapse from the index cancer, as well as the potential 61 cer. Although generally less toxic and easier to administer than role of natural extracts are also of interest. the parent drug, it is more bone marrow suppressive and may be There is no standard role for the use of HR-HPV vaccination in 62 somewhat less active. This last issue is more of a concern in the the prevention of head and neck cancer at this time, although the definitive treatment setting in which cure is a central endpoint, impact of current vaccination programs on the incidence of head as opposed to the palliative setting, when patients often seek a less and neck cancer warrants follow-up. toxic alternative treatment. Although taxanes as a class have significant activity in head and Chemotherapy for Recurrent or neck SCCs, hopes of clinically significant improvement in sur- Metastatic Disease vival in the palliative setting with the introduction of these agents have yet to be realized. Neither paclitaxel or docetaxel has been Single Agents demonstrated in random assignment trials to be clearly superior to methotrexate with regard to survival as an endpoint.70 Paclitaxel Patients with recurrent or metastatic head and neck SCCs have dosed at 250 mg/m2 intravenously over 24 hours with growth fac- a median survival of 6 to 9 months, and a 1-year survival rate of tor support in an Eastern Cooperative Oncology Group (ECOG) 20% to 40% when treated with chemotherapy alone.63,64 Although trial, yielded a major response in 12 of 30 patients (40%) including 428 Practice of Oncology / Cancer of the Head and Neck four complete responses; grade 3 or greater neutropenia occurred Major response rates and median survivals ranged from 0% to 15% in 91% of patients, and there were two deaths.71 Less cumbersome and 5.9 to 8.1 months, respectively.78,84–86 A large randomized trial to administer and less toxic schedules are commonly used in prac- (486 patients) compared gefitinib (250 or 500 mg daily) to meth- tice (e.g., 135 to 225 mg/m2 intravenously over 3 hours every 3 otrexate and demonstrated no survival improvement with either weeks; 80 to 100 mg/m2 weekly), although their relative efficacies gefitinib dose.87 have not been well evaluated. A paclitaxel schedule that provides With both of these classes of agents, the development of rash more prolonged exposure to the drug may be more efficacious,72 was associated with clinical benefit, but this association is not fully although a phase II trial of 120 to 140 mg/m2 over 96 hours yielded explained by simple pharmacokinetics.78 There is no established disappointing results even in treatment-naïve patients (major re- molecular predictor of response to these agents currently in head sponse rate, 13%).65 Other toxicities besides myelosuppression and neck SCC. include sensory neuropathy, alopecia, allergic reactions, and ar- The successful development and approval of cetuximab in rhythmia, although cardiac monitoring is not required. head and neck SCC highlights the potential for therapies to ex- Docetaxel appears less neuropathic than paclitaxel, but fluid ploit specific molecular pathways with therapeutic effect. A num- retention and hematologic toxicity may be more problematic. A ber of other new agents, often with multitarget capability, are typical dose is 60 to 100 mg/m2 intravenously over 1 hour. Initial entering clinical trials. There is a good rationale for agents that studies evaluated the efficacy of the 100 mg/m2 dose level, with target angiogenesis in head and neck SCC.88 However, the de- major response rates ranging from 21% to 42%73; an excellent velopment of bevacizumab has been cautious given the reported performance status is required for this higher dose. Lower doses toxicity concerns, specifically bleeding, in patients with squamous may offer similar efficacy and better tolerance.74 As with paclitaxel, cell lung cancer.89 Alterations in the PI3K/Akt/mammalian target weekly schedules are applied in practice, but the relative efficacy of rapamycin (mTOR) pathway are common in head and neck of weekly versus a schedule of every 3 weeks is not well-studied. cancer, and activation appears independent of EGFR activation, Although initial studies evaluated a bolus schedule for 5-fluo- making targeting of this pathway of great interest.90 Cancer gene rouracil, an infusional program of 1,000 mg/m2 per day over 96 therapy, whereby genetic sequences are introduced via viral or to 120 hours appears more efficacious in head and neck cancer.75 nonviral vectors, is well-suited to head and neck tumors given the Infusional 5-fluorouracil is associated with more mucositis and local–regional character of head and neck tumors that facilitates diarrhea than a bolus schedule, so the shorter infusion (i.e., 96 direct injection and the monitoring of gene expression. The tumor hours) is typically applied in patients who are pretreated and have suppressor gene p53 has been one target, because somatic mu- received prior head and neck RT. tations of it are common in head and neck cancers, particularly EGFR is highly expressed in most head and neck SCCs, and among patients who have smoked cigarettes and used alcohol.91 the degree of expression is inversely associated with prognosis.76–78 For example, in a phase II study of Onyx-015, a replication-com- As such, there has been a keen interest in drugs that target the petent adenovirus absent the E1B gene, major responses, includ- receptor itself or steps downstream. Cetuximab, a chimeric immu- ing some complete regressions, occurred in 10% to 14% of treated noglobulin G antibody that binds the receptor, has been approved patients.92 In other studies, treatment with Onyx-015, or a similar by the U.S. Food and Drug Administration for use in patients virus such as H101, improved the efficacy of chemotherapy.93,94 with disease refractory to platin-based therapy. As summarized in Table 38.7, the response rates in this refractory setting are similar Combination Therapy (10% to 13%) whether cetuximab is used alone or combined with platin-based therapy; median survivals remained disappointing, Given the disappointing track record for single-agent therapy in ranging from 5.2 to 6.1 months.79–82 Another EGFR antibody, za- the palliative setting, combinations of drugs have been extensively lutumumab, was compared in a randomized trial to best supportive evaluated. In the early 1980s, investigators from Wayne State, build- care alone in patients with cisplatin-refractory squamous cell head ing upon potential synergy between cisplatin and 5-fluororuacil, and neck cancer. Among 286 entered patients, there was no sig- reported a major response rate of 70% with a complete response nificant improvement in the primary endpoint of overall survival rate of 27% using a regimen of cisplatin 100 mg/m2 intravenously (median 6.7 versus 5.2 months, p = 0.0648) although a significant and a 5-fluorouracil 1,000 mg/m2 per day continuous infusion over difference was found on an exploratory post hoc analysis done 12 96 hours recycled every 3 weeks in patients with recurrent or dis- months after the last patient was randomized. Response rate (6.3% seminated disease.95 Other investigators confirmed the significant versus 1.1%) and progression-free survival (median 9.9 versus 8.4 activity of the regimen, albeit with a somewhat lower major and weeks, p = 0.0012) were improved with zalutumumab.83 complete response rate on average (50% and 16%, respectively),96 The small molecule tyrosine–kinase inhibitors, gefitinib and establishing it as the standard regimen to which new therapies are erlotinib, offer no efficacy advantage in similar refractory patients. compared.

TABLE 38.7 Cetuximab for Recurrent or Metastatic Head and Neck Cancer: Selected Studies

No. of Median PFS Median OS Author Patients Cancer Chemotherapy RR (Months) (Months) Herbst et al.80,a 79 SCC–POD on CDDP based CDDP based + cetuximab 6%–20% 2.0–3.0 4.3–6.1 Baselga et al.79,a 96 SCC–POD on platin based CDDP based + cetuximab 10%–11% 2.4–2.8 4.9–6.0 Trigo et al.81 103 SCC–POD on platin based Cetuximab 13% 2.3 5.9 Burtness et al.82,b 117 SCC CDDP 10% 2.7 8.0 No chemo for R/M CDDP + cetuximab 26% 4.2 9.2 a Range related to how POD was defined in different subgroups. b Response rates were significantly different (p = 0.03): PFS (p = 0.09) and OS (p = 0.21) did not reach statistical significance. RR, response rate; PFS, progressive-free survival; OS, overall survival; SCC, squamous cell cancer; POD, progression of disease; CDDP = cisplatin; R/M = recurrent or metastatic disease.

tahir99 - UnitedVRG Chapter 38 Cancer of the Head and Neck 429

Despite an improvement in response rates associated with a randomized trial.109 Phase II studies of a taxane with cisplatin or the use of combination therapies like cisplatin and 5-flurouracil, carboplatin and a third drug (e.g., 5-fluorouracil, ifosfamide), have demonstrating a statistically or clinically significant improvement yielded major response rates of 55% to 86%110–113; whether these in survival compared to single-agent therapy has proven elusive. regimens translate into better survival outcomes compared to a cis- Table 38.8 summarizes the results of three randomized trials that platin-based doublet that may be less toxic await further evaluations. compared treatment with cisplatin and infusion 5-fluorouracil to There is great interest in the combination of standard chemo- that with different single agents.97–99 Treatment with combination therapy with newer targeted agents. One ECOG study compared chemotherapy led to a significant increase in response rate, albeit cisplatin versus cisplatin and cetuximab as first-line treatment in at the cost of greater toxicity. Overall survival did not significantly 123 patients. The arm including the cetuximab had a significantly improve. The meta-analysis reported by Browman et al.100 yielded higher response rate (10% versus 26%, p = 0.03), but no signifi- similar conclusions. These data do not support the routine use of cant difference was found in the primary endpoint of progression- cisplatin-based combinations for patients with recurrent or meta- free survival (2.7 versus 4.2 months, p = 0.09) or in overall survival static SCC. Combination therapy seems most appropriate for pa- (8.0 versus 9.2 months, p = 0.21), although the trends favored the tients with a good performance who have significant symptoms combination arm.82 In a larger trial (EXTREME),114 442 patients (e.g., pain) for which the higher anticipated response rate will were randomized to cisplatin or carboplatin and 5-fluorouracil translate into better palliation. with or without cetuximab for six cycles. Subsequent maintenance The activity of paclitaxel and docetaxel in head and neck can- with cetuximab alone was allowed on the investigational arm, but cer has fostered the development and evaluation of taxane and cis- there was no crossover to cetuximab on the standard arm. Both platin combinations. Docetaxel with cisplatin is associated with a median progression-free (5.6 versus 3.3 months) and overall (10.1 major response rate of 40% to 53%, with complete response rates versus 7.4 months) survivals were significantly improved on the approximating 6% to 18%101,102; a weekly schedule of paclitaxel triplet arm, at the cost of more sepsis (nine patients versus one (80 mg/m2) and carboplatin (area under concentration [AUC] ver- patient; p = 0.02), grade 3 skin reactions (9%), and grade ≥3 infu- sus time curve, 2) appeared more efficacious than an every 3-week sion reactions. Quality of life outcomes were reported to not be sig- dosing (paclitaxel 175 to 200 mg/m2 intravenously over 3 hours nificantly different between the treatment arms.115 Tumor EGFR followed by carboplatin AUC 6) in two separate phase II stud- copy number and degree of EFGR expression were not predictive ies.103–105 ECOG compared a high-dose (200 mg/m2) and mod- of benefit with cetuximab.116,117 Of interest, data from Vermorken erate-dose (135 mg/m2) paclitaxel, both by 24-hour infusion and and colleagues118 suggests that the therapeutic effect of cetuximab, followed by the same dose of cisplatin (75 mg/m2) in a random- when combined with chemotherapy, is mainly additive rather ized study (E1393). No significant difference in response rate or than synergistic. Whether allowing patients to crossover to cetux- survival was found between the arms.106 Another randomized trial imab on the doublet arm at progression would have decreased or done under the auspices of ECOG compared standard cisplatin eliminated the observed survival difference is of interest for future 2 and 5-fluorouracil with paclitaxel 175 mg/m intravenously over research. A similar but not identically designed randomized trial OF ONCOLOGY PRACTICE 3 hours and cisplatin 75 mg/m2 (E1395).107 Objective major re- evaluated cisplatin and 5-flourouracil with or without the EGFR sponse rates (27% versus 26%) and median survivals (8.7 versus 8.1 antibody panitumumab in 657 patients. Response rate (36% versus months) were no different between the arms. The reported quality 25%, p = 0.0065) and median progression-free survival (5.8 versus of life was better on the paclitaxel arm over the first 16 weeks of 4.6 months, p = 0.0036) were significantly improved with the in- treatment.108 corporation of the panitumumab, but not the primary endpoint of Attempts have been made to improve the efficacy of combina- overall survival (11.1 versus 9.0 months, p = 0.1403). Overall sur- tion chemotherapy through the development of a variety of triplets. vival was improved in the p16-negative subgroup (p = 0.0115).119 The addition of interferon-alpha2b (IFN-α2b) to cisplatin and With advances in RT techniques that facilitate reirradiation 5-fluorouracil failed to significantly improve response or survival in with acceptable morbidity, this approach has been increasingly

