On the Road to Precision Cancer Medicine: Analysis of Genomic Biomarker Actionability in 439 Patients Maria Schwaederle, Gregory A

Published OnlineFirst April 7, 2015; DOI: 10.1158/1535-7163.MCT-14-1061

Companion Diagnostics and Cancer Biomarkers Molecular Cancer Therapeutics On the Road to Precision Cancer Medicine: Analysis of Genomic Biomarker Actionability in 439 Patients Maria Schwaederle, Gregory A. Daniels, David E. Piccioni, Paul T. Fanta, Richard B. Schwab, Kelly A. Shimabukuro, Barbara A. Parker, and Razelle Kurzrock

Abstract

Despite the increased use of molecular diagnostics, the extent to patients (11%) had only one alteration; and 372 patients had two which patients who have these tests harbor potentially actionable or more abnormalities (85%). The median number of potentially aberrations is unclear. We retrospectively reviewed 439 patients actionable anomalies per patient was 2 (range, 0–8). Most with diverse cancers, for whom next-generation sequencing patients (393/439; 90%) had at least one potentially actionable (mostly 236-gene panel) had been performed. Data pertaining alteration, and in all these cases the aberration could at least be to the molecular alterations identified, as well as associated targeted by an experimental drug in a clinical trial. A total of 307 treatment suggestions (on- or off-label, or experimental), were patients (70%) had an alteration that was actionable with an extracted from molecular diagnostic reports. Most patients (420/ approved drug, but in only 89 patients (20%) was the drug 439; 96%) had at least one molecular alteration: 1,813 alterations approved for their disease (on-label). Next-generation sequencing (in 207 distinct genes) were identified [the majority being muta- identified theoretically actionable aberrations in 90% of our tions (62%) or amplifications (29%)]. The three most common patients. Many of the drugs are, however, experimental or would gene abnormalities were TP53 (44%), KRAS (16%), and PIK3CA require off-label use. Strategies to address drug access for patients (12%). The median number of alterations per patient was 3 harboring potentially actionable mutations are needed. Mol Cancer (range, 0–16). Nineteen patients (4%) had no alterations; 48 Ther; 14(6); 1–7. 2015 AACR.

Introduction an survival (11–13). In the phase I setting, molecular matching was associated with improved outcomes in multivariate anal- The strategy of matching targeted drugs to biologically rel- ysis (14). Further, a systematic review of phase II clinical trials evant targets using molecular profiling techniques is becoming in advanced/metastatic non–smallcelllungcancershowedthat better established, though many challenges remain (1–3). molecular matching of patients' tumors to drugs was indepen- Indeed, the presence of genomic aberrations in tumors may dently associated with better outcomes, including higher medi- be critical to achieving response, especially when using agents an response rate (48.8% vs. 9.7%; P ¼ 0.005), longer median with specific molecular targets. For instance, KIT kinase inhi- progression-free survival (6 vs. 2.8 months; P ¼ 0.005), and bitors are effective in patients with KIT mutations (4, 5). overall survival (11.3 vs. 7.5 months; P ¼ 0.05), as compared Similarly, preclinical and emerging clinical data suggest that with those of unselected patients (15). PI3K inhibitors may be most effective in patients with PI3K or The percentage of patients with cancer who have potentially PTEN aberrations (6, 7); RAF inhibitors, in patients with RAF "actionable" aberrations remains a matter of debate. In order to mutations (8); MEK inhibitors, in patients with RAS or RAF determine the proportion of individuals who have druggable mutations (9, 10), and so on. In some cases, matching patients alterations, we analyzed the molecular profile results of patients with targeted therapies has resulted in transformative changes. who had had a targeted next-generation sequencing panel per- For instance, treatment of chronic myelogenous leukemia, a formed on their tumor. disease driven by an aberrant BCR-ABL kinase, with imatinib BCR-ABL (a kinase inhibitor) has dramatically increased medi- Patients and Methods Patients Center for Personalized Cancer Therapy, and Division of Hematology We retrospectively reviewed the medical charts of 439 patients and Oncology, UCSD Moores Cancer Center, La Jolla, California. with diverse cancers, for whom molecular testing had been Note: Supplementary data for this article are available at Molecular Cancer performed, and who were seen at the UCSD Moores Cancer Therapeutics Online (http://mct.aacrjournals.org/). Center (La Jolla, CA) from October 2012 until July 2014. This Corresponding Author: Maria Schwaederle, Center for Personalized Cancer study was performed and consents were obtained in accordance Therapy, UCSD Moores Cancer Center, 3855 Health Sciences Drive, La Jolla, CA with UCSD Institutional Review Board guidelines. 92093-0658. Phone: 858-822-2171; Fax: 858-822-2300; E-mail: [email protected] Next-generation sequencing doi: 10.1158/1535-7163.MCT-14-1061 Next-generation sequencing was performed by Foundation 2015 American Association for Cancer Research. Medicine (FoundationOne, http://www.foundationone.com),

www.aacrjournals.org OF1

Schwaederle et al.

