And Cyclophosphamide in Patients With

Cancer Therapy: Clinical

A Phase I Study of Pemetrexed (ALIMTA) and Cyclophosphamide in Patients with Locally Advanced or Metastatic Breast Cancer Christian Dittrich,1Lubos Petruzelka,3 Pavel Vodvarka,4 Margit Gneist,1Filip Janku,3 Tamara Kysela,4 Allen Melemed,5 Jane Latz,5 Lorinda Simms,5 and Kurt Krejcy2

Abstract Purpose: Determine the maximum tolerated dose (MTD) of pemetrexed and cyclophosphamide combination therapy for patients with locally advanced or metastatic breast cancer. Experimental Design: Patients with locally advanced or metastatic breast cancer and WHO performance status 0 to 2 were eligible. Pemetrexed (range, 400-2,400 mg/m2) was administered on day 1 of a 21-day schedule followed by cyclophosphamide (range, 400-800 mg/m2). Folic acid and vitamin B12 supplementation began 1to 2 weeks before the first pemetrexed dose. Results: Fifty-seven pretreated patients were enrolled and received 342 cycles (median, 4 cycles; range, 1-26) through 14 dose levels. The MTD of pemetrexed was 2,400 mg/m2 (combined with cyclophosphamide, 600 mg/m2) with dose-limiting toxicities of grade 4 neutropenia with grade 4 infection and grade 3 diarrhea. Other grade 3 or 4 toxicities included (febrile) neutropenia, thrombocytopenia, anemia, elevated alanine aminotransferase/aspartate aminotransferase, and diarrhea. Pharmacokinetic analysis indicated that pemetrexed clearance and central volume of distribution were 40% lower than single-agent reference data, yielding a 68% increase in total systemic exposure and a 56% increase in maximal plasma concentration. Among the 50 patients evaluable for efficacy, 13 (26%) patients had a partial response and 17 (34%) patients had stable disease. Conclusions: Pemetrexed was generally well tolerated. The observed toxicities were consistent with the known toxicity profiles of pemetrexed and cyclophosphamide. Considering the MTD and the toxicity and efficacy results in this and prior studies, a low (600 mg/m2)andahigh (1,800 mg/m2) dose of pemetrexed with cyclophosphamide (600 mg/m2) will be evaluated in the consecutive prospective randomized phase II study.

Breast cancer is the most common malignancy to affect women multiple tumor types (5). The substance has a manageable in developed countries. Approximately 36% of women with toxicity profile that includes dose-limiting neutropenia and breast cancer are expected to develop fatal metastatic disease thrombocytopenia as well as lower grades of reversible (1). In the metastatic setting, anthracyclines and/or taxanes hepatotoxicity and gastrointestinal toxicity. Pemetrexed- have become standard therapeutics for first-line treatment if associated toxicity is increased with a high level of plasma they have not already been used in the adjuvant setting. After homocysteine, which in turn has been associated with nutri- these agents fail, non-anthracycline-based and non-taxane- tional folate deficiency (6). Folate and vitamin B12 supplemen- based chemotherapy regimens must be developed as the next tation have been found to reduce plasma homocysteine and, logical treatment option for these patients. thus, pemetrexed-associated toxicities, allowing dose escalation Pemetrexed, an inhibitor of thymidylate synthase, dihydro- without apparently compromising antitumor activity (7, 8). folate reductase, and glycinamide ribonucleotide formyl The antitumor activity of pemetrexed against breast cancer transferase (2–4), has shown broad-spectrum activity in has been shown in various preclinical models (9, 10). In addition, pemetrexed is expected to have antitumor activity against breast cancer because other thymidylate synthase inhibitors, such as 5-fluorouracil (5-FU), and other dihydrofo- Authors’ Affiliations: 1Ludwig Boltzmann Institute for Applied Cancer Research and Applied Cancer Research-Institution forTranslational Research Vienna, Kaiser late reductase inhibitors, such as methotrexate, are active Franz Josef-Spital; 2Eli Lilly and Company,Vienna, Austria; 3Charles University, substances and well established in the treatment of breast Prague, Czech Republic; 4Faculty Hospital and Policlinic Ostrava, Ostrava, Czech cancer. In addition, because thymidylate synthase and dihy- 5 Republic; and Eli Lilly and Company, Indianapolis, Indiana drofolate reductase are not targeted by anthracyclines or Received 12/27/05; revised 5/16/06; accepted 6/27/06. The costs of publication of this article were defrayed in part by the payment of page taxanes, cross-resistance of pemetrexed with those agents would charges. This article must therefore be hereby marked advertisement in accordance not be anticipated, and therefore, pemetrexed may be expected with 18 U.S.C. Section 1734 solely to indicate this fact. to inhibit growth of breast cancer that relapses following Requests for reprints: Christian Dittrich, 3rd Medical Department, Center for exposure to those two classes of agents. Oncology and Hematology, Kaiser Franz Josef-Spital, Kundratstrasse 3, A-1100 Single-agent pemetrexed has shown activity in patients with Vienna, Austria. Phone: 43-1-60191-2303; E-mail: [email protected]. F 2006 American Association for Cancer Research. advanced breast cancer; most of whom were previously treated doi:10.1158/1078-0432.CCR-05-2829 in the metastatic and/or adjuvant settings, yielding response