TABLE 38.8 Chemotherapy for Recurrent or Metastatic Head and Neck Cancer: Selected Phase III Trials of Cisplatin/5-Fluorouracil Versus Other Options

Study No. of Patients Agents Response Ratesa Median Survivalsb (Months) Jacobs et al.98 249 CDDP/FU 32% 5.5 CDDP 17% 5.0 FU 13% 6.1 Forastiere et al.97 277 CDDP/FU 32% 6.6 CBDCA/FU 21% 5.0 MTX 10% 5.6 Clavel et al.99 382 CDDP/MTX/BLEO/VCR 34% 8.2 CDDP/FU 31% 6.2 CDDP 15% 5.3 Schrijvers et al.109 122 CDDP/FU/IFN-α2b 38% 6.0 CDDP/FU 47% 6.3 Gibson et al.107 218 CDDP/FU 27% 8.7 CDDP/PAC 26% 8.1 a The following response rate differences were statistically significant at p< 0.05: Jacobs et al., CDDP/FU versus both CDDP and FU; Forastiere et al., CDDP/FU versus MTX; Clavel et al., both combinations versus CDDP. b All survival differences were not statistically significant. CDDP, cisplatin; FU, 5-fluorouracil; MTX, methotrexate; CBDCA, carboplatin; BLEO, bleomycin; VCR, vincristine; IFN-α2b, interferon-alpha 2b; PAC, paclitaxel. 430 Practice of Oncology / Cancer of the Head and Neck

explored in patients with unresectable local or regional recur- The advantages of postoperative compared with preoperative rence, often with integrated chemotherapy. The observed median RT include less operative morbidity, more meaningful margin survivals in these series are similar to those obtained in phase II checks at the time of the surgery, a knowledge of tumor spread trials of chemotherapy alone, but more durable responses occur for RT planning, safe use of a higher RT dose, and no chance the in selected patients and there is a clearer plateau on the survival patient will refuse surgery. The disadvantages of postoperative RT curve. In two larger series involving 169 and 115 patients, respec- include the larger treatment volume necessary to cover surgical tively, among patients treated with a variety of RT fractionation dissections, a delay in the start of RT with possible progression, and schedules and concurrent chemotherapy regimens, 2-year survival the higher dose required to accomplish the same rates of local– rates exceeded 20%.120,121 In two sequential RTOG studies, a regi- regional control. men of daily paclitaxel (20 mg/m2) and cisplatin (15 mg/m2) added concurrently to split-course RT (total dose 60 Gy, 1.5 Gy twice- Preoperative Radiation Therapy daily fractions; granulocyte colony-stimulating factor support dur- Preoperative RT should be considered for the following situations: ing off weeks) (RTOG 96-11) yielded a better 2-year survival rate (1) fixed-neck nodes, (2) delayed initiation of postoperative RT by than concurrent 5-fluororuacil and hydroxyurea added to the same >8 weeks, (3) use of the gastric pull-up for reconstruction, and (4) RT schedule (RTOG 96-10) (24.9% versus 16.9%, p = 0.44).122,123 open biopsy of a positive neck node. The reported 2-year survival rates exceed the rate of 10.5%, which was observed in a subgroup of 124 patients with local disease only Postoperative Radiation Therapy who had previously received RT and participated in E1393 or E1395.124 Randomized trials comparing chemotherapy alone to Postoperative RT is considered when the risk of recurrence above reirradiation and chemotherapy are needed. the clavicles exceeds 20%. The operative procedure should be one stage and of such magnitude that RT is started no later than 6 to 8 Nasopharynx Cancer weeks after surgery. The operation should be undertaken only if it Many of the same drugs and regimens used in the treatment of is believed to be highly likely that all gross disease will be removed head and neck SCC are also active in NPC. There are reports of and margins will be negative. a small proportion of patients with recurrent or metastatic disease Although no definitive randomized trials have addressed the being controlled long term with chemotherapy alone.125 Available efficacy of postoperative RT in the treatment of head and neck data support the use of cisplatin-based combination chemotherapy cancer, excellent data that has bearing on this issue is available (e.g., cisplatin/5-fluorouracil; cisplatin/bleomycin/5-flurouracil from the Medical College of Virginia. Two groups of surgeons +/− epirubicin), although there is a lack of randomized studies operated on patients with head and neck cancer: general sur- to clarify the relative efficacies and toxicities of different options. gical oncologists who used surgery alone and reserved RT for Site-specific phase II studies report major response rates of 70% treatment of recurrent disease, and otolaryngologists who rou- 126–128 tinely sent patients with locally advanced disease for postopera- or higher with regimens containing cisplatin. In a review 143 of the Princess Margaret Hospital experience, single-agent or tive RT. Of 441 patients, 125 were treated surgically between noncisplatin-based combination chemotherapy was associated 1982 and 1988 and had ECE and/or positive margins, 71 were with a major and complete response rates of 25% and 8%, respec- treated with surgery alone, and 54 received postoperative RT. Local control rates at 3 years after surgery alone compared with tively, in 40 patients, whereas cisplatin-based combination therapy = produced major and complete response rates of 70% and 23%, surgery and RT were: for ECE, 31% and 66% (p 0.03); positive 126 margins, 41% and 49% (p = 0.04); and ECE and positive mar- respectively, in 30 patients. The substitution of carboplatin may = be associated with less activity.127 gins, 0% and 68% (p 0.001). A multivariate analysis of local 2 control revealed that the use of postoperative RT (p = 0.0001), With regard to other agents, paclitaxel as a 175 mg/m 3-hour = = infusion is active with a response rate of 22% in a series of 24 pa- macroscopic ECE (p 0.0001), and margin status (p 0.09) tients with undifferentiated NPCs.128 The combination of it with were of independent significance. Cause-specific survival rates at 3 years were 41% for surgery alone and 72% for surgery and carboplatin has yielded response rates consistently greater than = 50%.129–131 Gemcitabine is active in NPCs,132,133 and combina- postoperative RT (p 0.0003). A multivariate analysis of cause- specific survival showed that postoperative RT (p = 0.0001) tions including it appear promising, with response rates exceeding = 70%.134,135 Capecitabine, prolonged 5-fluorouracil infusion, and and the number of nodes with ECE (p 0.0001) significantly influenced this endpoint. cetuximab all have modest activity in the refractory setting, and 144 no major responses were seen in one study with gefitinib.136–139 In another series, Lundahl et al. reported on 95 patients with There is keen interest in looking to exploit the association NPCs node-positive SCC who were treated with a neck dissection and have with EBV for therapeutic purposes. Potential gene therapy postoperative RT at the Mayo Clinic. A matched-pair analysis was approaches are discussed elsewhere.140,141 performed using a series of patients treated with surgery alone; 56 matched pairs of patients were identified. The recurrence rates in the dissected neck (relative risk [RR] = 5.82, p = 0.0002), recur- GENERAL PRINCIPLES OF COMBINING rence in either side of the neck (RR = 2.21, p = 0.0052), and = = MODALITIES death from any cause (RR 1.67, p 0.0182) were significantly higher for patients treated with surgery alone. Thus, it appears that postoperative RT may significantly im- Surgery Plus Radiation Therapy prove both local–regional disease control and survival for patients who are at high risk for failure after surgery. RT may be administered preoperatively or postoperatively. An Indications for postoperative RT include close (<5 mm) or analysis of available data suggests there is no compelling difference positive margins, ECE, multiple positive nodes, invasion of the in survival rates comparing the two sequences42; local–regional soft tissues of the neck, endothelial-lined space invasion, PNI, and control may be improved with postoperative treatment.142 more than 5 mm of subglottic invasion.42 The authors currently Combined modality therapy should be avoided for lesions recommend 60 Gy in 6 weeks to 66 Gy in 6.5 weeks for patients with a high cure rate (70% or greater) by either surgery or RT with negative margins and fewer than three indications for RT. For alone. The increased morbidity from combined treatment is not patients with close (<5 mm) or positive margins, we recommend associated with a significantly improved control rate, and many 70 Gy in 7 weeks or 74.4 Gy at 1.2 Gy twice a day. Concomitant patients with local or regional failure can be salvaged by secondary cisplatin chemotherapy should be considered for patients with procedures. positive margins and/or ECE.43–45

tahir99 - UnitedVRG Chapter 38 Cancer of the Head and Neck 431

Given the appreciation that HPV-related cancers have a better survival benefit when induction chemotherapy was combined prognosis than HPV-unrelated disease, there is interest in potential with surgery or RT is understandable. Yet, in the original report of deescalation of therapy with the intent of decreasing toxicity with- the MACH-NC analysis, which included 31 induction studies, all out compromise in survival.145 Although this is an important area but 2 suggested no survival benefit.149 of future research, modification in standards of care based on the However, a more careful look at these and other data do pro- HPV status of the tumor is not recommended outside of a clinical vide grounds for continued interest in this approach. Many of trial at present.146 the included studies had significant methodologic limitations by more contemporary trial standards. A subset analysis, limited to the 15 trials that used cisplatin and infusional 5-fluorouracil, sug- gested survival benefit (hazard ratio [HR], 0.88; 95% confidence CHEMOTHERAPY AS PART OF 149 CURATIVE TREATMENT interval [CI], 0.79 to 0.97). Even in the absence of survival im- provement, there seemed to be a correlation between response to chemotherapy and subsequent response to RT, which provided a Systematically designed and randomized studies have established basis for subsequent organ preservation initiatives.151,152 Finally, a role for drug therapy as part of the standard combined modality patterns of failure were affected with less distant metastases in management of head and neck SCC in several settings. These in- certain studies when induction chemotherapy was incorporated. clude the therapy of unresectable disease, for organ preservation, As local–regional control improves, the rate of clinically apparent and for patients with poor risk pathologic features after surgery. distant metastases is increasing,153 and induction chemotherapy is, Chemotherapy has been shown to improve the likelihood of dis- on average, better tolerated than maintenance therapy as a way to ease control compared to RT alone in patients with advanced dis- give additional systemic therapy. ease, albeit with increased acute toxicity. In certain circumstances, The study reported originally by Paccagnella et al.154 is illustra- response to chemotherapy has been used to triage patients to dif- tive of these types of trials, and provides further insights.155 Two ferent local–regional treatments. hundred and thirty-seven patients with stage III or IV head and Chemotherapy has been integrated with surgery or RT in a neck cancer were randomized to four cycles of induction cis- variety of ways including induction, concurrent with RT, and/ platin and infusional 5-fluorouracil followed by standard local– or maintenance. Unlike outcome studies of surgery and/or RT regional treatment (i.e., surgery plus RT if resectable, RT alone in which site-specific results are reported, albeit typically using a if unresectable). Resectability was assessed pretreatment, not after retrospective methodology, many of the trials evaluating the role chemotherapy, and was a stratification criteria. Overall, there was of chemotherapy enrolled patients with a variety of head and neck no significant difference between the arms with regard to overall SCCs. Even when site specific, although prospective, subsites are survival or local–regional control, although the incidence of distant combined. This is less of an issue for studies evaluating therapy metastases was lower among patients treated with chemotherapy. for NPC.147,148 Nonetheless, important lessons have been learned