which is a clinical-grade CLIA-approved next-generation Molecular diagnostic test results sequencing test that sequences the entire coding sequence of Overall, 1,813 alterations (found in 207 distinct genes) were 236 cancer-related genes and 47 introns from 19 genes often identified (Supplementary Table S1). The three most commonly rearranged in cancer [9 patients (2.3%) were tested with prior altered genes were TP53 (44%), followed by KRAS (16%) and version of the panel comprising 182 cancer–related genes]. PIK3CA (12%; Fig. 1A). Most of the alterations identified were Amplification was noted when there was 8-fold change in mutations (62%) or amplifications (29%; Fig. 1B). copy number. Four hundred and twenty patients had at least one molecular alteration (96%). The median number of alterations per patient was 3 (range, 0–16). Nineteen patients (4%) had no alterations Definition of actionability and 48 patients (11%) had only one alteration. Interestingly, the An actionable alteration was defined as an alteration that was majority of our patients had different molecular profiles. Indeed, either the direct target or a pathway component that could be only 7 patients (1.6%) had precisely the same molecular profile, targeted by at least one approved or investigational drug. For when looking at both the gene and the location/type of alteration. consistency purposes, the interpretations provided on the next- If only the gene was taken into consideration (and not the generation sequencing reports were utilized as a basis for action- location/type of alteration), 49 patients (11.2%) had the same ability determination. profile.

Data extraction Actionable aberrations Demographic information such as gender, age at diagnosis, In total, 393 patients of 439 (90%) had at least one potentially race, and clinical information such as cancer histology, pres- actionable alteration identified, for which at least one or several ence of metastasis at diagnosis, and presence of metastatic experimental drugs in clinical trial were usable (Fig. 2). Among the disease at the time of the biopsy used for molecular testing 10% of patients who had no actionable alterations, 4% had no were extracted from patients' electronic medical charts. The reportable genetic alteration found, and 6% had one or more biopsy site used for molecular testing was also recorded. From alterations, but none actionable. The median number of poten- the molecular testing reports, the following information were tially actionable alteration per patient was 2 (range, 0–8; Fig. 3A). extracted: number of total alterations, number of actionable In considering the drug options for actionable aberrations, the alterations, and, more specifically, the number of alterations total number of patients who had an aberration targetable by an with an approved drug available in the disease (on-label use), approved drug was 307 [70%; 296 patients (67%) had at least one the number of alterations with an approved drug in another drug that was approved for another disease (off-label use), and 89 disease (off-label use), and the number of alterations with patients (20%) had at least one or more approved agents in their experimental drug(s) available (clinical trials). Table 1. Patient characteristics Characteristics Total patients (N ¼ 439) Statistical analysis Age at diagnosis, years Most of the analysis was descriptive in nature. When appro- Median (CI 95%) 54.3 (52.6–55.8) priate, linear or binary logistic regression analyses were per- Gender formed; coefficients and 95% confidence intervals (95% CI) were Women 248 (57%) reported. Spearman's rho coefficients were computed to assess the Men 191 (43%) correlation between two continuous variables. The sample size Race was determined by the available patients with genetic testing Caucasian 320 (73%) Other 59 (13%) information. The medians and 95% CI were reported for contin- Asian 29 (6%) uous variables. Demographic and genetic characteristics were African American 12 (3%) compared between groups. All statistical analyses were performed Unknown 11 (3%) by MS with SPSS version 22.0. Hispanic 6 (1.4%) Native American/Eskimo 2 (0.5%) Type of cancer Results Gastrointestinal 110 (25%) Patients' characteristics Breast 83 (19%) Our population comprised 439 patients who were seen at Brain 62 (14%) Gynecologic 37 (8%) the cancer center and had molecular testing performed. Patient Head and neck 34 (8%) characteristics are listed in Table 1. The median age at diag- Hematologic 36 (8%) nosis was 54 years (95% CI, 52.5–55.8 years); 57% were Melanoma 32 (7%) women. The majority of our patient population was Caucasian Lung 27 (6%) (73%), followed by other (13%) and Asian (6%). The most Othera 18 (4%) common primary tumor sites were gastrointestinal (n ¼ 110, Metastatic disease at diagnosis 70 (16%) Metastatic disease at time of biopsy 257 (58.5%) 25%), followed by breast (n ¼ 83, 19%) and brain tumors (n ¼ b n ¼ Biopsy site 62, 14%). Seventy patients ( 70, 16%) had metastatic Primary 255 (58%) disease at the time of diagnosis, and 257 patients (58.5%) Metastatic 159 (36%) hadmetastaticdiseaseatthetimeofthebiopsyusedfor Unknown 25 (6%) molecular testing. Biopsy used for molecular testing mostly aSarcoma, n ¼ 8; carcinoma, n ¼ 5; carcinoid tumor, n ¼ 1; sarcomatoid neoplasm, originated from the primary tumor (n ¼ 255, 58%) versus a n ¼ 1; nerve sheath tumor, n ¼ 1; unknown origin, n ¼ 2. b metastatic site. Used for molecular testing.