www.aacrjournals.org 7071 Clin Cancer Res 2006;12(23) December 1, 2006 Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2006 American Association for Cancer Research. Cancer Therapy: Clinical rates of 28% (11), 26% (12), or 20% (13, 14). In a study of including active infection, second primary malignancy, and symptom- heavily pretreated patients with metastatic breast cancer who atic central nervous system metastases; inability to take folic acid or had received three or more regimens, including prior anthracy- vitamin B12 supplementation or dexamethasone (or an equivalent clines, taxanes, and capecitabine, pemetrexed showed only corticosteroid); and inability to stop treatment with nonsteroidal anti- inflammatory drugs. In addition, patients were excluded from the study modest antitumor activity (8% response rate) but was well for pregnancy, breast-feeding, and/or childbearing potential without tolerated and efficacious in stabilizing disease and providing adequate contraception. symptom palliation (15). In a single study testing pemetrexed All patients were required to provide written informed consent in previously untreated breast cancer, a response rate of 35% before study entry. This study was conducted according to the was reported (16). Given these single-agent results, the guidelines of each participating institution and was in compliance pemetrexed and cyclophosphamide combination was expected with Good Clinical Practice and the tenets of the Declaration of to offer additional efficacy benefits. Helsinki. Among the many cytotoxic agents registered for breast cancer, Treatment and dose escalation scheme. Beginning at the low dose of 2 2 cyclophosphamide is one of the most widely used alkylating 400 mg/m and increasing to a maximum of 2,400 mg/m , pemetrexed agents with antineoplastic activity (17). Single-agent cyclo- was administered i.v. over 10 minutes (and up to 30 minutes for doses z1,800 mg/m2 to avoid hot flashes observed in some patients) on day phosphamide is associated with objective response rates 1 of a 21-day schedule. Cyclophosphamide was administered i.v. over ranging from 10% to 62% in patients with breast cancer 2 f 30 minutes starting at 400 mg/m and increased to a maximum of (18, 19) and with an average response rate of 34% (20). The 800 mg/m2 on day 1 of a 21-day cycle beginning f10 minutes after the cyclophosphamide-based combination regimen cyclophospha- completion of the pemetrexed infusion. To reduce pemetrexed- mide, methotrexate, and 5-FU is well established in breast associated toxicities, patients were required to take oral folic acid and cancer both as adjuvant therapy (21, 22) and as therapy for to get i.m. injections of vitamin B12. Patients were to take oral folic acid advanced disease with an overall response rate of f50% (23). (350-1,000 Ag) daily beginning approximately 1 to 2 weeks before the To determine whether pemetrexed-containing combination start of cycle 1 and continuing daily until 3 weeks after the last dose of therapy might provide an effective alternative to cyclophos- study therapy. Patients were also required to receive vitamin B12 A phamide, methotrexate, and 5-FU, we proposed replacing the injections (1,000 g) approximately 1 to 2 weeks before the start of thymidylate synthase inhibitor 5-FU and the dihydrofolate cycle 1, and injections were repeated approximately every 9 weeks until 3 weeks after the last dose of study therapy. Prophylactic dexametha- reductase inhibitor methotrexate with the multitargeted anti- sone (4 mg twice daily) was administered on the day before, the day of, folate pemetrexed. Thus, we designed a phase I/II study to and the day after each dose of pemetrexed to prevent rash. Patients evaluate the combination of pemetrexed and cyclophospha- continued treatment until disease progression; however, a patient could mide in patients with locally advanced or metastatic breast have discontinued earlier in case of the following: unacceptable toxicity; cancer. The phase Istudy results are reported here. pregnancy or failure to use adequate birth control; failure to stop using The primary objective of this multicenter phase Istudy was to nonsteroidal anti-inflammatory drugs 2 to 5 days before, during, and determine the maximum tolerated dose (MTD) of pemetrexed 2 days after therapy administration; in case of patient request; or at the and cyclophosphamide combination therapy. Secondary objec- discretion of the investigator, the patient’s physician, or Eli Lilly and tives included the evaluation of quantitative and qualitative Company. The decision to escalate the doses of pemetrexed and cyclophospha- toxicities, the assessment of the effect of a patient’s baseline 2 nutritional status on safety as measured by homocysteine levels, mide from the respective starting doses of 400 mg/m to the next dose levels was based on toxicities observed during cycle 1. If a patient did a pharmacokinetic assessment of systemic exposure to the not have a dose-limiting toxicity (DLT) at cycle 1 at a given dose level, chemotherapy based on a limited blood sampling scheme then two additional patients were treated at that dose level before during dose escalation, and a pharmacodynamic assessment of patients were allocated to the next higher dose level. If a DLT was the relationship between systemic drug exposure and toxicities. observed in one of three patients at cycle 1, then a maximum of three additional patients were entered at that dose level. The same pattern of enrollment was followed for all dose levels. The MTD was reached if the Materials and Methods same DLT occurred in two or more of six patients at a given dose level; the dose level recommended for phase II studies was defined as the next Patient selection. Major eligibility criteria included histologic or lower dose level than the MTD. cytologic diagnosis of locally advanced or metastatic breast cancer, A DLT was defined as follows: (a) Common Toxicity Criteria grade performance status 0 to 2 on the WHO scale, and an estimated life 4 neutropenia lasting z5 days, (b) febrile neutropenia, (c) grade expectancy of at least 12 weeks. Patients were not allowed to receive 4 thrombocytopenia or grade 3 thrombocytopenia with bleeding other forms of treatment for at least 4 weeks before study enrollment requiring platelet transfusion, and (d) grade z3 nonhematologic with the exception of radiotherapy (to <25% of the bone marrow) and toxicity, excluding alopecia, nausea, and vomiting manageable with endocrine therapy (except luteinizing hormone-releasing hormone antiemetic therapy and excluding isolated Common Toxicity Criteria analogues). Patients were required to have adequate bone marrow grade 3 ALT or AST that returned to grade 1 or pretreatment values function (absolute neutrophil count z1.5 109/L, platelet count within 3 weeks. z100 109/L, and hemoglobin z9 g/dL), adequate liver function Baseline and treatment assessments. Assessments at baseline includ- [bilirubin V1.5 times the upper limit of normal, alkaline phosphatase, ed a complete medical history and physical examination, evaluation of aspartate aminotransferase (AST), and alanine aminotransferase (ALT) performance status (WHO), tumor measurement of palpable or visible V3.0 times the upper limit of normal], and normal renal function lesions, hematologic and blood chemistry assessment, and vitamin (calculated creatinine clearance z45 mL/min). metabolite panel. Hematology and blood chemistry were monitored Patients were excluded from the study for the following reasons: weekly throughout the study and within 4 days before each cycle. chemotherapy within 4 weeks before study enrollment, treatment with Hematologic and nonhematologic data were assessed according to the any drug within the last 30 days that had not received regulatory National Cancer Institute Common Toxicity Criteria rating scale version approval at the time of study entry; previous participation in a clinical 2.0 (24). Patients were considered evaluable for MTD and safety if they study involving pemetrexed; serious concomitant systemic disorders, received at least one dose of either study drug.