On a subset analysis, however, patients with unresectable disease OF ONCOLOGY PRACTICE from these studies, further enhanced by use of a random assign- benefitted from the incorporation of induction chemotherapy for ment methodology. In this section, general principles for the in- all outcomes, including local–regional control, distant control, tegration chemotherapy with local–regional treatment will be and overall survival (3-year survival 24% versus 10%, p = 0.04). discussed with a focus on the results of randomized trials. Among resectable patients, improvement in distant control was The Meta-Analysis of Chemotherapy on Head and Neck Can- offset by a decrement in local–regional control with the integra- cer (MACH-NC) included 63 randomized trials published from tion of induction chemotherapy, and reported survival rates in this 1965 through 1993, all of which compared local–regional treat- subgroup were similar on both treatment arms. ment with or without chemotherapy.149 Individual patient data Historically, then, there was no established role for induction was available on 10,741 patients. The absolute improvement in chemotherapy prior to planned surgery and postoperative RT, and 5-year survival overall was 4% (p <0.001). However, the signifi- a limited role only in selected settings prior to RT. However, with cant improvement appeared limited to those patients who received the incorporation of taxanes into induction regimens containing concomitant treatment (absolute difference of 8% at 5 years, p cisplatin and 5-fluororuacil, newer data suggest that the indica- <0.001). Neither the difference seen at 5 years with induction tions for induction chemotherapy may further evolve. (2%, p = 0.10), nor maintenance (1%, p = 0.74) chemotherapy Three randomized trials have compared the relative efficacies was statistically significant.149 In an update of this analysis, now of induction chemotherapy with standard cisplatin and 5-fluo- including trials through 2000 and totaling 17,346 patients,149 the rouracil versus a triplet including a taxane and these same two superior efficacy of concurrent therapy was confirmed, and was drugs with one or both being dose adjusted.156–158 All three stud- greater than that seen with induction chemotherapy. Survival ben- ies randomized patients with advanced M0 head and neck can- efit diminished with patient age and, on subset analysis, was not cer to either cisplatin and 5-fluororuacil or a triplet, followed by significant in patients over 70 years of age. the same RT-based treatment. In one study, this was RT alone, Tumor HPV status has emerged as an important predictor whereas, in the other two, concurrent therapy with carboplatin of favorable treatment response and survival, particularly for pa- and cisplatin, respectively, were employed. In general, the taxane- tients with oropharynx cancer.7 In ECOG 2399, the response containing triplet was associated with a higher response rate to to chemotherapy to all protocol treatment, progression free-sur- induction chemotherapy, and improved both progression-free vival, and overall survival were all improved in the HPV-positive and overall survival. More neutropenia was observed with triplet group.7 Subsequent analysis of RTOG 0129 demonstrated that therapy but, overall, it was as well-tolerated as standard cisplatin tobacco use (>10 pack-years) and the extent of nodal disease and 5- fluorouracil. (N2b-N3) both adversely affect the prognosis associated with HPV- These studies were designed to determine which induction positive tumors.8 chemotherapy was more efficacious, and provide convincing evi- dence that the triplet of a taxane with cisplatin and 5-flurouracil is Induction Chemotherapy superior to standard cisplatin and 5-fluorouracil alone as induction therapy. However, an alternative design is necessary to define the In untreated patients with local or regionally advanced M0 head role of induction with such triplets in standard practice. For this and neck SCC, treatment with cisplatin-based combination che- population, as discussed in the next section, concurrent chemo- motherapy will yield major response rates approximating 90%, therapy and RT alone without induction chemotherapy is the more with clinical complete response rates in the 30% range.150 Enthu- established standard therapy. Randomized studies are necessary to siasm that response rates of this magnitude should translate into determine whether a sequential approach using induction with a 432 Practice of Oncology / Cancer of the Head and Neck triplet followed by RT-based treatment (typically with concurrent function preservation and overall survival between the arms; more chemotherapy), is superior to concurrent chemotherapy and RT local failures occurred on the cetuximab arm.163 alone such that the added duration of treatment and potential tox- icity is justified. To date, available randomized trials have failed to demon- Concurrent Chemotherapy and Radiation for strate a clear overall survival benefit with the incorporation of Gross Disease induction chemotherapy. The combination of docetaxel, cis- platin, and 5-fluorouracil has been the focus of these investiga- Although there are a number of ways to integrate chemotherapy tions. One study available only in abstract form was confounded with RT, available data most strongly support concurrent chemo- by the lack of an intention to treat an analysis with unequal ex- therapy. Given proven efficacy in patients with poor prognostic clusions among treatment arms. Even with those methodologic and unresectable disease, more recent investigations have applied limitations, it failed to demonstrate a significant improvement in the approach in better prognostic, organ preservation, and adju- overall survival with the incorporation of induction docetaxel, vant settings. cisplatin, and 5-fluorouracil or cisplatin, and 5-fluorouracil.159 Concurrent chemoRT programs vary in many ways, of which In the PARADIGM study, 145 patients with local or region- the type of chemotherapy (i.e., specific agents, single, combina- ally advanced SCC were randomized to induction docetaxel, tion) and RT schedule (i.e., dose, fractionation) are the most cisplatin, and 5-fluorouracil followed by carboplatin or docetaxel apparent variables. In general, three main approaches can be dis- concurrent with RT versus concurrent cisplatin with concomi- cerned: single-agent or combination chemotherapy with continu- tant boost radiation. Patients could have unresectable disease or ous-course RT; combination chemotherapy with split-course RT, be resectable, with the intent of therapy being organ preserva- often with altered fractionation; and chemotherapy alternating tion. The study was closed early because of slower than expected with RT.164 Although continuous course RT may be desirable and accrual, so it was somewhat underpowered. There was no dif- more attractive from a radiobiologic perspective, local toxicities ference in overall or progression-free survival between the arms may preclude it depending on the concurrent agents used. The with a median follow-up of 49 months; the 3-year overall survival first two approaches are the most common. rates were 73% on the induction arm and 78% on the concur- A variety of drugs and combinations have been utilized con- rent arm (p = 0.77); and the 3-year progression-free survival rates currently with RT. When only one drug is used, the MACH-NC were 67% and 69%, respectively (p = 0.82). A subset analysis of indicates that the impact is largest with a platin, of which cisplatin the group with advanced neck disease (N2b/N2c,N3), felt to be is the predominant one studied, a conclusion shared in another at increased risk of distant metastases, demonstrated no advan- meta-analysis reported by Browman and colleagues.165 Of interest, tage with the incorporation of induction chemotherapy.160 The platin plus 5-fluorouracil (HR, 0.75) offered no clear advantage DECIDE trial used a similar design, but only patients with N2/ compared to platin alone (HR, 0.74).166 The results of a three- N3 were eligible and, for the concurrent therapy, hydroxyurea arm randomized study comparing concurrent cisplatin and RT, and 5-fluorouracil was used. Among 280 patients accrued with concurrent cisplatin, 5-fluorouracil and split-course RT (with pos- minimum 24-months follow-up, there was no significant differ- sible resection depending on response), and definitive RT alone in ence between the sequential and concurrent arms with regard patients with unresectable disease reported by Adelstein et al.,167 to overall survival (75% versus 73%, p = 0.70) or disease-free in the E1392 study, are consistent with this assessment. Although survival (69% versus 64%, p = 0.39); the cumulative incidence daily,168 weekly,46,169 and every 3 week schedules of cisplatin intra- of distant failure, however, was lower in the induction arm (10% venously concurrent with RT have been applied, the last schedule versus 19%, p = 0.025).161 Finally, in 256 enrolled patients with is the one most studied and is a widely accepted standard. If weekly stage III or IVA oral cavity cancer, induction with docetaxel, cis- dosing is used, 20 mg/m2 weekly appears too low because it did platin, and 5-fluorouracil prior to surgery and postoperative radia- not significantly improve overall survival or failure-free survival tion failed to improve overall survival (p = 0.918) or disease-free in one randomized study.169 Attempts to improve the efficacy of survival (p = 0.897) compared to proceeding directly to surgery concurrent cisplatin through intra-arterial administration170,171 did and postoperative radiation alone.162 not prove more efficacious in a randomized trial when compared The optimal role of induction chemotherapy is currently to intravenously delivered cisplatin, although toxicity profiles controversial. A review of the National Comprehensive Can- differed.172 In absence of a proven efficacy advantage with intra- cer Network (NCCN) guidelines highlights this reality, because arterial delivery, intravenous cisplatin is preferred because it is lo- concurrent chemoRT alone and induction followed by RT-based gistically easier to administer. therapy are both listed as treatment options for certain disease sce- Most randomized trials to date have compared chemoRT to RT narios.146 Although concurrent chemoRT alone remains the stan- alone. As such, studies evaluating the efficacy of different chemoRT dard to which new treatments are compared for local or regionally programs are limited. For example, for purposes of the MACH-NC advanced disease and generally receives the higher category rating analysis, “platin” included both cisplatin and carboplatin. Yet the in these guidelines, induction is well-suited for certain settings in relative efficacy of these agents, when given concurrently, is not patients who are medically fit. Examples include when immediate well-studied. In one three-arm randomized study by Jeremic et al.168 therapy is needed in the hope of avoiding a tracheostomy or PEG, using a daily schedule for each drug with RT and a control arm of in organ preservation settings where the degree of response affects RT alone, both the cisplatin (6 mg/m2 per day) and carboplatin the decision to proceed with surgery versus RT-based therapy, or (25 mg/m2 per day) arms appeared comparable, and superior in in patients with advanced neck disease at higher risk for distant efficacy to RT alone. However, in a randomized study reported by metastases. the Hellenic Cooperative Oncology Group using an every 3 week After induction chemotherapy, there is some controversy as schedule for each drug (cisplatin 100 mg/m2, carboplatin AUC, 6), to whether to proceed with RT alone or concurrent chemoRT, their equivalence seemed less clear.173 The RTOG reported a ran- and if the latter, which drug to use.156–158 Concern exists regarding domized phase II study comparing three different chemotherapy the tolerance to high-dose cisplatin after cisplatin-based induction regimens, all delivered concurrently with 70 Gy in 2 Gy fractions: treatment. In a randomized phase II trial, concurrent cetuximab ARM 1, cisplatin 10 mg/m2 per day and 5-fluorouracil 400 mg/m2 and RT was compared with high-dose cisplatin and RT after per day continuous infusion for the final 10 days of treatment; ARM induction docetaxel, cisplatin, and 5-fluorouracil in 116 patients 2, hydroxyurea 1 g every 12 hours and 5-fluorouracil 800 mg/m2 per with advanced hypopharynx or larynx cancer. Toxicity was sub- day continuous infusion every other week; or ARM 3, weekly pa- stantial on both arms, but treatment compliance was better with clitaxel 30 mg/m2 and cisplatin 20 mg/m2. Among 231 analyzable cetuximab therapy. There was no significant difference in larynx patients, 2-year disease-free and overall survival rates were for ARM