OF2 Mol Cancer Ther; 14(6) June 2015 Molecular Cancer Therapeutics

Genomic Biomarker Actionability in 439 Patients with Cancer

Patients with the A Patients harboring the alteration (%) Alterations (n) alteration (n) TP53 44 195 218 KRAS 16 71 72 PIK3CA 12 52 54 CDKN2A/B 11 50 50 MYC 10 42 42 PTEN 10 43 48 CDKN2A 8 34 36 EGFR 8 33 41 NF1 7 30 35 APC 6 28 40 CCND1 6 28 28 FGF3 6 26 26 MCL1 6 26 26 SMAD4 6 26 26 ARID1A 5 23 25 BRAF 5 22 24 Figure 1. BRCA2 5 22 25 Frequency and type of molecular ERBB2 alterations identiﬁed in 439 patients 5 22 25 with cancer. A, the percentage of FGF19 5 23 23 patients harboring the most frequent FGF4 5 24 24 alterations. Only the most frequent IDH1 5 24 24 alterations, numbering at least 20, are MDM2 5 21 21 represented in this table. Some MLL2 4 18 21 patients had different alterations in RB1 4 19 20 the same gene. The full list can be found in Supplementary Table S1. B, pie chart displaying the different types of alterations found in 439 patients B Other, 2% with cancer (N ¼ 1,813 total Splice site alterations). Other category mutation, 4% comprises truncation (n ¼ 15), fusion Loss, 6% (n ¼ 10), duplication (n ¼ 8), deletion (n ¼ 5), and rearrangement (n ¼ 5).

Amplification 29% Nonsense mutation, 12%

Frameshift Missense mutation, mutation, 33% 13%

disease available (on-label use); Table 2 and Fig. 2]. There was a A subanalysis revealed that across all different malignancies positive correlation between the number of alterations found and tested, there consistently was a majority of patients that had the number of actionable alterations (Spearman's rho coefficient potentially actionable alteration (Table 2 and Fig. 3C). Of ¼ 0.808, P < 0.0001; Fig. 3B). A multiple linear regression model, note, breast cancer cases had a significantly higher median of including breast cancer versus not (breast cancer being the only molecular alterations (median of 5 alterations vs. 3; P ¼ histology type that was statistically significantly correlated with a 0.0002). A multivariable analysis, including both the breast higher number of actionable alterations in univariable analysis), cancer histology and whether or not the disease was already the number of alterations, the origin of biopsy used for the testing metastatic at the time of biopsy, confirmed that breast cancers (primary vs. metastatic), whether the disease was metastatic or not were associated with a higher number of alterations (P ¼ at the time of biopsy that was used for testing, confirmed that only 0.003); the presence of metastatic disease at the time of the number of alterations was an independent predictor of a biopsy also correlated with a higher number of alterations higher number of actionable alterations (P < 0.0001). (P ¼ 0.024).

www.aacrjournals.org Mol Cancer Ther; 14(6) June 2015 OF3

Schwaederle et al.