Clin Cancer Res 2006;12(23) December 1, 2006 7072 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2006 American Association for Cancer Research. Pemetrexed and Cyclophosphamide in Breast Cancer

For determination of plasma pemetrexed concentration, two blood 1 patient experienced a hypersensitivity to the folic acid or samples were collected from each patient. Samples were taken at the vitamin supplementation required by protocol. Thus, a total of end of the 10-minute pemetrexed infusion (30-minute infusion for 2 57 patients finally received treatment according to the protocol dosages from 1,800 to 2,400 mg/m ) and 6 to 8 hours later. Plasma between January 2001 and September 2004 (Table 1). The samples obtained during this study were analyzed for pemetrexed at patients ranged in age from 32 to 81 years (median, 55 years), Taylor Technology, Inc. (Princeton, NJ). Heparinized human plasma samples were analyzed for pemetrexed using a high-performance liquid with a median WHO performance status of 0 (range, 0-2). More chromatography electrospray ionization tandem mass spectrometry than three quarters of patients had at least two sites of method over the concentration range of 10.0 to 200,000.0 ng/mL (25). metastatic disease at baseline. Liver, lung, and bone metastases Efficacy was not a primary or secondary end point of the phase I were each present in about half of the patients. In general, the portion of this study; however, lesions were followed with appropriate patients were heavily pretreated; two thirds had already radiological imaging every other cycle to exclude progressive disease. received chemotherapy in the metastatic setting, with about Antitumor activity was documented using Southwest Oncology Group half of them receiving two lines or more. The majority of the criteria (26). A complete response was defined as complete absence of patients had received a prior anthracycline-, cyclophospha- measurable, evaluable, or nonevaluable disease with no new lesions mide-, and/or taxane-based therapy. and no disease-related symptoms. A partial response was defined as Therapy administration. The patients received a total of 342 z50% reduction from baseline of the sum of the products of perpendicular diameters of all measurable lesions, with no progression cycles of therapy through 14 dose levels; the median number of of evaluable disease and no new lesions. Any complete or partial cycles administered per patient was 4 (range, 1-26 courses). response required confirmation 4 weeks later. Tumor progression was There were two pemetrexed dose reductions due to diarrhea defined as a 50% increase or an increase of 10 cm2 (whichever was (both cycle 2, dose levels 1 and 14) and two pemetrexed and smaller) in the sum of products of all measurable lesions over the cyclophosphamide dose reductions due to neutropenia (dose smallest sum observed or a clear worsening of any evaluable disease or level 7, cycle 4 and dose level 10, cycle 7). There were no the reappearance of any lesion that had disappeared or the appearance pemetrexed or cyclophosphamide dose omissions. Over the of any new lesion/site. Stable disease was that which did not qualify for complete response, partial response, or progression. Patients were considered evaluable for tumor response (during treatment and N poststudy follow-up) if they had measurable disease and had received Table 1. Patient characteristics ( = 57) at least one complete cycle of both study drugs. Statistical and pharmacokinetic analyses. Plasma pemetrexed con- Median age (range), y 55 (32, 81) n centration versus time data from the current study (pemetrexed WHO performance scale, (%) 0 42 (73.7) administered in combination with cyclophosphamide) was compared 1 13 (22.8) with reference data of 1,388 pemetrexed concentration results from 2 2 (3.5) 252 patients in 10 phase II single-agent pemetrexed studies (25). A No. metastatic sites, n (%) previously established open two-compartment population pharmaco- 1 12 (21.1) kinetic model parameterized in terms of clearance (CL), central volume 2 18 (31.6)