tahir99 - UnitedVRG Chapter 38 Cancer of the Head and Neck 433

1, 38.2% and 57.4%; for ARM 2, 48.6% and 69.4%; and for ARM 3, median follow-up was 54 months. The median duration of sur- 51.3% and 66.6%, respectively.174 vival was 49 months after combined therapy compared with 29 Because anemia may adversely affect the efficacy of RT, the in- months after RT alone (p = 0.03). Other than an acneiform rash tegration of an appropriate hematopoietic growth factor has been and infusion reactions, grade 3 or greater complications were simi- investigated. In a multicenter, double-blind, randomized, placebo- lar in the two groups of patients. The results of this trial have been controlled trial, the addition of erythropoietin 300 IU/kg three times confirmed with longer follow-ups.186 Patients with oropharynx weekly during postoperative RT was evaluated in 351 patients with cancer appeared to derive the largest benefit with the integration head and neck SCC.175 Although target hemoglobulin levels were of cetuximab, suggesting that HPV-related disease may have an reached in 82% of patients receiving erythropoietin compared to improved outcome with the use of this agent; other studies, how- 15% receiving placebo, local–regional progression-free survival ever, suggest greater activity of EGFR-directed antibody therapy (adjusted RR, 1.62 [95% CI, 1.22 to 2.14]; p = 0.0008), local– in HPV-negative disease.114,119 Randomized data comparing cetux- regional progression (RR, 1.69 [CI, 1.16 to 2.47]; p = 0.007), and imab and RT to other chemoRT programs are not available. One survival (RR, 1.39 [CI, 1.05 to 1.84]; p = 0.02) were all inferior RTOG study evaluating concurrent cisplatin and RT versus ce- on the erythropoietin arm. Consistent with the current FDA alert, tuximab and RT in patients with HPV-related cancers is currently an erythropoietin-stimulation agent is contraindicated during in progress. Investigators at Memorial-Sloan Kettering reported a curative-intent RT-based therapy.176 Transfusion then, is the pre- phase II trial of cisplatin and cetuximab concurrent with RT in ferred approach to address potential radiation resistance attributed patients with local–regionally advanced head and neck SCC. Ef- to anemia, but a recent analysis of two randomized trials failed ficacy was impressive, although there were toxicity concerns.187 to demonstrate that prophylactic transfusion improved overall sur- The RTOG completed accrual to a large randomized trial (RTOG vival or other disease control endpoints.177 Use of a hypoxic radio- 0522, n = 895 evaluable patients) intended to assess the efficacy sensitizer represents another strategy to potentially address tumor and safety of this regimen compared to concurrent cisplatin and hypoxia. The results of one meta-analysis were consistent with RT. With a median follow-up of 2.4 years among surviving pa- potential benefit178; recent randomized trials, however, have not tients, the incorporation of cetuximab failed to significantly im- been convincing in terms of improved disease control with such prove 2-year progression-free survival (63% versus 64%, p = 0.66) a strategy.179,180 or overall survival (83% versus 80%, p = 0.17), but mucosal and An important question is whether the use of newer, more skin toxicity were increased.188 efficacious, and altered fractionated RT programs181 obviates the Choosing among the numerous concurrent programs can be benefits accrued with the addition of chemotherapy. A single difficult. In the NCCN guidelines,146 concurrent cisplatin with RT institution study reported by Brizel et al.182 compared a more is the preferred choice, although several other options are listed. aggressive RT schedule with or without concomitant 5-fluorouracil It is important to emphasize that concurrent chemoRT may be and cisplatin. Patients who underwent RT alone received 75 Gy associated with significant toxicity; treatment-related mortality, al-

in 60 twice-daily fractions; those who underwent concomitant beit infrequent (<5% in the cooperative group setting), may occur. OF ONCOLOGY PRACTICE chemotherapy received 70 Gy in 56 twice-daily fractions with a Morbidity from chemotherapy (dependent on the agent chosen) 7-day split. Chemotherapy consisted of two cycles of concomitant and RT are possible, and there are both acute (e.g., mucositis, cisplatin 12 mg/m2 per day and 5-fluorouracil 600 mg/m2 per day blood count suppression) and chronic (e.g., dry mouth, swallow- each for 5 days, followed by two cycles of maintenance chemother- ing dysfunction, fibrosis) toxicities. Selected studies have begun apy. Among the 116 patients included, chemoRT was associated to report long-term, not just acute, toxicities.189 Appropriate infra- with improved 3-year rates of local–regional control (70% versus structure, an experienced multidisciplinary team, and a coopera- 40%, p = 0.01), relapse-free survival (61% versus 41%, p = 0.08), tive patient are necessary to optimize both efficacy and safety. and overall survival (55% versus 34%, p = 0.07). In another randomized study reported by Jeremic et al.,183 the addition of daily cisplatin to hyperfractionated RT also leads to incremental Nasopharynx Cancer benefits. The results of these studies are consistent with the MACH-NC analysis, which demonstrated significant HRs, that are Current practice has been particularly affected by the Intergroup consistent with benefits among patients receiving postoperative RT Study 0099 (Table 38.9).190 In it, 147 patients with stage III or IV (HR, 0.79), conventional RT (HR, 0.83), or altered fractionated NPC were randomized to definitive RT (70 Gy, 35 fractions over RT (HR, 0.73), suggesting a benefit for adding concomitant che- 7 weeks) versus cisplatin 100 mg/m2 intravenously on days 1, 22, motherapy regardless of the type of RT schedule.166 and 43 concurrent with the same dose of RT followed by three Of note, the converse—once concurrent chemotherapy is planned cycles of cisplatin and infusional 5-fluorouracil. Although added, does an altered fractionation RT schedule further improve only 63% and 53% of patients received all the planned concurrent outcome compared to that seen with standard fractionation—has and maintenance treatments, respectively, local–regional control, not been established in randomized trials. Neither RTOG 0129 distant control, progression-free, and overall survivals were all sig- (standard versus concomitant boost RT both with concurrent nificantly improved with chemoRT. high-dose cisplatin)8 nor GORTEC 99-02 (accelerated RT with One of the potential limitations of the Intergroup Study was or without concurrent carboplatin and 5-flourouracil, standard how generalizable its results would be to endemic NPCs, because fractionated RT with concurrent carboplatin and 5-flourouracil)184 24% of patients entered in the trial had World Health Organiza- demonstrated improved overall survival with the incorporation tion (WHO) type I histology. However, subsequent reports of ran- of altered fractionated RT with concurrent chemotherapy versus domized trials in which WHO types II and III predominated have standard fractionation with concurrent chemotherapy to justify the similarly shown a survival advantage with concurrent cisplatin- added logistical complexity and potential added toxicity. based concurrent chemotherapy without191–193 or with mainte- For patients who are not cisplatin candidates, using a carbo- nance chemotherapy.194,195 platin-based program (e.g., carboplatin/5-fluorouracil)147 or other One relatively small randomized trial (n = 206) with a median concurrent programs that have different side effect profiles and follow-up of 26.3 months, suggested that weekly carboplatin dosed that withstood the scrutiny of a randomized trial is recommended. at 100 mg/m2 concurrent with RT when compared to standard There has been great interest in cetuximab and concurrent RT in high-dose cisplatin, did not yield inferior disease-free or overall this regard.185 In a randomized study reported by Bonner et al.,186 survival. Both arms received adjuvant platin and 5-fluorouracil.196 patients with local–regionally advanced head and neck cancer However, another randomized study involving 408 patients who were randomized to RT alone (213 patients) or combined with received concurrent carboplatin at AUC 6 every 3 weeks after weekly cetuximab dosed in a standard fashion (211 patients); induction chemotherapy failed to demonstrate an improvement 434 Practice of Oncology / Cancer of the Head and Neck

TABLE 38.9 Selected Randomized Trials Evaluating Concurrent Chemoradiotherapy Versus Radiotherapy for Advanced Nasopharynx Cancer

Study No. of Patients Maintenance Chemotherapy Treatment Arms PFSa,b (p value) OSa (p value) Al-Saraff et al.190 147 Yes, on RT 24% 47% CDDP/RT arm CDDP/RT 69% 78% (<0.001) (0.005) Lin et al.191 284 No RT 53% 54% CDDP/FU/RT 72% 72% (0.0012) (0.0022) Chan et al.198 350 No RT 52% 59% CDDP/RT 60% 70% (0.06) (0.049) Wee et al.194 221 Yes, on RT 53% 65% CDDP/RT arm CDDP/RT 72% 80% (0.0093) (0.0061) Lee et al.195 348 Yes, on RT 62% 78% CDDP/RT arm CDDP/RT 72% 78% (0.027) (0.97) a Five-year rates for Lin et al. and Chan et al.; otherwise, 3-year rates. b Disease-free rate provided for Wee et al., and failure-free rate for Lee et al. PFS, progression-free survival; OS, overall survival; CDDP, cisplatin; FU, 5-fluorouracil; RT, radiation therapy.

in overall survival at 5 years or other disease outcomes compared cisplatin prior to concurrent cisplatin and RT was consistent with to patients receiving radiation alone after the same induction a benefit compared with concurrent cisplatin and RT alone.200 regimen.197 Programs incorporating newer taxane- and cisplatin-based triplet Another limitation of the Intergroup Study was that it was not induction regimens warrant further study.206 There is also interest designed to delineate the proportional benefits of concurrent and in the role of plasma EBV–DNA assays as a way to assess disease maintenance chemotherapy. Although current NCCN guidelines and monitor response.199 recommend concurrent and maintenance chemotherapy in M0 patients with a more advanced disease based on the Intergroup experience,146 in reviewing available data, the benefits of mainte- Organ Preservation nance chemotherapy appear more controversial. As noted, other randomized studies have demonstrated a survival improvement Organ preservation therapy is intended to control disease without with concurrent therapy alone.191,198 Earlier randomized trials, compromise in survival while optimizing function or cosmesis.207 summarized elsewhere, failed to demonstrate a survival benefit The term implies that the tumor is potentially resectable for cure, when either maintenance or induction chemotherapy was added and that the morbidity from surgery is significant. Although con- to definitive RT.199,200 Furthermore, a meta-analysis of updated in- servation surgical procedures can achieve the same goals, the label dividual patient data on 1,753 patients enrolled in eight random- of organ preservation is more commonly applied to nonsurgical ized trials, besides confirming an absolute survival benefit of 6% approaches. In that regard, the role of chemotherapy integrated at 5 years with incorporation of chemotherapy with RT (HR, 0.82, with RT is best established for more advanced primary tumors. In 95% CI, 0.71 to 0.91; p = 0.006), also reported a significant asso- this setting, conservation surgical procedures become less feasible, ciation between the timing of chemotherapy and overall survival and local control rates with RT alone are lower than seen with (p = 0.005), with the largest benefit being attributed to concomi- earlier stage disease. tant therapy.201 A recently reported randomized trial involving 508 Total laryngectomy is one of the surgical procedures most patients with nonmetastatic, stage III to IV nasopharyngeal cancer feared by patients.208 Thus, larynx preservation has been a cen- failed to demonstrate a significant difference in 2-year failure-free tral focus of many organ preservation studies, including those survival, overall survival, local–regional failure-free survival, or dis- that established integrated chemotherapy and RT as a standard tant failure-free survival after a median follow-up of 37.8 months. organ preservation treatment option. Studies commonly focused Of note, the direction of each of the previous endpoint compari- on patients with advanced tumors of the larynx, hypopharynx, and sons favored the adjuvant arm, albeit not significantly so, with as- oropharynx (particularly the base of tongue), in whom primary sur- sociated p values of 0.13, 0.32, 0.10, and 0.12, respectively.202 A gical management would jeopardize the voice box.209 longer term follow-up of this trial will be of interest. Currently, Initial chemoRT approaches to larynx preservation utilized in- the NCCN guidelines list both concurrent chemoRT followed by duction chemotherapy. The response to initial chemotherapy was maintenance chemotherapy and concurrent chemoRT alone as used to triage patients to either definitive RT (a partial response treatment options for advanced disease, with the former having the or better at the primary site; surgery to the primary site was re- higher category rating.146 served for salvage) or primary surgical management (lower than Selected randomized studies have demonstrated evidence of a a partial response). The randomized and landmark Veterans Ad- positive biologic effect with the use of induction chemotherapy, ministration (VA) Larynx Preservation Study demonstrated that but no survival benefit has been documented.200,203–205 Such such an approach could be pursued in patients with advanced promising results have engendered interest in the potential for laryngeal cancer without compromise in survival when com- enhanced efficacy with newer drugs and combinations. A ran- pared to primary treatment with surgery and RT.151 Over 60% of domized phase II trial evaluating induction with docetaxel and patients on the chemoRT arm avoided total laryngectomy. Among