N = 439 patients Figure 2. Actionability in 439 patients with diverse cancers. There may be some overlapping as some patients might n = 393 (90%) patients have n = 46 (10%) patients have no actionable alteration(s) actionable alterations have approved agents on-label and off-label, as well as experimental drug options for their disease. All patients with actionable alterations had at least n = 19 (4%) have no reportable genomic one clinical trial available. In total, n = 89 (20%) patients n = 296 (67%) patients n = 393 (90%) patients alterations 307 patients (70%) had one or more have at least one have at least one have at least one n = 27 (6%) have alterations, but none are alteration for which an alteration for which an alteration for which there actionable approved drugs as option: 11 patients approved drug in the approved drug in is an experimental drug had on-label only, 218 had off-label disease is available (on- another disease is available (clinical trial) label use) available (off-label only, and 78 had both on-label and off- use) label options.

A binary logistic regression model showed that there were approved inhibitor treatments identiﬁed, whereas patients with signiﬁcantly more patients with breast cancers (P < 0.0001) and melanoma frequently had RAS/RAF/MEK inhibitors (P < 0.0001) melanoma (P ¼ 0.0002) who had approved matched targeted as treatment options. drug options in their disease (on-label use). Patients with breast Table 3 provides the potentially actionable aberrations cancer had more often mTOR (P < 0.0001) and ERBB2 (P ¼ 0.005) and examples of the FDA-approved drugs for them. The mTOR A B 120 8.0 Total alterations (median = 3 per patient) 100 Actionable alterations (median = 2 per patient) 80 6.0

60 4.0 Figure 3.

40 alterations Overall alterations and potentially 2.0 actionable alterations. A, the blue bars Number of patients 20 represent the number of patients who had the designated total number of 0 Total number of actionable 0123456789 10 11 12 13 14 15 16 0.0 alterations (actionable or not; median 0.0 5.0 10.0 15.0 20.0 ¼ 3 total alterations per patient); the Number of alterations Number of alterations orange bars represent the number of patients who had the designated Spearman rho coefficient = 0.808, P < 0.0001 number of potentially actionable alterations (median ¼ 2 actionable alterations per patient). B, scatterplot C a Other actionable, 14 depicting the linear trend of the Othera not actionable, 4 Brain not actionable, 9 number of actionable alterations in function of the total number of Skin/melanoma not alterations (Spearman's rho actionable, 3 Skin/melanoma coefﬁcient ¼ 0.808, P < 0.0001). C, pie actionable, 29 chart representing the number of Brain actionable, 53 Breast not actionable, 5 patients with and without potentially Lung actionable alterations in each Lung actionable, 25 a not actionable, 2 malignancy type. "Other" category comprised patients with sarcoma (n ¼ 8), carcinoma (n ¼ 5), carcinoid Hematologic tumor (n ¼ 1), sarcomatoid neoplasm actionable, 27 (n ¼ 1), nerve sheath tumor (n ¼ 1), and unknown origin (n ¼ 2). For patients Hematologic not Breast actionable, 78 with no actionable alterations, reasons actionable, 9 could be that they had no reportable n ¼ Head and neck alterations: brain ( 3), breast actionable, 31 (n ¼ 1), gastrointestinal (n ¼ 4), head and neck (n ¼ 2), hematologic (n ¼ 5), skin/melanoma (n ¼ 2), and other Head and neck not (n ¼ 2). Gynecologic actionable, 3 actionable, 36 Gastrointestinal actionable, 100