of distribution (V1), intercompartmental clearance (Q), and peripheral 3 17 (29.8) z4 10 (17.5) volume of distribution (V2) with calculated creatinine clearance Location of metastases, n (%) estimated by the Cockcroft-Gault method (CrCLCG,std) as a covariate with respect to CL and body surface area as a covariate with respect to Liver 29 (50.9) Lung 27 (47.4) V were included in the model (25). Potential differences in pemetrexed 1 Bone 32 (56.1) pharmacokinetics between the two populations were identified as those Estrogen/progesterone receptors, n (%) which resulted in a significant decrease in the minimum value of the +/+ 25 (43.9) NONMEM objective function (z10.828 points for 1 degree of freedom; +/ 6 (10.5) P < 0.001) on inclusion of a dichotomous variable as a covariate / 19 (33.3) differentiating the current study population from the reference U/U 3 (5.3) population. Plots of individual empirical Bayesian estimates of model +/N 2 (3.5) variables (e.g., CL and V ; obtained using the POSTHOC option in +/U 1 (1.8) 1 NONMEM) by dose were used to identify potential dose-related U/ 1 (1.8) Prior therapy, n (%) alterations in pemetrexed pharmacokinetics. Surgery 53 (93.0) To examine possible pharmacodynamic relationships, pemetrexed Chemotherapy 52 (91.2) total systemic exposure [area under concentration curve (AUC)] was Anthracycline 51 (89.5) estimated based on individual empirical Bayesian estimates of CL. The Cyclophosphamide 45 (78.9) relationship between pemetrexed AUC, cyclophosphamide total dose Taxane 39 (68.4) administered, and occurrence of any grade 3 or 4 events and grade 3 or Trastuzumab 8 (14.0) 4 neutropenia was summarized graphically (scatter plots and associated Adjuvant setting 36 (63.2) box plots) to identify any exposure-toxicity/response relationship. Neoadjuvant setting 9 (15.8) Locally advanced setting 1 (1.8) Metastatic setting 39 (68.4) Results 1 line of therapy 8 (14.0) 2 lines of therapy 18 (31.6) z Patient demographics. A total of 65 patients, all female, 3 lines of therapy 13 (22.8) Radiotherapy 47 (82.5) signed informed consent and were registered into the study at Hormonaltherapy 39 (68.4) one of three study centers in Austria and the Czech Republic. Immunotherapy 2 (8.5) Of these 65 patients, 8 patients were not entered in the study: 5 patients failed to meet inclusion criteria, 1 patient had a Abbreviations: N, not assessed; U, unknown. personal conflict, 1 patient revised her previous decision, and

www.aacrjournals.org 7073 Clin Cancer Res 2006;12(23) December 1, 2006 Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2006 American Association for Cancer Research. Cancer Therapy: Clinical

342 cycles of therapy, there were 48 cycles delayed (14.0%) for experienced a DLT grade 3 diarrhea at cycle 1; this patient also clinically significant reasons; these included 25 delays due to had grade 3 vomiting. Because of the two DLTs observed at this neutropenia (52.1% of delays). Although neutropenia caused dose level (grade 4 neutropenia and associated infection in one some dose reductions and delays, there was no evidence of patient, grade 3 diarrhea in another patient), the possibly study cumulative hematologic toxicity. Other causes of cycle delays drug-related death, the mucosal-related toxicities experienced by were transient increases in ALT or AST (13 incidents, 27.1% of two of the patients, and the substantial toxicities experienced delays), transient decreases in creatinine clearance (4 incidents, at the previous dose level, enrollment was closed, and dose 8.3% of delays), insufficient prophylactic dexamethasone and level 14, pemetrexed at 2,400 mg/m2 and cyclophosphamide nasopharyngitis (each 2 incidents, each 4.2% of delays), and at 600 mg/m2, was designated the MTD. The dose formally to infection and cough (each 1 incident, each 2.1% of delays). be recommended for future phase II studies, defined prospec- Patient compliance with daily oral folic acid supplementation tively as the next lower dose level, was therefore pemetrexed at was reported as 100% throughout the study. 2,100 mg/m2 and cyclophosphamide at 600 mg/m2. MTD and toxicity. As detailed in Table 2, DLTs occurred in six As previously mentioned, there is evidence that high baseline patients at dose levels 1, 6, 13, and 14. At dose level 14, the first plasma homocysteine levels are associated with increased patient experienced grade 4 neutropenia associated with grade 4 toxicity in nonsupplemented patients (7). Therefore, pre-folic infection at cycle 1, as DLT. This patient subsequently died due acid supplementation homocysteine levels were plotted against to streptococcal sepsis on day 12 of cycle 1, an event considered chemotherapy dose levels to assess whether there was any possibly related to study drug. In addition, the patient evidence of a difference in homocysteine levels across chemo- experienced grade 1 stomatitis and vomiting during cycle 1. therapy dose levels (Fig. 1). Although a slight trend to lower Because of the DLT, an additional cohort of up to three patients homocysteine levels (P = not significant) was observed at higher was to be enrolled. Another patient at this dose level also dose levels, this trend seems insufficient to affect the MTD.

Table 2. Summary of DLTs and other NationalCancer Institute Common Toxicity Criteria grade 3 and grade 4 toxicities by dose level

Dose P/C (mg/m2) Patients with DLTs Other maximum grade 3and 4 toxicities * level DLTs/enrolled patients 1 400/400 1/6 G3 diarrhea G3 ALT G3 anemia G3 ALT G3 fatigue 2 500/400 0/3 None G3 AST 3 500/600 0/3 None G4 neutropenia G3 neutropenia G3 ALT 4 500/800 0/7c None G4 febrile neutropenia, G3 neutropenia G3 ALT, G3 abdominalpain 5 600/800 0/3 None G4 neutropenia, G3 AST G3 AST 6 800/600 1/6 G4 AST/ALT, G3 bilirubin G3 neutropenia G3 ALT 7 900/600 0/3 None G4 neutropenia, G3 ALT G4 neutropenia G3 hypokalemia 8 1,000/600 0/3 None G3 neutropenia G3 anemia 9 1,100/600 0/4b None G4 neutropenia G4 neutropenia, G3 ALT 10 1,300/600 0/3 None G4 neutropenia, G3 weight loss 11 1,500/600 0/3 None G4 neutropenia, G3 leukopenia, G3 infection 12 1,800/600 0/3 None G3 neutropenia 13 2,100/600 2/6 G4 thrombocytopenia G4 leukopenia, G3 anemia G3 rash/desquamation G3 neutropenia 14 2,400/600 2/4x G4 neutropenia with G4 infection G3 fatigue G3 diarrhea G3 vomiting

Abbreviations: C, cyclophosphamide; G3, grade 3; G4, grade 4; P, pemetrexed. *Toxicities appearing on the same line occurred in the same patient. cBecause one patient died from pneumonia 7 days after starting study drug therapy, an additional patient was enrolled. Of this four patient cohort, a second patient died (respiratory and heart failure, cycle 2). Three additional patients were subsequently enrolled, bringing the total to seven patients. Both deaths were unrelated to study drug. bBecause one patient discontinued the study due to progressive disease 5 days after starting study drug therapy, an additionalpatient was enrolled. xBecause of the DLTs and subsequent possibly study drug-related death (streptococcal sepsis) observed in the first patient enrolled, up to three additional patients were to be enrolled. However, the fourth patient experienced DLTs similar to the grade 1 mucosal toxicities reported in the first patient, so enrollment was stopped and this dose level was said to be the MTD.