tahir99 - UnitedVRG Chapter 38 Cancer of the Head and Neck 435

long-term survivors, patients treated on the chemoRT arm had although the difference was not statistically significant (median, better emotional well-being, were less depressed, and also reported 2.3 years versus 1.6 years; HR, 0.85; 95% CI, 0.68 to 1.06).213 less pain.210 Available randomized phase III data support a concurrent A similarly designed randomized trial in patients with pyriform chemotherapy–RT strategy administered with organ preservation sinus and aryepiglottic fold tumors reported by the European Or- intent for patients with advanced oropharynx cancer. A phase III ganization for Research and Treatment of Cancer (EORTC) con- study from the Groupe d’Oncologie Radiothérapie Tête et Cou firmed these findings.152 However, a small randomized study (n = (GORTEC) demonstrated improved local–regional control (66% 68) limited to patients with T3 disease with a fixed cord done by the versus 42% at 3 years, p = 0.03) in patients with advanced orophar- Groupe d’Etude des Tumeurs de la Tete et du Cou (GETTEC) ynx cancer who received concurrent chemotherapy (carboplatin reported that survival was superior on the primary surgery arm and 5-fluorouracil) and RT (70 Gy) compared to RT alone.147 A (84% versus 69% at 2 years, p = 0.006).211 When the MACH-NC non–site-specific trial from the Cleveland Clinic, which included performed a collective analysis of the VA, EORTC, and GETTEC a high proportion of patients with advanced oropharynx cancer, studies, the rate of larynx preservation among survivors was 58%. yielded similar results.214 A nonsignificant (6%) decrement in survival at 5 years was seen in Although concurrent chemotherapy is the current cornerstone the chemoRT group (39% versus 45%; pooled HR, 1.19; 95% CI, of organ preservation treatment of advanced disease, other para- 0.97 to 1.46; p = 0.10).149 digms deserve mention. In an Italian study, 195 patients with T2 The data reviewed in the prior section highlighting the thera- to T4 oral cavity cancer were randomized to either primary surgical peutic benefits of a concurrent chemoRT relative to an induction management or induction chemotherapy with cisplatin and 5-fluo- or RT alone approach have obvious implications for the larynx pres- rouracil followed by a surgical procedure, which could be modified ervation setting. RTOG 91-11 was designed to assess the impacts based on response. Overall survival was similar on both arms, but less of adding chemotherapy to RT and its timing (concurrent versus postoperative RT was necessary (33% versus 46%) and fewer man- induction) with regard to achieving larynx preservation. Four hun- dible resections were performed (31% versus 52%) on the chemo- dred and ninety-seven patients with larynx cancer were randomized therapy arm.215 As a further extension of this concept, Laccourreye to one of three arms: primary RT, 70 Gy to the primary site, 50 to and colleagues216 have pioneered the selective observation without 70 Gy to nodes; induction chemotherapy with cisplatin and infu- local–regional treatment of patients with laryngeal cancer who have sional 5-fluorouracil for three cycles followed by RT in responders, a complete response to induction chemotherapy. Durable tumor surgery in nonresponders; and cisplatin 100 mg/m2 on days 1, 22, control without the addition of surgery or RT has been reported in and 43 concurrent with RT. Surgical salvage was an option on all a small subset of patients with early stage tumors.216 The University three arms. The recently updated 10-year results are summarized of Michigan has developed a larynx preservation program whereby in Table 38.10.148,212 As anticipated, the rate of grade 3 or 4 mucosal the triage to RT-based treatment or surgery occurs after only one toxicity was highest on the concurrent arm; however, this did not cycle of chemotherapy.217 The intent is to improve survival and

translate into more significant speech or swallowing impairment at minimize morbidity through the timely selection of appropriate OF ONCOLOGY PRACTICE 2 years compared to the other treatment arms. Noteworthy is that, therapy, including referral to surgery if indicated. The implication although the larynx preservation rate and local–regional control is that induction chemotherapy has little other therapeutic benefit, was highest and statistically superior with concurrent treatment, and some patients who are slow to respond may be triaged unneces- there was no significant difference in overall survival rates among sarily to total laryngectomy.218 Conversely, newer sequential strate- the arms. Deaths not attributed to larynx cancer were highest in the gies of induction chemotherapy followed by planned concurrent concurrent arm (30.8%) versus 20.8% on the induction arm and chemotherapy are looking to optimize both local–regional and dis- 16.9% in the RT alone group, raising concern regarding the long- tant control. Induction with more efficacious triplet chemotherapy, term morbidity of concurrent therapy. However, late effects were including a taxane combined with cisplatin and 5-fluorouracil, is similar among the groups, and there were no substantial differences already being incorporated into larynx preservation strategies with in speech or swallowing function reported.212 evidence of improved larynx preservation rates.219 In another randomized larynx-preservation study, induction chemotherapy followed by RT and alternating chemotherapy and RT approaches were compared in 450 patients with advanced lar- Adjuvant Therapy after Surgery ynx or hypopharynx cancer. Both treatment arms used cisplatin and 5-flourouracil and allowed surgical salvage. Overall and pro- The use of maintenance chemotherapy after the completion of gression-free survival rates were similar on both arms. Survival with local–regional treatment has been evaluated in several random- a functional larynx in place was higher with alternating chemoRT, ized trials, but with disappointing results. Suboptimal compliance

TABLE 38.10 Intergroup 91-11: Updated Results at 10 Years

RX Arms No. of Patients LP Rate LRF DMFa DFS OS RT only 171 63.8% 47.2% 76.0%b 14.8%b 31.2%b p <0.0012 p = 0.0015 Induction PF→RT 173 67.5% 48.9% 83.4% 20.4% 38.8%a p = 0.005 p = 0.0037 p = 0.06 p = 0.06 p = 0.29 Concurrent P/RT 171 81.7%b 65.3%b 83.9% 21.6% 27.5% p = 0.08 p = 0.04 p = 0.53 a Trend, p = 0.08 for induction versus concurrent arm. b Comparison group. RX, treatment; LP, larynx preservation; LRF, local-regional failure; DMF, distant metastatic failure; DFS, disease-free survival; OS, overall survival; RT. radiation therapy; PF, cisplatin/5-fluorouracil; ,P cisplatin. From Forastiere AA, Zhang Q, Weber RS, et al. Long-term results of RTOG 91-11: a comparison of three nonsurgical treatment strategies to preserve the larynx in patients with locally advanced larynx cancer. J Clin Oncol 2013;31:845–852. 436 Practice of Oncology / Cancer of the Head and Neck with maintenance treatment may in part explain the lack of ben- Some investigators have questioned whether extracapsular efit, because tolerance of chemotherapy can be poor after surgery nodal spread has the same adverse prognostic implications in pa- and RT.220 Despite these limitations and the lack of convincing tients with HPV-related disease.226 The related therapeutic ques- survival benefit, patterns of failure were affected in selected stud- tion is whether this pathologic risk factor should drive the addition ies, with a decrease in distant metastases, consistent with a biologic of concurrent chemotherapy to radiation in the postoperative set- effect of chemotherapy.220,221 ting for patients with HPV-related disease. These issues are the The results of Intergroup 0034 highlight these points. In this focus of ongoing investigations. For now, the NCCN does not rec- trial, 442 analyzable patients were randomized after definitive sur- ommend different approaches to treatment in the poor-risk adju- gical therapy to either postoperative RT alone (50 to 60 Gy) or vant setting based on tumor HPV status.146 three cycles of standard dose cisplatin and 5-fluorouracil by infu- sion followed by the same RT-dosing scheme. The randomization was stratified by risk, with surgical margins less than 5 mm, cancer Toxicity Reduction in situ (CIS) at the margins, and ECE being considered high-risk features. Overall, there was no significant difference in overall sur- Xerostomia is one of the most troubling side effects of RT-based treat- vival, disease-free survival, or local–regional control between the ment. Extra oral fluids, artificial saliva, other topical measures, and treatment arms, although there was a significant decrease in inci- humidity are commonly utilized. Available data indicates that IMRT 34,227,228 dence of distant metastases on the investigational arm (p = 0.03). applied with salivary-sparing intent decreases this symptom. Interestingly, on a subset analysis, adjuvant chemotherapy had Historically, the emphasis has been on parotid sparing. Sparing no significant impact in the low-risk group, but a more dramatic other salivary glands when possible further improves salivary out- 229,230 impact on survival and tumor control was seen among high-risk comes. Pharmacotherapy may also help. The cholinomimetic patients. Given the success of concurrent chemoRT as definitive and muscarinic agent, pilocarpine, at a dose of 5 mg three times a treatment, its application in the adjuvant setting was a logical ex- day was shown in a randomized trial to improve the production of tension. A pilot study done by the RTOG demonstrated that con- saliva as well as symptoms of dry mouth compared to placebo in pa- 231 current high-dose cisplatin every 3 weeks with RT was feasible in tients treated with at least 40 Gy to the head and neck. Excessive the adjuvant setting.222 Early randomized studies demonstrated an sweating was the most common side effect. Cevimeline, a similar improvement in local–regional control with the incorporation of agent with a more selective mechanism of action, was associated concurrent mitomycin.223,224 A study comparing weekly cisplatin with a significant increase in unstimulated salivary flow at dosing of 232 concurrent with postoperative RT versus RT alone with ECE, 30 to 45 mg three times a day. Amifostine is a thiol with chemo- yielded a significant improvement in both local–regional control and radioprotectant properties. Objective and subjective measures of and survival with combined modality treatment.46 salivary function were improved in an open-label, randomized study 2 Two randomized studies, both published in 2004, have further among patients who received 200 mg/m of amifostine intravenous 233 clarified the indications for postoperative chemoRT in the poor daily 15 to 30 minutes before RT. Grade 3 toxicities were infre- risk adjuvant setting. RTOG 9501 and EORTC 22931 had very quent, but nausea, vomiting, hypotension, and allergic reactions similar designs.43–45,209 Patients were randomized after surgery if were more common among patients treated with amifostine. The 234 they had poor risk features to either standard postoperative RT benefits when given with chemoRT are more controversial. There alone (60 to 66 Gy, over 6 to 6.5 weeks, standard fractionation) have been concerns regarding the potential of tumor protection by or the same RT with three planned cycles of concurrent cisplatin amifostine, but a recent meta-analysis did not demonstrate any ad- at 100 mg/m2 every 3 weeks. What constituted poor risk differed verse impact on progression-free or overall survival in patients treated 234 somewhat between the studies: the RTOG required the presence with RT or chemoRT. Finally, acupuncture may be of benefit in 38,235 of two or more positive lymph nodes, ECE, or positive margins; the management of xerostomia in selected patients. the EORTC required ECE, positive margins, pT3 or pT4 with any Mucositis is a troubling symptom typically exacerbated by N, N2, or N3 disease, level IV nodes or stage IV disease in patients aggressive altered fractionated RT programs and the use of con- with oral cavity or oropharynx primaries, PNI, or vascular embo- current chemoRT. A variety of rinses (e.g., topical anesthetics, an- lism.43–45 Both studies demonstrated a significant improvement in tifungals) and systemic pain medications are used for symptomatic local–regional control and disease-free or progression-free survival relief. Concurrent amifostine with RT has not been clearly shown 233,234 with combined modality therapy. These improvements translated to decrease mucositis. Iseganan hydrochloride, a synthetic into a significant advantage in overall survival in the EORTC study peptide with broad spectrum antibacterial activity, was evaluated (p = 0.02), but only a trend (p = 0.19) in the RTOG study. Nei- in a multinational, double-blind, placebo-controlled trial among ther study showed a significant impact on distant control with the patients receiving definitive or postoperative RT-based therapy. No 236 addition of chemotherapy. Acute toxicity was greater with the ad- improvement in oral mucositis was found compared to placebo. dition of the cisplatin, but there was no difference in late toxicity. Recombinant human keratinocyte growth factor has been shown A subsequent analysis of the EORTC and RTOG studies was to decrease the incidence and duration of mucositis in the trans- 237 performed to better understand which pathologic subgroups may plant setting. Two randomized studies demonstrated improve- benefit the most from the concurrent addition of cisplatin to RT.203 ment in observer-assessed mucositis with palifermin use, but no Patients having evidence of ECE or a positive margin derived the clear difference in patient-reported pain, opioid analgesic use, or 238,239 largest benefit from combined modality adjuvant therapy. Con- treatment breaks when compared to placebo. versely, patients in whom their only poor risk factor was two or more positive lymph nodes without ECE seemed to do just as well with RT alone. Long-term follow-up of RTOG 9501 yielded results FOLLOW-UP consistent with this conclusion.45 The EORTC and RTOG studies focused on patients who were How to optimally follow-up with patients after treatment for previously untreated with the exception of prior surgery. Janot et head and neck cancer is less well studied than how to treat it. Ap- al.,225 on behalf of the GETTEC and GORTEC groups, addressed proaches are informed by patterns of failure. Most relapses occur the potential role of concurrent chemotherapy and reirradiation after within the first 3 years and are front loaded, and relapses above the salvage surgery. In this randomized study enrolling 130 patients, the clavicles are potentially curable. The lung is the most common standard arm was salvage surgery alone. The combined modality site of distant spread. treatment significantly improved disease-free survival (HR, 1.68 [95% The schedule proposed in the NCCN practice guidelines is rea- CI, 1.13 to 2.5], p = 0.01), although overall survival was not signifi- sonable.146 Patients are seen for follow-up head and neck examina- cantly improved and both acute and chronic toxicities were worse. tions every 1 to 3 months during year 1, every 2 to 6 months during