Gastrointestinal not actionable, 10

Gynecologic not actionable, 1

OF4 Mol Cancer Ther; 14(6) June 2015 Molecular Cancer Therapeutics

Genomic Biomarker Actionability in 439 Patients with Cancer

Table 2. Alterations and actionability in 439 patients with diverse cancer types Approved drug(s) in Experimental Patients had alteration(s), Approved drug(s) in another disease treatmentb No reportable but none actionable, the disease available, available, n (%)b (clinical trials), Histology (patients) alteration, n (%) n (%) n (%)b (on-label) (off-label) n (%) Brain (n ¼ 62) 3 (5%) 6 (10%) 1 (2%) 47 (76%) 53 (85%) Breast (n ¼ 83) 1 (1%) 4 (5%) 47 (57%) 61 (73%) 78 (94%) Gastrointestinal (n ¼ 110) 4 (4%) 6 (5%) 13 (12%) 72 (65%) 100 (91%) Gynecologic (n ¼ 37) 0 1 (3%) 2 (5%) 25 (68%) 36 (97%) Head and neck (n ¼ 34) 2 (6%) 1 (3%) 1 (3%) 23 (68%) 31 (91%) Hematologic (n ¼ 36) 5 (14%) 4 (11%) 2 (6%) 19 (53%) 27 (75%) Lung (n ¼ 27) 0 2 (7%) 8 (30%) 21 (78%) 25 (93%) Skin/melanoma (n ¼ 32) 2 (16%) 1 (3%) 15 (47%) 20 (63%) 29 (91%) Othera (n ¼ 18) 2 (11%) 2 (11%) 0 7 (39%) 14 (78%) Overall (N ¼ 439) 19 (4%) 27 (6%) 89 (20%) 296 (67%) 393 (90%) aSarcoma, n ¼ 8; carcinoma, n ¼ 5; carcinoid tumor, n ¼ 1; sarcomatoid neoplasm, n ¼ 1; nerve sheath tumor, n ¼ 1; unknown origin, n ¼ 2. bThere may be some overlapping as some patients might have approved agents on-label and off-label, as well as experimental drug options for their disease. All patients with actionable alterations had at least one clinical trial available. In total, 307 patients (70%) had one or more approved drugs as option: 11 patients had on-label only, 218 had off-label only, and 78 had both on-label and off-label options. inhibitors everolimus and temsirolimus, targeting the PI3K/Akt/ ERBB2 and mTOR inhibitors are approved in breast cancer, and mTOR pathway alterations, were the most commonly potentially BRAF/MEK inhibitors are approved for melanoma. For drugs matched drugs (38%; 151/393 patients with 1 actionable alter- approved in other disease types (off-label use), logistical pro- ation), followed by trametinib (26%, 101/393 patients with 1 blems such as insurance coverage may limit access (18). Beyond actionable alteration; Table 3). logistics, a fundamental question relates to the amount and For malignancies comprising the most patients, we were able to quality of data needed to qualify a patient for a drug, or inversely identify the more common doublets of actionable alterations. For to determine that withholding that drug might be detrimental. patients with gastrointestinal disease, KRAS alterations were Because many molecular abnormalities do not appear to segregate found with either APC or PIK3CA alterations in 14 and 6 patients, well by histology, and because many mutations and amplifica- respectively. For patients with breast cancer, the most frequent tions can be found in a subset of patients with almost any cancer actionable co-alterations were PIK3CA with either ERBB2 or (14), it appears unlikely that classical methods of approval that PTEN, and CCND1 with ERBB2 (n ¼ 5 patients each doublet). require high-level scientific evidence (usually randomized phase Patients with brain cancers frequently harbored CDKN2A/B loss III trials) will be feasible for the subset of patients with a particular with either PTEN (n ¼ 7) or EGFR alterations (n ¼ 12). molecular anomaly in each histology. On the other hand, new classifications based on molecular diagnosis or the use of geno- mically driven bucket trials that cross canonical disease bound- Discussion aries may be able to provide adequate scientific evidence to Herein, we studied the molecular alterations identified in 439 determinate actionability. As an example, a bucket trial using patients with diverse cancers (58.5% of whom had metastatic imatinib (19; reported in 2008) assessed a variety of uncommon disease at the time of biopsy) using next-generation sequencing. disorders that harbored an imatinib target (KIT, PDGFR, Bcr-Abl); We found that 96% of our patients demonstrated at least one the study led to approval for rare disorders bearing these targets, molecular alteration, and 85% had two or more alterations. The including myeloproliferative/myelodysplastic disorders, derma- most frequent alteration was TP53 mutation, found in 44% of our tofibrosarcoma protuberans, aggressive systemic mastocytosis, patients, similar to previous reports (16). and hypereosinophilic syndrome, with only 5 to 14 patients in Ninety percent of patients had an actionable aberration. Sim- each subgroup (20). ilarly, a recent report on a smaller number of patients suggested There were several limitations to our study. First, it included a that 83% (of 103 tested individuals with cancer) had actionable limited number of patients and comprised different malignancies. abnormalities (17). The median number of actionable aberra- However, the latter may suggest that the results are generalizable tions in our patients was 2 (range, 0–8). In regard to the types of across malignancies. Second, the definition of "actionable" and agents that could be used, 296 patients (67%) had abnormalities the level of evidence needed for such a determination are a matter that could be prosecuted by at least one drug that was approved of debate and in constant evolution. Finally, whether or not the for another disease (off-label use); 89 patients (20%) had abnor- patients would have responded to these drugs was not ascertained malities that could be prosecuted by at least one approved agent in in the study. Also, we were able to observe frequent actionable co- their disease (on-label use). The total number of patients who alterations, although the number of patients with each disease had an aberration targetable by an approved drug was 307 (70%). subgroup limited this analysis; still, this type of data may inform All patients who had a least one potentially actionable alteration the development of combination therapy protocols. also had one or more experimental drugs that targeted the In conclusion, our observations suggest that the vast majority anomaly as possible treatment options. However, previous expe- of patients (90%) have theoretically actionable aberrations. Fur- rience suggests that patient eligibility for these clinical trials or thermore, 70% of the patients had an aberration that can be their conduct at a limited number of distant enrolling sites might targeted by an approved drug. However, only a minority of severely limit patients' access (18). individuals (approximately 20%) had aberrations that could be Actionability by drugs that were approved in the same disease targeted by drugs approved for their type of cancer (on-label). was more common in breast cancer and melanoma. Indeed, Only 6% of patients had no aberration that was actionable by