Clin Cancer Res 2006;12(23) December 1, 2006 7074 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2006 American Association for Cancer Research. Pemetrexed and Cyclophosphamide in Breast Cancer

from 400 to 2,400 mg/m2 as evidenced by the lack of a dose- related trend in CL, V1,orVSS (Fig. 3). Therefore, Cmax and AUC are also expected to be dose proportional over this dose range. Furthermore, pemetrexed pharmacokinetics seemed to be independent of cyclophosphamide over a cyclophosphamide dose range of 400 to 800 mg/m2 (data not shown). Pharmacodynamic analysis did not identify a significant association between higher pemetrexed AUCs and the occurrence of grade 3 or 4 toxicities [Wilcoxon Ps for the com- parisons, P = 0.59 (all toxicities) and P = 0.18 (neutropenia)]. Antitumor activity. Of the 57 enrolled patients, 50 (87.7%) were evaluable for response. Of the seven patients not evaluable for response, six were not followed for tumor evaluation at the discretion of the investigator because they did not have measurable lesions at baseline. One patient was not evaluable because the baseline tumor assessment was incomplete. Of the 50 evaluable patients, there were no complete responses and there were 13 (26%) confirmed partial responses at the Fig. 1. Plot of baseline homocysteine level (pre-folic acid supplementation) and following dose levels: DL1:1, DL3:2, DL4:2, DL5:1, DL6:1, fitted linear regression line of pemetrexed and cyclophosphamide by dose level. DL7:1, DL9:1, DL10:1, DL12:2, and DL13:1. Responses Each patient’s pre-folic acid supplementation (baseline) homocysteine level was plotted against dose level (1-14), and the linear regression line was fitted to the data. occurred at most dose levels, with 8 of the 13 responders occurring at dose levels 1 to 7 (n = 31; pemetrexed dose range, 400-900 mg/m2). Of the 37 nonresponders, 17 had SD at the Grade 3 or 4 neutropenia was the most frequent hematologic following dose levels: DL1:3, DL2:1, DL5:2, DL6:3, DL8:3, toxicity (a total of 16 incidences) occurring in at least one DL9:1, DL10:1, DL11:1, and DL13:2. patient in every cohort from dose level 3 onwards (Table 2). Eight patients received RBC transfusions for grade 2 or 3 Discussion anemia, and one patient received platelet transfusions prophy- lacticly for grade 4 thrombocytopenia to prevent bleeding. To identify non-anthracycline-based and non-taxane-based Grade 3 and 4 nonhematologic toxicities were infrequent chemotherapy alternatives for patients with metastatic breast (Table 2). Grade 3 or 4 AST and ALT occurred most often in cancer, the combination of pemetrexed and cyclophosphamide four (7.0%) and eight (14.0%) patients, respectively. The one was examined in a phase Idose escalation study. The primary patient who had both grade 4 AST and ALT also developed objective of the study was the determination of the MTD of grade 3 bilirubin elevation. pemetrexed and cyclophosphamide combination therapy. The Three patients were discontinued from the study (dose levels MTD was determined to be pemetrexed at 2,400 mg/m2 and 5, 11, and 13) due to hematologic toxicities not defined as cyclophosphamide at 600 mg/m2. DLTs (two patients, neutropenia; one patient, leukopenia). The designation of these dosages to represent the MTD for There were no discontinuations due to study drug-related the drug combination under investigation was done despite the nonhematologic toxicities. Pharmacokinetic and pharmacodynamic evaluations. A total of 96 pemetrexed determinations from 51 patients was evaluable for inclusion in the pharmacokinetic evaluations. As depicted in Fig. 2, dose-normalized concentrations of pemetrexed in plasma were generally within the range of the reference data (25), indicating that there were no overt alterations in pemetrexed pharmacokinetics in the current population. However, the dose-normalized concentrations seemed to be somewhat higher on average than those from patients given single-agent pemetrexed, which reflects a 40% decrease in pemetrexed clearance (CL) and central volume of distribution (V1; P < 0.001 for both), as summarized in Table 3. These decreases in pemetrexed clearance and volume of distribution correspond to a 68% increase in AUC and a 56% increase in Cmax for the patients enrolled in the current study relative to single-agent administration of the same pemetrexed dose. The alterations in CL and V1 are also reflected in a modest increase in half-life in this study population compared with the reference population of 252 patients (3.93 versus 3.38 hours). As previously observed over a dose range of 126 to 838 mg/m2 Fig. 2. Pemetrexed concentration in plasma versus time. Main graph plots, pemetrexed concentration versus time out to 8 hours for the study and reference (25), pemetrexed pharmacokinetics also seem to be independent populations; subfigure plots, pemetrexed concentration versus time out to 42 hours of pemetrexed dose within the current trial where doses ranged for the reference population.

www.aacrjournals.org 7075 Clin Cancer Res 2006;12(23) December 1, 2006 Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2006 American Association for Cancer Research. Cancer Therapy: Clinical

Table 3. Pemetrexed population pharmacokinetic model variables and comparison of the model estimated pharmacokinetic variables for the reference population (pemetrexed administered as a single agent) and the current population (pemetrexed administered in combination with cyclophosphamide)

