Study Population

Patients with platinum-sensitive (defined as a cancer initially platinum sensitive followed by a progression-free interval from the patient's last exposure to platinum of > 6 months) recurrent EOC, peritoneal serous cancer, or fallopian tube cancer were eligible. Patients ≥ 18 years of age with a life expectancy of ≥ 12 weeks were required to have measurable cancer according to Response Evaluation Criteria in Solid Tumors (RECIST)¹⁵ via computed tomography or magnetic resonance imaging scan and an Eastern Cooperative Oncology Group (ECOG) performance status of 0 or 1. Patients could have received up to two prior chemotherapy regimens in the recurrent setting (platinum or nonplatinum containing). Maintenance taxanes and hormonal therapies did not count as prior treatment, but antiangiogenic therapy did count as a prior treatment.

Eligibility criteria included a total WBC count ≥ 2,000/μL, absolute neutrophil count (ANC) ≥ 1,000/μL, platelet count ≥ 100,000/μL, total bilirubin less than 1.5× the upper limit of normal (ULN), serum creatinine less than 1.5× ULN, calculated creatinine clearance more than 45 mL/min (using Cockcroft-Gault formula¹⁶), ALT/AST less than 3× ULN (no liver metastases), and ALT/AST less than 5× ULN (with liver metastases). Exclusion criteria included prior radiotherapy to more than 25% of bone marrow, uncontrolled medical and/or psychiatric illness, history of active CNS metastases, history of prior malignancy (except treated cervical carcinoma in situ, stage I endometrial cancer, treated basal or squamous cell skin cancer, or breast cancer for which the patient had been disease free for ≥ 5 years), grade ≥ 2 peripheral neuropathy, or history of anaphylactic shock with prior platinum chemotherapy.

Study Design

This study was conducted at the Dana-Farber Cancer Institute, Brigham and Women's Hospital, and Massachusetts General Hospital (all in Boston, MA). The study was investigator initiated; research funding was provided by Eli Lilly & Co (Indianapolis, IN). The primary objective was to determine the tumor response rate by RECIST criteria.¹⁵ Secondary objectives included toxicity assessment, progression-free survival (PFS), and overall survival (OS).

Patients were required to take at least five daily doses of folic acid (at least 400 μg/dose) 1 week before treatment; dosing continued throughout treatment and for 21 days after treatment finished. Patients received one intramuscular injection of vitamin B₁₂ 1,000 μg 1 week before starting treatment and every three cycles (9 weeks). Additional premedication included dexamethasone 4 mg orally twice per day the day before, the day of, and the day after treatment for rash prophylaxis. Antiemetic agents were decided by the treating physician. Erythropoietic stimulating agent use was dictated by hospital and US Food and Drug Administration policy. Myeloid growth factor was not permitted for cycle 1 of study treatment, but thereafter, myeloid growth factor use was determined by the treating physician.

Starting doses of chemotherapy were pemetrexed 500 mg/m² intravenous over a 5- to 10-minute bolus (administered first), followed by carboplatin area under the curve (AUC) 5 (calculated via the Cockcroft-Gault equation¹⁶) intravenous over 30 to 60 minutes administered every 21 days for six cycles or for up to eight cycles if clinical benefit occurred and toxicities were manageable. The duration of treatment was chosen to mirror the study design of the ICON4 study³ and the study by Pfisterer et al.⁴ Patients who developed a carboplatin hypersensitivity reaction (HSR) were allowed to follow a desensitization protocol if agreed on by the patient and the treating physician and were under the care of an allergist.¹⁷

Toxicity and Efficacy

Before starting treatment, patients underwent a history (including concomitant medications), physical examination, determination of ECOG performance status, CBC (including WBC, ANC, platelet count, and hemoglobin levels), chemistry panel, CA125, and either a computed tomography or magnetic resonance imaging scan. Toxicities were reported using the Common Terminology Criteria for Adverse Events version 3. Patients underwent weekly CBCs during treatment, and all laboratory tests, including CBC, chemistry panel, and CA125, were rechecked on day 1 of each cycle.