tahir99 - UnitedVRG Chapter 38 Cancer of the Head and Neck 437 year 2, every 4 to 8 months during years 3 to 5, and every 12 months Clinical Picture thereafter. Thyroid function tests are obtained every 6 to 12 months if the neck was irradiated. Many practitioners obtain annual chest The vermilion of the lower lip is the most common site of origin. x-rays or other chest imaging to monitor for a second primary lung SCC may present as an enlarging discrete lesion that is not tender cancer and to document distant metastases. The impact of this until it ulcerates. Some lesions develop slowly on a background of imaging on outcomes is not well established, which is consistent leukoplakia or CIS and present as superficially ulcerated lesions with a vague recommendation from the NCCN of chest imaging with little or no bulk. Erythema of the adjacent skin suggests der- “as clinically indicated.” Tobacco history should be considered for mal lymphatic invasion. Palpation of the lip will reveal the extent purposes of lung cancer screening.146 Posttreatment baseline imag- of induration. Paresthesia of the skin of the lip indicates PNI. ing within 6 months of therapy is recommended. Additional studies such as CT, MRI, and PET scans may subsequently be necessary to determine whether there is a recurrence or a complication, but Treatment otherwise are not routinely performed in surveillance. Speech and swallowing evaluations and rehabilitation are obtained as indi- Selection of Treatment Modality cated. Counseling is indicated for patients in whom tobacco or Early lesions may be cured equally well with surgery or RT. Surgi- alcohol contributed as a risk factor for tumor development. cal excision is preferred for the majority of lower lip lesions up to 2 cm in diameter that do not involve the commissure; the treatment is simple and the cosmetic result is satisfactory. Removal of more ORAL CAVITY of the lip with simple closure usually results in a poor cosmetic and functional result and, therefore, requires reconstructive proce- The oral cavity consists of the lips, the floor of mouth, the anterior dures. RT is often preferred for lesions involving the commissure, two-thirds of the tongue, the buccal mucosa, the upper and lower for lesions over 2 cm in length, and for upper lip carcinomas. Ad- alveolar ridges, the hard palate, and the retromolar trigone. vanced lesions with bone, nerve, or node involvement frequently The AJCC staging system is used.26 require a combined modality approach. The regional lymphatics are not treated electively for early cases. Advanced lesions, high-grade lesions, and recurrent lesions LIP should be considered for elective neck treatment. Clinically posi- tive nodes are managed as previously discussed in Clinically Posi- The ratio between men and women with lip cancer is approxi- tive Neck Lymph Nodes. mately 15:1.240 Persons with light-colored skin and/or prolonged exposure to sunlight are most prone to develop lip carcinoma. Surgical Treatment PRACTICE OF ONCOLOGY PRACTICE Surgical treatment for early lesions (0.5 to 1.5 cm) utilizes a V- Anatomy or W-shaped excision, depending on the size of the defect, which facilitates cosmetic primary closure. If the vermilion is diffusely in- The lips are composed of the orbicularis oris muscle with skin on volved with little or no involvement of the muscle, a vermilionec- the external surface and mucous membrane on the internal surface. tomy may be performed and the mucosa from the labial vestibule The transition from skin to mucous membrane is the lip vermilion. of the oral cavity advanced to cover the defect. The blood supply is from the labial artery, a branch of the facial artery. The motor nerves are branches of cranial nerve (CN) VII. Irradiation Technique The sensory nerve to the upper lip is the infraorbital branch of Lip cancer may be successfully treated by EBRT, interstitial CN V (V 2), and the mental nerve (V 3) supplies the lower lip. brachytherapy, or a combination of both. Interstitial brachytherapy may be accomplished with removable sources such as iridium-192 Pathology (192Ir). EBRT techniques use orthovoltage (55.8 Gy at 1.8 Gy per fraction) or electrons (60 to 66 Gy at 2 Gy per fraction) with lead The most common neoplasms are SCCs. Basal cell carcinomas shields behind the lip to limit exit EBRT. IMRT is not indicated arise on the skin of the lip and may secondarily invade the vermil- except for the occasional patient with advanced neck disease and/ ion. Keratoacanthoma occurs on the skin of the lips and may be or clinical PNI where it is necessary to extend the dose distribution mistaken grossly and histologically for SCC. to the skull base and reduce the dose to the contralated parotid. Leukoplakia and CIS are common problems on the lower lip For more advanced lesions, combining chemotherapy with EBRT and may precede the appearance of carcinoma by many years. Pri- is appropriately considered.146 mary lesions arising from the moist mucosa of the lip are consid- ered under the section Buccal Mucosa. Results of Treatment

Patterns of Spread MacKay and Sellers242 reviewed 2,864 patients with all stages of lip cancer, of whom 92% were managed initially by RT. The primary SCC can originate from the skin of the lip or the vermilion, which lesion was controlled by the initial treatment in 84% of cases; an may invade the adjacent skin and orbicularis muscle. Advanced additional 8% were salvaged by later treatment for an overall local lesions invade the adjacent commissures of the lip, the buccal mu- control rate of 92%. Of those who presented with clinically involved cosa, the skin and wet mucosa of the lip, the adjacent mandible, nodes, 58% had control of disease, but only 35% had control of dis- and eventually the mental nerve. PNI occurred in 2% of the cases ease when neck nodes appeared later. The 5-year, cause-specific reported by Byers and coworkers241 and was related to recurrent survival rate was 89%; the 5-year absolute survival rate was 65%. lesions, large tumor size, mandibular invasion, and poorly differen- Mohs and Snow243 reported the results for 1,448 patients tiated histology. Lymphatic spread is to the submental (IA) and sub- treated with microscopically controlled surgery for SCCs of the mandibular (IB) lymph nodes and then to the jugular chain. The lower lip between 1936 and 1976. Eighty-three percent had can- risk for lymph node metastases is approximately 5% at diagnosis cers less than 3 cm in diameter, with a 5-year cure rate of 96.6%. and is increased by high-grade histology, large lesions, invasion of For 192 patients with cancers that measured 2 cm or more, the the mucosa of the lip, and for patients with recurrent disease. cure rate dropped to 60%. 438 Practice of Oncology / Cancer of the Head and Neck

Complications of Treatment from 20% to 35%.39,244 For T1 or superficial T2 lesions, the risk for occult metastasis is probably 10% to 15%.39,244 Oral competence, which permits patients to control oral secretions The first nodes involved are in levels IB and II; the risk for bi- and effectively suck, speak, and swallow, requires the sphincteric lateral spread is fairly high. function of an intact orbicularis oris muscle. Hence, disruption of the sphincteric function resulting from division of the orbicularis Clinical Picture oris should be restored. Microstomia and drooling secondary to oral incompetence may occur after a large flap reconstruction. If On physical examination, the earliest lesions appear as a red area, the oral opening is too small, the patient may not be able to inset slightly elevated, with ill-defined borders, and very little induration. As a denture. the lesion enlarges, the edges of the tumor become distinct, elevated, There will be some atrophy of the irradiated tissues; this pro- and “rolled,” with a central ulceration and induration. Some lesions gresses with time. Soft tissue necrosis may occur, but this problem start with a background of leukoplakia. Bimanual palpation will de- is reduced by plans that prolong the treatment. termine the extent of the induration and the degree of fixation to the periosteum. Large lesions bulge into the submental space and rarely grow through the mylohyoid muscle into the soft tissues of the neck. FLOOR OF THE MOUTH Gross invasion of the mandible may be detected, especially when the anterior teeth have been removed. A tumor may grow through the Anatomy mandible to involve the gingivolabial sulcus and lip. The subman- dibular duct and gland are evaluated by bimanual palpation. The floor of the mouth is a U-shaped area bounded by the lower gum and the oral tongue; it terminates posteriorly at the anterior Treatment tonsillar pillar. The paired sublingual glands lie immediately below the mucous membrane; the paired genioglossus and geniohyoid muscles separate them. Bony protuberances, the genial tubercles, Selection of Treatment Modality occur at the point of insertion of these two muscle groups at the symphysis. The mylohyoid muscle arises from the mylohyoid ridge Early Lesions. Surgery or RT are equally effective treatments for of the mandible and is the muscular floor for the oral cavity; it T1 or T2 lesions. Most patients are treated surgically because of ends posteriorly at about the level of the third molars. The sub- the risk of soft tissue or bone necrosis after RT. mandibular gland rests on the external surface of the mylohyoid A few patients are seen after excisional biopsy of a tiny lesion, muscle between the mandible and the insertion of the mylohyoid. and the only finding is a surgical scar with varying degrees of in- The submandibular duct (the Wharton duct) is about 5 cm long. It duration under the scar (TX). The margins are often equivocal. courses between the sublingual gland and the genioglossus muscle These patients are treated with reexcision or brachytherapy. and exits in the anterior floor of the mouth near the midline. Moderately Advanced Lesions. The usual recommendation for moderately advanced anterior midline lesions is rim resection Pathology or segmental mandibulectomy and osteomyocutaneous free flap reconstruction; postoperative RT or chemoRT is added depending Most neoplasms are SCC, usually of moderate grade. Adenoid on the pathologic findings. The clinically N0 neck is usually man- cystic and mucoepidermoid carcinomas account for about 5% of aged by a bilateral functional neck dissection for midline lesions. malignant tumors in this area. Advanced Lesions. Massive lesions have a poor prognosis with Patterns of Spread combined surgery and postoperative chemoRT. Only palliation can be offered in some cases. Primary Surgical Treatment Approximately 90% of neoplasms originate within 2 cm of the anterior midline floor of the mouth, penetrating early beneath Wide Local Excision. Small lesions (5 mm or less in size) may the mucosa into the sublingual gland and eventually into the ge- be excised transorally with a 1-cm margin with primary closure or nioglossus and geniohyoid muscles. The mylohyoid muscle acts a skin graft. If the duct is involved, the submandibular gland and as an effective barrier until the lesion becomes advanced. Exten- duct are removed in continuity. sion toward the gingiva and periosteum of the mandible occurs early. When the tumor reaches the periosteum, the tumor usu- Rim Resection. Rim resection of the mandible in continuity ally spreads along the periosteum rather than through it. Mandible with excision of the primary lesion preserves the arch and may be invasion is a late manifestation. The skin of the lower lip may combined with postoperative RT. Periosteal invasion is often an be involved in advanced cases. Posterior extension occurs in the indication for this procedure. Patients who have been edentulous muscles of the root of the tongue. One or both submandibular for a long time may have an atrophic mandible and are not suit- ducts are frequently obstructed by the tumor or after the biopsy; it able because the mandible is likely to fracture. may be difficult to distinguish between tumor extension and infec- tion in an obstructed duct. The submandibular gland frequently Segmental Mandibulectomy. A partial segmental mandibu- enlarges, becoming firm and occasionally painful when the duct lectomy with resection of the floor of the mouth is done for le- is obstructed. Extensive lesions may follow the anatomic plane of sions invading bone. An osteomyocutaneous flap is usually used the mylohyoid muscle to its posterior extremity and emerge in the to repair the defect. submandibular space of the neck. Irradiation Technique Lymphatic Superficial T1 cancers are treated with either brachytherapy or in- Approximately 30% of patients will have clinically positive nodes traoral cone RT to approximately 65 Gy, and the neck is observed. on presentation; 4% will have bilateral nodes. The reported inci- Larger lesions are treated with EBRT to 45 to 50 Gy over 5 weeks fol- dence of conversion from N0 to N+ with no neck treatment varies lowed by an interstitial implant for an additional 20 to 30 Gy. Lesions