Pemetrexed population pharmacokinetic model variables Variable estimate Between-patient variability (% SEE) (% SEE) Clearance (CL) TVCL, base variable for CL (mL/min) 44.0 (12.9) 19.1% (12.6)

Q1, variable for effect of CrCLCG,std on CL 44.2 (12.0) Q2, variable comparing CL for current population 0.595 (5.21) with that for the reference population

Centralvolumeof distribution (V 1) TVV1, base variable for V1 (L) 6.06 (8.86) 16.6% (22.0) Q3, variable for effect of BSA on V1 1.15 (13.1) Q4, variable comparing V1 for current population 0.636 (5.80) with that for the reference population Intercompartmentalclearance(Q) Variable for Q (mL/min) 19.3 (16.0) —*

Peripheralvolumeof distribution (V 2) Variable for V2 (L) 4.04 (9.78) 24.6% (19.6) Residualerror (proportional) 28.7% (7.01)

Model estimated pharmacokinetic variablesc Reference data: single-agent Pemetrexed/ pemetrexed cyclophosphamide

CL (mL/min) 88.2 52.5

V1 (L) 11.5 7.31 b C max (Ag/mL) 72.3 113 AUC (h Ag/mL)b 165 278

t 1/2 (h) 3.38 3.93

NOTE: CL for the reference population = TVCL + Q1 (CrCLCG,std / 90). CL for the pemetrexed/cyclophosphamide study population = CL for the reference population Q2.V1 for the reference population = TVV1 BSA^Q3.V1 for the pemetrexed/cyclophosphamide study population = V1 for the reference population Q4. Abbreviations: BSA, body surface area; Cmax, maximum concentration; CrCLCG,std, calculated creatinine clearance estimated by the Cockcroft- Gault method; t1/2, terminal elimination half-life; SEE, standard error of the estimate. *Not estimated; the model did not support inclusion of between-patient variability with respect to Q. c Estimates are the population estimates (i.e., the ‘‘typical value’’ i.e., the central tendency corresponding to CrCLCG,std of 90 mL/min and BSA of 1.75 m2). bEstimates are the population estimates corresponding to a 500 mg/m2 dose. fact that the two patients’ DLTs were not the same as required patients at dose levels 8 to 14 showed stabilization of their by the protocol definition of MTD. Nevertheless, the decision disease. These results obviously do not suggest the presence of a was made because of the severity of toxicity, particularly the dose-response relationship for this drug combination. possibly study drug-related death in one patient. In addition, The high MTD could not be explained by significantly lower two patients developed mucosal-related toxicities: one in the initial homocysteine levels in the patients receiving the higher form of grade 3 diarrhea, representing a DLT, and the other in pemetrexed doses (Fig. 1) nor could it be explained by low the form of grade 1 mucositis in the patient who died, although homocysteine levels in the entire study population. Presupple- not representing a DLT. mentation homocysteine levels in this study cohort were The MTD of pemetrexed in combination with cyclophos- comparable with those of patients with other cancers participat- phamide was much higher than had been found in single-agent ing in phase Iand IItrials (31). Presupplementation homo- pemetrexed studies conducted in Japan (27) and in the United cysteine will be analyzed in the phase II study, and data from States (28) and higher than in all dual-agent phase Istudies a larger number of patients will be available to assess the (29). This contrasts with observations from studies of relationship between homocysteine level and pemetrexed dose. cyclophosphamide-methotrexate-5-FU and cyclophospha- A secondary objective of this study was qualitative and quan- mide-doxorubicin-5-FU (combination) therapy (30). The com- titative assessment of the toxicities of the combination therapy. mon toxicity (myelosuppression), which results when Toxicities associated with pemetrexed and cyclophosphamide in cyclophosphamide is combined with these agents, requires this study were found to be consistent with the known toxicity dose reductions of all the agents with little resultant increase in profiles of these agents. The combination was generally well overall dose intensity and therefore only a relatively moderate tolerated, including no apparent cumulative hematologic toxic- increase in (therapeutic) effect. ity; this suggests that future studies could administer a sufficient The assessment of clinical activity is not an aim of a phase I number of cycles with manageable toxicity. study, but nevertheless, it was systematically recorded. Thus, The lower dose levels had fewer nonlaboratory toxicities 8 of 31 patients at dose levels 1 to 7 and 5 of 26 patients at dose and a lower incidence of hematologic toxicities, with the levels 8 to 14 responded objectively with a partial remission. In exception of neutropenia, which was distributed generally addition, 9 of 31 patients at dose levels 1 to 7 and 8 of 26 evenly across all dose levels except the lowest 2. Interestingly,