Any patient receiving two or more cycles was assessable for tumor response by RECIST criteria,¹⁵ and if one or more cycle was received, patients were assessable for toxicity. Tumor response was assessed every two cycles.

Dose Modifications and Delays

To receive cycle 2 and beyond, patients needed to have an ANC of ≥ 1,000/μL, platelet count of ≥ 100,000/μL, and resolution of toxicities to ≤ grade 1. Doses were reduced for both drugs by 20% for any of the following: febrile neutropenia, platelet count less than 50,000/μL, ANC less than 500/μL lasting more than 7 days, or failure to recover either platelet count to more than 50,000/μL or ANC to more than 1,000/μL on planned day 22 of therapy. If counts were adequate by day 35, patients underwent dose reductions for both drugs by 20%. Study drugs were reduced by 20% for any episode of grade 3 diarrhea, mucositis, nausea, or vomiting. Patients were removed from study if more than two dose reductions occurred, counts were inadequate by day 35, or a grade 3 or 4 nonhematologic toxicity did not resolve by day 1 of the next cycle.

Discontinuation of Treatment

Treatment was discontinued when any of the following events occurred: radiographic or clinical evidence of cancer progression; deterioration of health or the treating physician believed it was in the patient's best interest to stop therapy; intolerable toxicities; or patient refusal. Once patients stopped treatment, poststudy treatment was dictated by the treating physician.

Statistical Analysis

The primary end point was to determine the overall response rate. Secondary end points included duration of response, PFS, and OS. The overall response rate included the combined rate of complete and partial responses, was reported as the proportion of the events among all patients, and was calculated via the intent-to-treat population. Expected response rates greater than 30% were based on the data by Markman et al² and on the fact that combination therapy was being used, which should result in a higher response rate than carboplatin alone. The study sample size was determined to achieve an 80% power to reject, at a 5% α level by using a binomial test, the null hypothesis that the observed response rate is less than 20% in favor of the alternative hypothesis that the response rate is more than 36%. The test would have rejected the null hypothesis if there were 14 or more complete or partial responses. The response rates and their 95% CIs were estimated based on the exact binomial distribution. Duration of response was measured from the time that measurement criteria were met for response until the first date that progressive disease was documented. PFS was measured from the start of treatment until evidence of progression of cancer. OS was measured from date of start of study treatment to the date of death from any cause. PFS and OS were estimated using the Kaplan-Meier method. Toxicity events were summarized descriptively by frequency distributions for ≥ grade 2 toxicity events.

Enrollment and Demographics

Between September 21, 2005 and July 26, 2007, 45 patients were enrolled. One patient withdrew consent but eventually reconsented at a later time point. Forty-four patients were assessable for toxicities, and 41 patients were assessable for response. Three patients withdrew from study before two cycles were completed; two patients developed unacceptable toxicities and were withdrawn (grade 3 fatigue and grade 3 diarrhea). One patient developed a severe HSR during cycle 2, and the patient withdrew before having her cancer reassessed for tumor response. The mean duration of follow-up for this study was 15.3 months.

Demographic and baseline characteristics are listed in Table 1. Ninety-three percent of patients had recurrent EOC (n = 41), whereas the remainder had either fallopian tube cancer (n = 1) or peritoneal serous cancer (n = 2). Most patients (68.2%) had papillary serous cancer, and 18.2% of patients had mixed histologies. The mean age was 59 years (range, 36 to 78 years), and all patients were white. Fifty-seven percent of patients had an ECOG performance status of 0, and the remainder (43%) had a performance status of 1. All patients received a prior platinum- and taxane-containing regimen as their initial treatment. Fifty percent of patients had no prior treatment for a recurrence, whereas 16 patients (36%) had received one prior regimen and six patients (14%) had received two prior regiments for recurrent cancer. Thirty percent of patients (n = 14) received at least one platinum-containing regimen in the recurrent setting. The median platinum-free interval was 19 months (range, 6 to 52 months). Thirteen patients (30%) had a previous platinum-free interval of between 6 and 12 months, whereas the remaining 31 patients (70%) had a ≥ 12-month platinum-free interval.