www.aacrjournals.org Mol Cancer Ther; 14(6) June 2015 OF5

Schwaederle et al.

Table 3. Examples of potentially actionable genesa that were aberrant in at least one patient and examples of approved drugs Actionable gene Examples of approved drugs ABL1 Bosutinib, dasatinib, nilotinib, ponatinib AKT1 Temsirolimus, everolimus AKT3 Temsirolimus, everolimus ALK Crizotinib ARAF Sorafenib BCR Bosutinib, dasatinib, imatinib, nilotinib, ponatinib BRAF Regorafenib, vemurafenib, trametinib, dabrafenib CRKL Dasatinib CSF1R Sunitinib, imatinib, nilotinib DNMT3A Azacitidine, decitabine EGFR Lapatinib, cetuximab, erlotinib, geﬁtinib, panitumumab, afatinib ERBB2 Ado-trastuzumab emtansine, lapatinib, pertuzumab, trastuzumab, afatinib ERBB3 Pertuzumab, afatinib ERBB4 Erlotinib, geﬁtinib, lapatinib, regorafenib FBXW7 temsirolimus, everolimus FGFR1 Pazopanib, ponatinib, regorafenib FGFR2 Ponatinib, regorafenib, pazopanib FGFR3 Pazopanib, ponatinib FGFR4 Ponatinib FLT1 Axitinib, bevacizumab, pazopanib, regorafenib, sorafenib, sunitinib, vandetanib FLT3 Sorafenib, sunitinib GNA11 Trametinib GNAS Trametinib HGF Cabozantinib, crizotinib HRAS Trametinib IDH1 Azacitidine, decitabine IDH2 Azacitidine, decitabine JAK2 Ruxolitinib KDR Axitinib, bevacizumab, pazopanib, sorafenib, sunitinib, vandetanib, ponatinib, ramucirumab, regorafenib KIT Nilotinib, pazopanib, everolimus, dasatinib, sunitinib, imatinib, sorafenib, temsirolimus, regorafenib, ponatinib KRAS Trametinib MAP2K1 Trametinib MET Cabozantinib, crizotinib NF1 Temsirolimus, everolimus, trametinib NF2 Temsirolimus, everolimus, trametinib, lapatinib NRAS Trametinib PDGFRA Dasatinib, everolimus, imatinib, nilotinib, pazopanib, sorafenib, sunitinib, temsirolimus PIK3CA Temsirolimus, everolimus PIK3CG Temsirolimus, everolimus PIK3R1 Temsirolimus, everolimus PTCH1 Vismodegib PTEN Temsirolimus, everolimus PTPN11 Trametinib RAF1 Regorafenib, trametinib, sorafenib RET Cabozantinib, sorafenib, sunitinib, vandetanib, ponatinib RPTOR Temsirolimus, everolimus SRC Bosutinib, dasatinib STK11 Dasatinib, everolimus, temsirolimus, bosutinib TET2 Azacitidine, decitabine TOP1 Irinotecan, topotecan TSC1 Temsirolimus, everolimus TSC2 Temsirolimus, everolimus VHL Axitinib, bevacizumab, everolimus, pazopanib, sorafenib, sunitinib, temsirolimus, vandetanib aThe levels of evidence for actionability remain a matter of discussion (22); although standards have been implemented for some aberrations (e.g., BRAF in melanoma), few guidelines exist for the vast majority of genomic abnormalities.