Clin Cancer Res 2006;12(23) December 1, 2006 7076 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2006 American Association for Cancer Research. Pemetrexed and Cyclophosphamide in Breast Cancer nonhematologic toxicities, primarily elevated AST and ALT, Istudy. As already detailed above, the MTD (2,400 mg/m 2) was occurred more frequently at the lower dose levels. nearly twice as high as the MTD observed with single-agent Population pharmacokinetic analyses using two time points pemetrexed (27, 28). Although no overt exposure-effect per patient from this study compared with reference single- relationship between pemetrexed AUC or cyclophosphamide agent data from 10 phase II studies (25) suggest that dose and the occurrence of grade 3 or 4 toxicities was observed, pemetrexed clearance and central volume of distribution were there was a trend toward more frequent grade 3 or 4 toxicities f40% lower in the current study. The decrease in pemetrexed with higher pemetrexed AUC (P = 0.59 for all toxicities; P = clearance and central volume of distribution is reflected in 0.18 for neutropenia). Taken together, these observations higher concentrations of pemetrexed. Increased pemetrexed support the decision to study a low and a high pemetrexed concentrations should be associated with an increase in toxicity dose during the phase II study to determine definitively if a for the combination of pemetrexed and cyclophosphamide. dose-effect relationship exists. Nevertheless, the latter seems not to be the case with this phase It is noteworthy that higher pemetrexed concentrations are evident for the end of the infusion time point [i.e., even before administration of cyclophosphamide (Fig. 2)]. This suggests that the difference in pharmacokinetics between the current study population and the reference population treated with single-agent pemetrexed may not be a reflection of cyclophosphamide administration. Cyclophosphamide pharmacokinetics were not evaluated in the phase Iportion of this study. Intracellular and intratumoral concentrations of pemetrexed and cyclophosphamide are also unknown. A drug interaction between the two substances might be occurring that affects the functionality of pemetrexed and its metabolites. This might take the form of suppression of adequate transport of pemetrexed into the cell via inhibition of folate carrier or a reduction/ inhibition of the degree of pemetrexed polyglutamation via folylpolyglutamate synthase. Alternatively, although not expected based on the differing elimination pathways for pemetrexed and cyclophosphamide, pemetrexed could be altering cyclophosphamide concentrations or functionality. More extensive pharmacokinetic analyses will be done in the subsequent phase II portion of the study to further evaluate the plasma pharmacokinetics of both pemetrexed and cyclophosphamide. Investigations of folate carrier activity or folylpolyglutamate synthase and the relative concentrations of their intermediate products (reduced pemetrexed, glutaminated pemetrexed) could also provide explanation for the observed MTD. Following from the identification of the MTD, the protocol- specified dose recommended for the phase II study was the prior dose level; in this case, 2,100 mg/m2 pemetrexed and 600 mg/m2 cyclophosphamide. However, because DLTs (grade 4 thrombocytopenia and grade 3 rash/desquamation) were reported in two different patients at this dose level and because the pemetrexed dose would be double compared with that of the recommended dose of single-agent pemetrexed (as already detailed), it was decided not to carry this dose forward in the phase IIsetting.The phase IIportionof the study will use a low pemetrexed dose of 600 mg/m2 because this dose of pemetrexed has shown activity in phase II studies of metastatic disease (11–13) as well as in this phase Istudy (6 of 13 responses occurred at this or a lower dose). Pemetrexed at 1,800 mg/m2, the dose level below the recommended dose, will be applied as the high dose in the phase II evaluation because no DLTs were reported at this dose level and even more tumor responses were observed (two of the three enrolled patients) compared with that of the protocol-defined recommended dose 2 Fig. 3. Pemetrexed pharmacokinetic variables (individual empirical Bayesian of 2,100 mg/m (one of six enrolled patients). In addition, at estimates) versus pemetrexed dose. To determine the relationship between the 1,800 mg/m2 dose level, no grade 3 or 4 nonhematologic pemetrexed pharmacokinetics and pemetrexed dose, the absolute dose of pemetrexedwasplottedagainstplasmaclearance(top), central volume of distribution toxicities were seen and only one grade 3 neutropenia was (middle),and steady state volume of distribution (bottom). reported.

www.aacrjournals.org 7077 Clin Cancer Res 2006;12(23) December 1, 2006 Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2006 American Association for Cancer Research. Cancer Therapy: Clinical

In summary, this phase I study has reached its goal in combination of cyclophosphamide (600 mg/m2) and peme- identifying the MTD for pemetrexed and cyclophosphamide trexed (600 or 1,800 mg/m2) to further elucidate whether combination therapy in the treatment of patients with locally there exists a dose dependency about tumor response, advanced or metastatic breast cancer (2,400 and 600 mg/m2, pharmacokinetics, and toxicities for this drug combination respectively). As an immanent consequence, a prospective in patients with metastatic breast cancer in a multicenter randomized phase II study is now under way that tests the setting.