tahir99 - UnitedVRG Chapter 38 Cancer of the Head and Neck 439 that are suitable for intraoral cone RT may be boosted with this tech- and repeat deep biopsies may be necessary. An enlarged submandib- nique prior to EBRT of the primary lesion and upper neck. Use of ular gland(s) may be a sequel to obstruction of the submandibular EBRT alone results in suboptimal cure rates and is discouraged.245 duct; a contrast-enhanced CT is useful to distinguish between an enlarged submandibular gland and a tumor in a lymph node. External-Beam Irradiation. Opposed lateral EBRT portals are The follow-up for surgical cases may be difficult if skin grafts used to treat anterior floor of the mouth carcinomas. The entire or flaps have been used because of the associated induration and width of the mandibular arch is included and the superior border thickness of the flaps. If the submandibular ducts have been reim- is shaped to spare part of the parotid gland. The level I and level II planted, stenosis may occur with subsequent enlargement of the nodes are included to the level of the thyroid notch if the neck is submandibular glands. clinically negative; the lower neck may be electively irradiated. If the neck is clinically positive, the portals are enlarged to include Complications of Treatment all of the upper neck nodes, and an en face low neck field is added. IMRT may be useful to reduce the dose to the contralateral parotid Radiation Therapy. A small soft-tissue necrosis may develop, in patients with positive nodes. usually in the site of the original lesion where the dose is high- est. These are moderately painful and respond to local anesthetics, Interstitial Irradiation. The implantation of T1 to T2 lesions antibiotics, and the tincture of time. Treatment with pentoxifylline confined to the floor of the mouth with minimal extension to the 400 mg three times daily may be beneficial. mucosa of the tongue can be accomplished with iridium using the If the ulceration develops on the adjacent gingiva, the underly- plastic tube technique. ing mandible is exposed. These areas are mildly painful. They are managed by discontinuing dentures, local anesthetics, antibiotics, Intraoral Cone Irradiation. Intraoral orthovoltage or electron and smoothing of the bone by filing, if needed. These small bone cone RT requires daily positioning by the physician and is preferable exposures do not often progress to osteoradionecrosis (ORN) and to interstitial RT because there is little or no irradiation of the mandi- either sequestrate a small piece of bone or are simply recovered by ble.246 An intraoral cone can be used for well-circumscribed anterior the mucous membrane. Severe ORN may require daily hyperbaric superficial lesions and is easiest to perform in the edentulous patient. oxygen (HBO) treatments for 4 to 6 weeks, either alone or in con- junction with surgical intervention. Combined Treatment Policies Surgical. Postoperative RT is preferred, because the risk of bone complica- These include bone exposure, orocutaneous fistula, tions and fistulae is higher with preoperative RT. Concurrent che- and failure of osteomyocutaneous flaps. Salvage procedures after motherapy may be necessary based on pathologic findings. RT are associated with an increased risk of complications.

Management of Recurrence OF ONCOLOGY PRACTICE ORAL TONGUE RT failures are treated by an operation. The salvage rate is good for patients with T1 to T2 lesions and poor for those with more advanced lesions. Anatomy Surgical treatment failures may be treated by a repeat operation and postoperative RT. The circumvallate papillae locate the division between the oral tongue and the base of the tongue. The arterial supply is mainly by way of paired lingual arteries that are branches of the external Results of Treatment carotid. The sensory pathway is by way of the lingual nerve to the gasserian ganglion. Rodgers et al.247 reported on 194 patients treated with surgery and/ or RT at the University of Florida between 1964 and 1987. The local control rates after RT versus surgery alone or combined with Pathology RT were: T1, 32 out of 37 (86%) versus 10 out of 11 (91%); T2, 25 out of 36 (69%) versus 16 out of 19 (84%); T3, 11 out of 20 (55%) More than 95% of oral tongue lesions are SCCs. Coexisting leuko- versus 9 out of 9 (100%); and T4, 2 out of 5 (40%) versus 6 out plakia is common. Verrucous carcinoma and minor salivary gland of 10 (60%).247 The 5-year cause-specific survival rates were com- tumors are uncommon. Granular cell myoblastoma is a benign parable for the treatment groups.247 Mild-to-moderate and severe tumor of uncertain origin that occurs on the dorsum of the tongue complications were observed as follows: RT alone, 49 out of 117 and may be confused histologically with carcinoma. (42%) and 6 out of 117 (5%); surgery alone, 3 out of 36 (8%) and 6 out of 36 (17%); and surgery and RT, 8 out of 41 (20%) and 6 out Patterns of Spread of 41 (15%), respectively.247 Two hundred seven patients treated with RT alone at the Cen- Primary tre Alexis Vautin between 1976 and 1992 were reviewed by Pernot and colleagues.248 Local control and cause-specific survival rates at Nearly all SCCs occur on the lateral and ventral middle and pos- 5 years were as follows: T1, 97% and 88%; T2, 72% and 47%; and terior thirds of the oral tongue. They tend to remain in the tongue T3, 51% and 36%, respectively. Six percent of patients developed until large unless they originate near the junction with the floor of complications necessitating surgical intervention and one patient the mouth. PNI and vascular space invasion may occur. experienced a fatal complication. Anterior third lesions usually are diagnosed early. Advanced le- sions invade the floor of the mouth and root of the tongue, produc- Follow-Up ing ulceration and fixation. Posterior third lesions grow into the anterior tonsillar pillar and base of tongue. There are two major difficulties in follow-up after RT: soft tissue ulcers and enlarged submandibular glands. An ulcer in the floor of Lymphatics the mouth within 2 years of treatment can be either a recurrence or necrosis. If the lesion appears to be soft tissue necrosis, a trial of con- The first-echelon nodes are the level Ib and II nodes.20 The sub- servative therapy is adequate. Failure to stabilize or resolve is an in- mental and level V lymph nodes are seldom involved. Rouvière13 dication for biopsy. A negative biopsy does not rule out recurrence, describes lymphatic trunks that bypass the level I to II nodes and 440 Practice of Oncology / Cancer of the Head and Neck

terminate in the level III lymph nodes. Byers et al.249 evaluated Advanced Lesions (T4). Near total or total glossectomy and nodal spread pattern in 277 patients treated surgically at the M.D. sometimes a laryngectomy is performed. Anderson Cancer Center and observed skip metastases to the level III or IV nodes without involvement of levels I and II in 16% of pa- Irradiation Technique tients. Of patients with oral tongue cancer, 35% have clinically posi- tive nodes at diagnosis and 5% are bilateral. The incidence of occult The ability to control the primary lesion is enhanced by giving all or part of the treatment with an interstitial RT or by intraoral disease is approximately 30%. The incidence of positive nodes in- 250–252 192 creases with T stage. Patients with N1 or N2 ipsilateral nodes have a cone. Superficial T1 tumors may be treated with Ir brachy- significant risk of developing node metastasis in the opposite neck. therapy alone using the plastic tube technique. Larger lesions that have an increased risk for subclinical neck disease may be treated with EBRT and a brachytherapy boost or with brachytherapy com- Clinical Picture bined with an elective neck dissection. The time factor is criti- cal for oral tongue cancer, and the EBRT part of the treatment is Mild irritation of the tongue is the most frequent complaint. As ul- shortened (30 Gy in 10 once-daily fractions, or 38.4 Gy in 1.6 Gy ceration develops, the pain worsens and is referred to the external twice-daily fractions) in order to increase the proportion of the RT ear canal. Extensive infiltration of the muscles of the tongue affects given by either interstitial or intraoral cone therapy. The intersti- speech and deglutition and is associated with a foul odor. tial therapy is given after the EBRT; the intraoral cone therapy The extent of disease is determined by visual examination and should be done prior to the EBRT. Elective neck RT is indicated palpation. The tongue protrudes incompletely and toward the side for nearly all lesions. of the lesion as fixation develops. Posterior oral tongue lesions may grow behind the mylohyoid and present as a mass in the neck at Combined Treatment Policies the angle of the mandible. Invasion of the hypoglossal nerve is rare. Postoperative RT or chemoRT is administered to the primary site Differential Diagnosis and neck for indications previously outlined. IMRT may be useful to reduce the dose to one or both parotids. The differential diagnosis includes granular cell myoblastomas, which are usually slow growing, nontender masses and 0.5 cm Management of Recurrence to 2.0 cm in size. The lesions are well circumscribed, firm, and Local recurrence after RT or surgery is heralded by ulceration, slightly raised; they may be multiple. Aggressive behavior is rare, pain, or increased induration. Recurrences have a slightly elevated and wide local excision is preferred. Pyogenic granulomas mimic or rolled border, whereas necroses do not. A biopsy should be done small exophytic carcinomas. Tuberculous ulcer and syphilitic as soon as ulceration appears if it is within the original tumor site. chancre are rare. Ulcers that appear on adjacent normal tissues are likely due to RT and not cancer. Treatment RT failure is managed by surgery. Surgical failure occasionally is salvaged by re-resection and postoperative RT-based treatment. Selection of Treatment Modality Recurrence in the soft tissues of the neck is rarely eradicated by any procedure. Both surgery and RT result in cure rates that are similar for similar Nodes appearing in a previously untreated neck are managed stages. The disadvantages of surgery include removal of part of the by neck dissection with or without postoperative RT or chemoRT. tongue and the decision of whether to do a neck dissection for the N0 neck. The disadvantage of RT is the risk of necrosis. Results of Treatment

Excisional Biopsy (TX). An excisional biopsy of a small lesion The local control rates for 170 patients treated with RT alone ver- may show inadequate or equivocal margins. An interstitial implant sus surgery alone or with RT between 1964 and 1990 at the Uni- = or reexcision will produce a high rate of local control. versity of Florida included: for T1, 79% versus 76% (p 0.76); for T2, 72% versus 76% (p = 0.86); for T3, 45% versus 82% (p = = 253 Early Lesions (T1 or T2). A partial glossectomy with primary 0.03); and for T4, 0% versus 67% (p 0.08). The differences in closure or a skin graft may be done transorally and is usually the 5-year survival between the two treatment groups were not statisti- preferred therapy. Depending on the depth of invasion, an elective cally significant. neck dissection may be indicated. Postoperative RT would only be The results of brachytherapy alone or combined with EBRT added for indications previously discussed. for 448 patients treated at the Centre Alexis Vautin were re- ported by Pernot et al.254 and revealed the following 5-year local Moderately Advanced Lesions (T2 or T3). The preferred control and survival rates: T1, 93% and 69%; T2, 65% and 41%; treatment for the majority of these patients is partial glossectomy, and T3, 49% and 25%, respectively. Shorter time intervals be- neck dissection, and postoperative RT-based treatment. tween brachytherapy and EBRT were associated with significantly improved local control and survival for those who received both Advanced Lesions (T4). Bi- or trimodality treatment will cure modalities. a minority of these patients. Some patients are best treated with palliative intent. Complications of Treatment