either an approved or experimental drug in clinical trials. Inter- bodies to provide an online registry where patients with a given estingly, the molecular portfolio of almost all patients was aberration would be provided free drug as long as baseline key unique, consistent with previously reported results (21). Indeed, information is documented. The registry could then track out- only 1.6% of patients harbored identical molecular aberrations. come parameters or time on drug (as a surrogate for clinical These observations suggest that individualization of therapy is beneﬁt). These challenges, in addition to the important question likely to become increasingly important. Challenges such as the regarding the level of evidence needed to deﬁne actionability, are expense and/or restriction in use of off-label drugs, as well as the crucial issues that must be addressed in order to fully deploy strict eligibility criteria and the distance of sites for clinical trials, precision medicine in patients with cancer. New paradigms for however, limit access to matched targeted medications. One clinical research and drug access are needed at the sponsor and possible solution is for pharmaceutical sponsors or government national levels.

OF6 Mol Cancer Ther; 14(6) June 2015 Molecular Cancer Therapeutics

Genomic Biomarker Actionability in 439 Patients with Cancer

Disclosure of Potential Conflicts of Interest Writing, review, and/or revision of the manuscript: M. Schwaederle, R. Kurzrock is founder of RScueRX, has ownership interest (including G.A. Daniels, D.E. Piccioni, P.T. Fanta, R.B. Schwab, B.A. Parker, R. Kurzrock patents) in RScueRX, and is a consultant/advisory board member for Sequenom. Administrative, technical, or material support (i.e., reporting or organizing No potential conflicts of interest were disclosed by the other authors. data, constructing databases): M. Schwaederle Other (final approval): R. Kurzrock

Authors' Contributions Grant Support Conception and design: M. Schwaederle R. Kurzrock received funding by the Joan and Irwin Jacobs Fund and My Development of methodology: M. Schwaederle Answer To Cancer philanthropic fund. Acquisition of data (provided animals, acquired and managed patients, The costs of publication of this article were defrayed in part by the payment of advertisement provided facilities, etc.): M. Schwaederle, G.A. Daniels, D.E. Piccioni, page charges. This article must therefore be hereby marked in P.T. Fanta, R.B. Schwab, K.A. Shimabukuro, B.A. Parker accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M. Schwaederle, D.E. Piccioni, P.T. Fanta, B.A. Parker, Received December 11, 2014; revised February 16, 2015; accepted March 26, R. Kurzrock 2015; published OnlineFirst April 7, 2015.