References 1. ParkinDM,BrayF,FerlayJ,PisaniP.Globalcancer 12. Spielmann M, Martin M, Namer M, duBois A, 22. Bonadonna G, Valagussa P, Moliterni A, Zambetti statistics, 2002. CA Cancer JClin 2005;55:74^ 108. Unger C, Dodwell DJ. Activity of pemetrexed M, Brambilla C. Adjuvant cyclophosphamide, metho- 2. Shih C, Grindey GB, Barnett CJ, et al. Structure- (ALIMTA, multitargeted antifolate, LY231514) in trexate, and fluorouracil in node-positive breast can- activity relationship studies of novel pyrrolopyrimidine metastatic breast cancer patients previously treated cer: the results of 20 years of follow-up. N Engl J antifolate LY231514 [abstract]. Proc Am Assoc with an anthracycline and a taxane: an interim analy- Med 1995;332:901 ^ 6. Cancer Res 1992;33:411. sis. Clin Breast Cancer 2001;2:47 ^ 51. 23. Joensuu H. Treatment of locally advanced and 3. Shih C, Gossett L, Gates S, et al. LY231514 and its 13. Martin M, Spielmann M, Namer M, et al. Phase II metastatic breast cancer. In: Souhami RL, Tannock I, polyglutamates exhibit potent inhibition against both study of pemetrexed in breast cancer patients pre- Hohenberger P, Horiot J-C, editors. Oxford textbook human dihydrofolate reductase (DHFR) and thymidy- treated with anthracyclines. Ann Oncol 2003;14: of oncology. 2nd ed. New York: Oxford University late synthase (TS): multiple folate enzymes inhibition 1246^ 52. Press; 2002. p. 1763 ^ 89. [abstract]. Ann Oncol 1996;7:85. 14. Llombart-Cussac A, Theodoulou M, Rowland K, 24. CancerTherapy Evaluation Program. Common Tox- 4. Grindey GB, Shih C, Barnett CJ, et al. LY231514, Lassus M, Cruciani S. A phase II trial of pemetrexed icity Criteria, version 2.0. DCTD, NCI, NIH, DHHS; a novel pyrrolopyrimidine antifolate that inhibits thymi- disodium (ALIMTA2, LY231514, MTA) in metastatic 19 9 8. dylate synthase (TS) [abstract]. Proc Am Assoc breast cancer (MBC) patients who have failed anthra- 25. Latz, JE, ChaudharyA, Ghosh A, Johnson RD. Pop- Cancer Res 1992;33:411. cyclines (A) and taxanes (T) (salvage chemotherapy) ulation pharmacokinetic analysis of ten phase II clini- 5. Calvert H, Bunn PA. Pemetrexed, a novel multitar- [abstract]. Breast Cancer ResTreat 2000;64:122. cal trials of pemetrexed in cancer patients. Cancer geted antifolate: current development and future 15. O’ShaughnessyJA, Clark RS, Blum JL, et al. Phase II Chemother Pharmacol 2006;57:401 ^11. directions. Semin Oncol 2002;29:1^ 61. study of pemetrexed in patients pretreated with an 26. Green S, Weiss GR. Southwest Oncology Group 6. Niyikiza C, Baker S, Johnson R,Walling J, Seitz D, anthracycline, a taxane, and capecitabine for ad- standard response criteria, endpoint definitions, and Allen R.MTA(LY231514):relationshipof vitaminmeta- vanced breast cancer. Clin Breast Cancer 2005;6: toxicity criteria. Invest New Drugs 1992;10:239 ^ 53. bolite profile, drug exposure, and other patient charac- 14 3 ^ 9. 27. Nakagawa K, Kudoh S, Matsui K, et al. A phase I teristics to toxicity [abstract].Ann Oncol1998;9:126. 16. Gomez HL, Hanauske A-R, Santillana S, et al. A study of pemetrexed supplemented with folic acid 7. Niyikiza C, Hanauske AR, Rusthoven JJ, et al. Peme- phase II trial of pemetrexed in previously untreated (FA) and vitamin b12 (VB12) in Japanese patients trexed safety and dosing strategy. Semin Oncol 2002; breast cancer [abstract]. Proc Am Soc Clin Oncol with solid tumors [abstract]. Eur JCancer 2004; 29:24 ^ 9. 2002;21:57a. 2:148. 8. Vogelzang NJ, Rusthoven JJ, Symanowski J, et al. 17. Colvin M. Cyclophosphamide and analogues. In: 28. Hammond LA, Forero L, Beeram M, et al. Phase 1 Phase III study of pemetrexed in combination with cis- Crooke ST, Prestayko AS, editors. Cancer and che- study of pemetrexed (LY231514) with vitamin supple- platin versus cisplatin alone in patients with malignant motherapy. New York: Academic Press; 1981. mentation in patients with locally advanced or meta- pleural mesothelioma. JClin Oncol 2003;21:2636^44. p. 25 ^ 38. static cancer [abstract]. Proc Am Soc Clin Oncol 9. Teicher BA, ChenV, Shih C, et al. Treatment regimens 18. Wall RL, Conrad FG. Cyclophosphamide therapy. Its 2003;22:133. including the multitargeted antifolate LY231514 in use in leukemia, lymphoma, and solid tumors. Arch 29. Hanauske A-R, Dittrich C, Otero J. Overview of human tumor xenografts. Clin Cancer Res 2000;6: Intern Med 1961;108:456 ^ 82. phase I/II pemetrexed studies. Oncology 2004;18: 1016 ^ 23. 19. Gordon I, McArthur J. Thiotepa and cyclophospha- 18 ^ 25. 10. TeicherBA,AlvarezE,LiuPC,etal.MTA mide in the treatment of advanced mammary cancer. 30. Frei E III, Antman KH. Principles of dose, schedule, (LY231514) in combination treatment regimens using Scott Med J1965;10:27 ^33. and combination chemotherapy. In: Bast RC, Kufe human tumor xenografts and the EMT-6 murine mam- 20. Carter SK, Livingston RB. Cyclophosphamide in DW, Pollock RE, et al, editors. Cancer medicine. 5th mary carcinoma. Semin Oncol 1999;26:55^ 62. solid tumors. Cancer Treat Rev 1975;2:295 ^ 322. ed.London:BCDecker;2000.p.556^68. 11. Miles DW, Smith IE, Coleman RE, Calvert AH, Lind 21. Bonadonna G, Brusamolino E, Valagussa P, et al. 31. Niyikiza C, Baker SD, Seitz DE, et al. Homocysteine MJ. A phase II study of pemetrexed disodium Combination chemotherapy as an adjuvant treatment and methylmalonic acid: markers to predict and avoid (LY231514) in patients with locally recurrent or meta- in operable breast cancer. N Engl JMed 1976;294: toxicity from pemetrexed therapy. Mol Cancer Ther static breast cancer. Eur JCancer 2001;37:1366 ^ 71. 405^10. 2002;7:545 ^ 52.

Clin Cancer Res 2006;12(23) December 1, 2006 7078 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 25, 2021. © 2006 American Association for Cancer Research. A Phase I Study of Pemetrexed (ALIMTA) and Cyclophosphamide in Patients with Locally Advanced or Metastatic Breast Cancer

Christian Dittrich, Lubos Petruzelka, Pavel Vodvarka, et al.

Clin Cancer Res 2006;12:7071-7078.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/12/23/7071

Cited articles This article cites 24 articles, 2 of which you can access for free at: http://clincancerres.aacrjournals.org/content/12/23/7071.full#ref-list-1

Citing articles This article has been cited by 2 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/12/23/7071.full#related-urls

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://clincancerres.aacrjournals.org/content/12/23/7071. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.