Surgical Treatment Surgical. Orocutaneous fistula, flap necrosis, and dysphagia are the most common complications after surgery. Damage to the lin- Early Lesions (T1 or T2). A partial glossectomy and primary gual nerve or the hypoglossal nerve is rare. Fistula and flap necrosis closure is performed. may result in carotid artery hemorrhage. Enunciation difficulties occur whenever the tongue is bound down by scarring. The inci- Moderately Advanced Lesions (T2 or T3). A partial glossec- dence of complications increases for surgical salvage attempts after tomy with primary closure, skin graft, or flap reconstruction is per- RT failure. Of 65 patients, 13 (20%) treated with surgery alone or formed. Frozen section control is essential. Positive margins are an combined with RT at the University of Florida developed signifi- indication for excision of additional tissue. cant complications.253

tahir99 - UnitedVRG Chapter 38 Cancer of the Head and Neck 441

Radiation Therapy. A minor soft-tissue necrosis is fairly com- Differential Diagnosis mon and is treated with broad-spectrum antibiotics, local anesthet- ics such as viscous lidocaine, and analgesics. Pentoxifylline 400 The differential diagnosis includes lues and tuberculosis; both are mg three times daily may be beneficial. Hyperbaric oxygen treat- rare. If the first biopsy reveals chronic inflammation or pseudoepi- ment may be tried in difficult cases. If the necrosis is persistent and theliomatous hyperplasia, a repeat biopsy may be necessary. the pain is uncontrollable, it must be resected. The edentulous person is less likely to develop bone com- plications compared with those who are dentulous.255 The most Treatment frequent problem involving the mandible is bone exposure. If the patient has dentures, they should be discontinued or altered to re- Selection of Treatment Modality lieve the pressure over the exposed bone. If sharp bony edges ap- Small lesions (≤1 cm) may be excised with primary closure; small pear, they are filed and the bone edge is lowered to speed healing. lesions that involve the lip commissure are sometimes treated by Healing may require months or even years. RT. Lesions 2 cm to 3 cm in size can be treated with surgery or If ORN develops, HBO has been used with some success. by RT (usually the former). Larger lesions are usually treated with If conservative measures are unsuccessful, segmental man- surgery, and postoperative RT or chemoRT. dibulectomy and an osteomyocutaneous flap reconstruction is performed. Surgical Treatment. Lesions that invade the mandible or max- Severe complications were observed in 9 of 105 patients (9%) illa require bone resection along with the soft tissues. Repair may treated with RT at the University of Florida.253 Pernot and col- 254 require a maxillary prosthesis. A myocutaneous flap repairs full- leagues observed the following soft tissue and/or bone complica- thickness removal of the cheek. tions in a series of 448 patients: Grade 1, 19%; Grade 2, 6%; and Grade 3, 3%. Irradiation Technique. Buccal mucosa lesions are suited for treatment with electrons, an intraoral cone, and interstitial tech- niques to spare the contralateral normal tissues. When tumors ex- BUCCAL MUCOSA tend into one of the gingivobuccal gutters or onto bone, treatment must be entirely by EBRT. Epidemiology Results of Treatment SCC is relatively uncommon in the United States. In Southern Diaz et al.257 recently reported the M.D. Anderson experience for India, it is common and is related to chewing a combination of 119 patients treated with surgery alone (84 patients) or combined tobacco mixed with betel leaves, areca nut, and lime shell.256

with adjuvant RT (35 patients) between 1974 and 1993. Tumor OF ONCOLOGY PRACTICE recurrence developed in 54 patients (45%): local recurrence in 27 Anatomy patients (23%); regional recurrence in 13 patients (11%); local and regional recurrence in 11 patients (9%); and distant metastases in The buccal mucosa is the mucous membrane covering the inner 3 patients (3%). The 5-year survival rates versus stage were stage I, 78%; stage II, 66%; stage III, 62%; stage IV, 50%; and overall, 63%. surface of the cheeks and lips, ending above and below with a tran- 256 sition to the gingiva. It ends posteriorly at the retromolar trigone. Nair and coworkers reported the definitive RT results for 234 The parotid duct opens into the buccal mucosa opposite the sec- cases of buccal mucosa cancer treated in southern India during the ond upper molar. The buccal mucosa is innervated by a branch of 1982 calendar year. The 3-year disease-free survival rates were stage the mandibular nerve. I, 85%; stage II, 63%; stage III, 41%; and stage IV, 15%. Thirty-two patients had verrucous carcinoma, and the 3-year disease-free sur- vival rate was 47%, similar to that for other grades of SCC. Pathology Complications of Treatment Most malignant tumors are low-grade SCCs that frequently ap- The buccal mucosa is tolerant of high-dose RT, and complications pear on a background of leukoplakia or lichen planus. Verrucous are uncommon. Bone exposure may appear on the mandible or carcinoma occurs. Minor salivary gland tumors and melanomas maxilla. Trismus may develop if the muscles of mastication receive are rare. high doses. Surgical injury of the Stensen duct may cause obstruction and Patterns of Spread parotitis. Injury to branches of the VII nerve may occur. Split- thickness skin grafts may shrink and produce a partial trismus. Early lesions are usually discrete and exophytic. As they enlarge, Resection of the lip commissure may produce oral incompetence. they penetrate the underlying muscles and eventually extend to the skin. Peripheral growth occurs into the gingivobuccal sulci and eventually onto the gingiva and into bone. GINGIVA AND HARD PALATE (INCLUDING The lymphatic spread is first to the level I and level II nodes. RETROMOLAR TRIGONE) The incidence of positive nodes on admission is 9% to 31%, and 20,244 the risk of occult disease is 16%. Anatomy

Clinical Picture The lower gingiva includes the keratinized masticatory mucosa covering the mandible from the gingivobuccal gutter to the ori- Small lesions produce the sensation of a lump that is felt with the gin of the nonkeratinized lining mucosa covering the floor of the tongue. Pain is minimal, unless there is posterior extension to in- mouth. The retromolar trigone lies behind the third molar and volve the lingual and dental nerves. Pain may be referred to the is contiguous superiorly with the maxillary tuberosity. Beneath ear. Obstruction of the Stensen duct will produce parotid enlarge- the keratinized mucosa of the retromolar trigone is the tendinous ment. Extension posteriorly, behind the pterygomandibular raphe pterygomandibular raphe, which is attached to the pterygoid ham- or into the buccinator and masseter muscles, causes trismus. ulus and the posterior mylohyoid ridge of the mandible and serves 442 Practice of Oncology / Cancer of the Head and Neck as the insertion of the buccinator, orbicular oris, and superior muscle produces trismus. Intra-alveolar SCC presents with a sub- pharyngeal constrictor muscles. Behind the pterygomandibular mucosal mass and dental symptoms. Roentgenograms show a lytic raphe and between the medial pterygoid muscle and the ascending lesion in the mandible. ramus is the pterygomandibular space, which contains the lingual Ameloblastoma exhibits few symptoms in the early stages. and dental nerves and is related posteriorly to the deep lobe of the Patients may notice a gradually increasing facial deformity or a parotid and the parapharyngeal space. There are no minor salivary loosening of teeth. An intraoral submucosal mass may be present glands in the mucous membranes of the alveolar ridges. initially; ulceration occurs as the mass increases in size. On roent- genograms, a radiolucent area is seen with the expansion of the Pathology overlying cortical plate, scalloped margins, a multilocular appear- ance, and/or resorption of the roots of adjacent teeth. Minor salivary gland tumors present as a submucosal mass, en- Most neoplasms are SCCs. Minor salivary gland tumors, usually 261 adenoid cystic carcinomas, often occur on the posterolateral hard large slowly, and may develop a central ulceration. palate.258 Verrucous carcinomas usually occur on the lower gin- giva. Melanoma has been reported.259 Differential Diagnosis SCC may arise within the body of the mandible or maxilla ei- ther from the odontogenic epithelium or from epithelium trapped The differential diagnosis includes dental disease and underlying during embryonic development. It is more frequent in the man- bony cysts or tumors. dible than the maxilla and is most common in the molar regions. It must be distinguished from metastatic SCC and ameloblastoma. Ameloblastoma is a rare, benign locally aggressive odontogenic Treatment tumor with an incidence of about 1% of all tumors of the maxilla and mandible; about 80% of cases occur in the mandible with the Selection of Treatment Modality molar–ramus region most commonly involved. Lower Alveolar Ridge. The majority of lesions are managed by Patterns of Spread surgery alone or followed by postoperative RT or chemoRT. Sur- gery entails a marginal mandibulectomy when there is, at most, Lower Gum saucerization of the underlying bone. Segmental mandibulec- tomy and free flap reconstruction is indicated for more advanced SCC invades the periosteum and the adjacent buccal mucosa and disease. floor of the mouth. Low-grade lesions tend to produce a smooth, saucerized defect before invading the mandible. Moderate to high- Ameloblastoma. The treatment is surgery; however, local grade lesions invade the bone directly or through recently opened recurrence is a problem. Sehdev and coworkers262 reported cu- dental sockets and produce a lytic defect. rettage was followed by local recurrence in 90% of mandibular Lymphatic spread is to the level I and level II nodes. Clinically ameloblastomas and in all maxillary ameloblastomas. Subsequent positive nodes occur in 18% to 52% of diagnoses; occult disease resection controlled 80% of the mandibular but only 40% of the occurs in 17% to 19%.20,244 maxillary tumors. The initial use of segmental mandibular resec- Ameloblastoma expands and destroys the bone and extends to tion controlled 78% (18 of 23 patients), with subsequent resection adjacent areas by contiguous growth. Ameloblastic carcinoma, a controlling those that recurred. The use of partial maxillectomy as rare malignant variant of ameloblastoma, may metastasize to re- the first treatment controlled 100% (7 of 7 patients) of maxillary gional nodes and distant sites.260 ameloblastomas as opposed to only 40% when a partial maxillec- tomy was performed for recurrence. Limited experience with RT Upper Alveolar Ridge and Hard Palate suggests that it may reduce the probability of progression and result in long-term local control in the occasional patient with incom- Most SCCs originate on the gingiva and spread secondarily to the pletely resectable disease.260 hard palate, soft palate, buccal mucosa, and underlying bone. The maxillary antrum is invaded late unless there are recent extractions Retromolar Trigone. Surgery is preferred for discrete early le- providing access. The risk for positive lymph nodes at diagnosis is sions. RT is recommended for superficial lesions involving a large 13% to 24%, and the incidence of occult disease is 22%.20,244 surface area.263 Advanced carcinomas are treated with surgery and Retromolar Trigone postoperative RT or chemoRT Carcinomas spread to the adjacent buccal mucosa, the anterior Upper Alveolar Ridge and Hard Palate. Resection alone or tonsillar pillar, and the maxilla. Posterior spread occurs into the followed by RT or chemoRT is the usual treatment for most le- pterygomandibular space and the medial pterygoid muscle. Pos- sions. However, if the lesion is superficial and extensively involves terolateral spread occurs into the buccinator muscle and fat pad. the hard palate or involves a significant portion of the soft palate, The first echelon lymphatics are the level I and level II nodes. then an RT-based approach should be considered for the initial The incidence of clinically positive nodes on presentation is about therapy. If the lesion is small and discrete and there is no bone 30%; the risk for occult disease is 15% to 25%. involvement, resection includes the periosteum or occasionally some underlying bone. Bone invasion requires a maxillectomy Clinical Picture that is tailored to optimally resect the cancer. The resulting defect is usually rehabilitated with a removable prosthesis. The patient with SCC may present to the dentist first with ill-fitting Irradiation Technique dentures, pain, loose teeth, or a sore that will not heal. A history of inappropriate dental extractions or root canal therapy is common. Small lesions of the lower alveolar ridge and retromolar trigone Invasion into the mandible may involve the inferior alveolar nerve may be treated by intraoral cone for all or part of their therapy. and produce paresthesia of the lower lip. A background of leuko- Well-lateralized lesions of the retromolar trigone and posterior al- plakia is frequently present. veolar ridge may be treated by either an ipsilateral mixed beam or Retromolar trigone lesions have pain referred to the external IMRT; the latter is preferred. Parallel-opposed portals treat ante- auditory canal and preauricular area. Invasion of the pterygoid rior gum lesions.

tahir99 - UnitedVRG