Response Rates

Response rates were calculated via intent-to-treat analysis. The overall response rate was 51.1% (95% CI, 35.8% to 66.3%), and all responses were partial responses; there were no complete responses (Table 2). Two patients had an unconfirmed partial response. Two patients (4.4%) had progressive disease after two cycles, and 14 patients (31.1%) had stable disease. Most partial responses were noted after two cycles of therapy (13 patients). The mean (± standard deviation) duration of response was 28.9 (± 18.14) weeks.

Seven patients completed eight cycles of chemotherapy and were removed from study because the maximum number of cycles was eight cycles. Of these seven patients, six had partial responses and one had stable disease after completion of therapy.

Toxicities

A total of 235 cycles was administered, and the mean number of cycles administered per patient was 5.34 cycles (range, one to eight cycles). Twenty-eight patients received six or more cycles of therapy. Table 3 lists any grade 2 or greater toxicities occurring in ≥ 5% of patients and any grade 3 or 4 toxicity. Toxicity analysis was based on the patients who were assessable for treatment toxicity. Grade 3 and 4 hematologic toxicities included neutropenia (41%), anemia (9%), and thrombocytopenia (23%). No patients experienced febrile neutropenia. Eight patients were prescribed myeloid growth factor support by their treating physician at some point during therapy.

For nonhematologic toxicities, 11% of patients experienced grade 3 fatigue, and 5% experienced grade 3 nausea, vomiting, diarrhea, pulmonary embolus, and/or syncope. Both episodes of pulmonary emboli were managed successfully with anticoagulation. Two percent of patients experienced an episode of grade 3 constipation or memory impairment. There were no grade 4 nonhematologic toxicities. Three patients experienced self-limited grade 2 conjunctivitis treated with corticosteroid eye drops. No patients experienced ≥ grade 3 mucositis or rash.

Sixteen patients (32%) experienced carboplatin HSR during treatment. Of these patients, nine patients discontinued treatment, and the remaining seven patients received a carboplatin desensitization protocol per the allergy service, allowing them to continue receiving study treatment.¹⁷ All of the HSRs were grade 3 and required intravenous medication during the reaction such as diphenhydramine and corticosteroids; none required epinephrine. Of the 16 patients who had a carboplatin HSR, six had received prior platinum in the recurrent setting, whereas the other 10 had only received platinum as part of their initial treatment.

Two hundred thirty-five cycles of chemotherapy were administered; 44 cycles were dose reduced, and in all of these cycles, doses were reduced for both carboplatin and pemetrexed. Thirteen patients (29%) accounted for these dose reductions (Table 4). Of these 13 patients, 10 patients were dose reduced by one dose level, and three patients required two dose reductions. Of the 13 patients who experienced dose reductions, nine were a result of thrombocytopenia, one was a result of prolonged neutropenia, and three were a result of other reasons. One patient came off trial because of persistent thrombocytopenia despite two dose reductions. Dose delays were observed in 19 (8%) of 235 cycles. Reasons for dose delays included carboplatin reaction (n = 5), surgical procedure (n = 1), prolonged neutropenia (n = 2) or thrombocytopenia (n = 6), fatigue (n = 1), vacation (n = 1), patient did not take folic acid (n = 1), pharmacy reasons (n = 1), and an episode of small bowel obstruction (n = 1).

The planned dose-intensity (planned mean dose per week) of carboplatin was AUC 1.67, and the actual dose-intensity (the dose the patient actually received per week) observed in this study, taking into account dose reductions and delays, was AUC 1.59, with a relative dose-intensity of 95.2% (actual dose divided by planned dose × 100%; Table 4).

PFS and OS

PFS is shown in Figure 1. Mean PFS time was 8.54 months (95% CI, 7.08 to 9.90 months), and median PFS time was 7.57 months (95% CI, 6.44 to 10.18 months). Mean OS time was 20.3 months (Fig 2), and the median OS time has not been reached.

View larger version:

• In this window
• In a new window

• PowerPoint Slide for Teaching

Fig 1.

Progression-free survival.

View larger version:

• In this window
• In a new window

• PowerPoint Slide for Teaching

Fig 2.

Overall survival.