References 1. Parkinson DR, Johnson BE, Sledge GW. Making personalized cancer mesylate in chronic myelogenous leukemia. N Engl J Med 2002;346: medicine a reality: challenges and opportunities in the development of 645–52. biomarkers and companion diagnostics. Clin Cancer Res 2012;18:619–24. 14. Tsimberidou A-M, Iskander NG, Hong DS, Wheler JJ, Falchook GS, Fu 2. Munoz J, Swanton C, Kurzrock R. Molecular profiling and the reclassifi- S, et al. Personalized medicine in a phase I clinical trials program: the cation of cancer: divide and conquer. Am Soc Clin Oncol Educ Book MD Anderson Cancer Center initiative. Clin Cancer Res 2012;18: 2013;2013:127–34. 6373–83. 3. Hoadley KA, Yau C, Wolf DM, Cherniack AD, Tamborero D, Ng S, et al. 15. Janku F, Berry DA, Gong J, Parsons HA, Stewart DJ, Kurzrock R. Outcomes Multiplatform analysis of 12 cancer types reveals molecular classification of phase II clinical trials with single-agent therapies in advanced/metastatic within and across tissues of origin. Cell 2014;158:929–44. non–small cell lung cancer published between 2000 and 2009. Clin Cancer 4. Demetri GD, von Mehren M, Blanke CD, Van den Abbeele AD, Eisenberg B, Res 2012;18:6356–63. Roberts PJ, et al. Efficacy and safety of imatinib mesylate in advanced 16. Said R, Hong DS, Warneke CL, Lee JJ, Wheler JJ, Janku F, et al. P53 gastrointestinal stromal tumors. N Engl J Med 2002;347:472–80. mutations in advanced cancers: clinical characteristics, outcomes, and 5. Hodi FS, Friedlander P, Corless CL, Heinrich MC, Mac Rae S, Kruse A, et al. correlation between progression-free survival and bevacizumab-contain- Major response to imatinib mesylate in KIT-mutated melanoma. J Clin ing therapy. Oncotarget 2013;4:705–14. Oncol 2008;26:2046–51. 17. Johnson DB, Dahlman KH, Knol J, Gilbert J, Puzanov I, Means-Powell J, 6. Janku F, Wheler JJ, Naing A, Falchook GS, Hong DS, Stepanek VM, et al. et al. Enabling a genetically informed approach to cancer medicine: a PIK3CA mutation H1047R is associated with response to PI3K/AKT/mTOR retrospective evaluation of the impact of comprehensive tumor profiling signaling pathway inhibitors in early-phase clinical trials. Cancer Res using a targeted next-generation sequencing panel. Oncologist 2014;9: 2013;73:276–84. 616–22. 7. Janku F, Hong DS, Fu S, Piha-Paul SA, Naing A, Falchook GS, et al. Assessing 18. Schwaederle M, Parker BA, Schwab RB, Fanta PT, Boles SG, Daniels GA, PIK3CA and PTEN in early-phase trials with PI3K/AKT/mTOR inhibitors. et al. Molecular tumor board: The University of California San Diego Cell Rep 2014;6:377–87. Moores Cancer Center Experience. Oncologist 2014;18:631–6. 8. Chapman PB, Hauschild A, Robert C, Haanen JB, Ascierto P, Larkin J, et al. 19. Heinrich MC, Joensuu H, Demetri GD, Corless CL, Apperley J, Improved survival with vemurafenib in melanoma with BRAF V600E Fletcher JA, et al. Phase II, open-label study evaluating the activity mutation. N Engl J Med 2011;364:2507–16. of imatinib in treating life-threatening malignancies known to be 9. Hofmann I, Weiss A, Elain G, Schwaederle M, Sterker D, Romanet V, et al. K- associated with imatinib-sensitive tyrosine kinases. Clin Cancer Res RAS mutant pancreatic tumors show higher sensitivity to MEK than to PI3K 2008;14:2717–25. inhibition in vivo. PLoS One 2012;7:e44146. 20. accessdata.fda.gov [database on the Internet]. U.S. Food and Drug Admin- 10. Solit DB, Garraway LA, Pratilas CA, Sawai A, Getz G, Basso A, et al. BRAF istration: imatinib package insert [updated 2015 Feb 2; cited 2015 Feb 2]. mutation predicts sensitivity to MEK inhibition. Nature 2006;439:358–62. Available from: http://www.accessdata.fda.gov/drugsatfda_docs/label/ 11. Westin JR, Kurzrock R. It's about time: lessons for solid tumors from 2006/021588s011s012s013s014s017lbl.pdf chronic myelogenous leukemia therapy. Mol Cancer Ther 2012; 21. Wheler JJ, Parker BA, Lee JJ, Atkins JT, Janku F, Tsimberidou AM, et al. 11:2549–55. Unique molecular signatures as a hallmark of patients with metastatic 12. Druker BJ, Guilhot F, O'Brien SG, Gathmann I, Kantarjian H, Gattermann breast cancer: implications for current treatment paradigms. Oncotarget N, et al. Five-year follow-up of patients receiving imatinib for chronic 2014;5:2349–54. myeloid leukemia. N Engl J Med 2006;355:2408–17. 22. Vidwans SJ, Turski ML, Janku F, Garrido-Laguna I, Munoz J, Schwab R, et al. 13. Kantarjian H, Sawyers C, Hochhaus A, Guilhot F, Schiffer C, Gambacorti- A framework for genomic biomarker actionability and its use in clinical Passerini C, et al. Hematologic and cytogenetic responses to imatinib decision making. Oncoscience 2014;1:614–23.

www.aacrjournals.org Mol Cancer Ther; 14(6) June 2015 OF7

On the Road to Precision Cancer Medicine: Analysis of Genomic Biomarker Actionability in 439 Patients

Maria Schwaederle, Gregory A. Daniels, David E. Piccioni, et al.

Mol Cancer Ther Published OnlineFirst April 7, 2015.

Updated version Access the most recent version of this article at: doi:10.1158/1535-7163.MCT-14-1061

Supplementary Access the most recent supplemental material at: Material http://mct.aacrjournals.org/content/suppl/2015/04/07/1535-7163.MCT-14-1061.DC1

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mct.aacrjournals.org/content/early/2015/05/21/1535-7163.MCT-14-1061. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.