Initial therapy for women with advanced ovarian cancer in the United States usually consists of a combination of a platinum and a taxane agent. The majority of patients will relapse after this therapy, and their cancers eventually develop resistance to these classes of drugs. 1 A number of compounds have been reported to produce response rates between 10% and 20% in platinum-resistant patients, including topotecan, 2 oral etoposide, 3 liposomal doxorubicin (Doxil, Sequus Pharmaceuticals, Menlo Park, CA, U.S.A.), 4 ifosfamide, 5 gemcitabine, 6 and vinorelbine. One strategy for improving response rates is to combine active agents.
Ifosfamide is an analog of cyclophosphamide. Earlier trials 5,7 repeatedly documented its activity against carcinoma of the ovary even though all patients in those studies had been pretreated with cyclophosphamide, because cyclophosphamide was previously included in standard primary chemotherapy for ovarian cancer. We postulated that the activity of ifosfamide might be greater in patients who had not been pretreated with cyclophosphamide.
Vinorelbine is a semisynthetic vinca alkaloid approved by the Food and Drug Administration for use in non–small-cell carcinoma of the lung. It has demonstrated significant antitumor activity in a number of malignancies 8,9 and has produced response rates of 15% to 30% in a number of studies in the setting of heavily pretreated ovarian cancer. 10–13
The primary side effect of both ifosfamide and vinorelbine is neutropenia, which we believed could be circumvented with the use of colony-stimulating factors. We therefore chose to test a regimen of ifosfamide and vinorelbine, which had already been developed and tested in patients with non–small-cell lung cancer, 14 in patients with previously treated ovarian cancer. Because there have been published reports of the development of severe peripheral neurotoxicity as the result of vinorelbine treatment in patients who had received prior paclitaxel therapy, 15–17 we used a brief questionnaire to monitor peripheral neuropathy.
PATIENTS AND METHODS
Eligible patients had epithelial ovarian cancer or primary peritoneal cancer of papillary serous histology. Tumors of low malignant potential were excluded. Disease must have persisted or relapsed after at least one prior chemotherapy regimen, and patients with prior ifosfamide or vinorelbine treatment were excluded. Although it was not required for entry, all patients had received prior therapy with both paclitaxel and a platinum compound. Measurable or evaluable disease was required. Evaluable disease included disease that could be followed, if not precisely quantitated, radiologically or on physical examination. Patients with only an elevated CA-125 and no evidence of disease on radiologic or physical examination were not accepted. A central venous access device was required. Other eligibility criteria included age 18 years or more, Gynecologic Oncology Group performance status less than or equal to 2, absence of active central nervous system disease, platelet count greater than or equal to 100,000/μl, absolute neutrophil count greater than or equal to 1,500/μl, serum creatinine less than or equal to 1.5 mg/dl, and bilirubin within normal limits. All patients signed informed consent in accordance with federal and institutional guidelines.
Chemotherapy Treatment Plan
Ifosfamide was diluted in 500 ml 5% dextrose or 0.9% sodium chloride and infused intravenously during 30 minutes to 2 hours on days 1, 2, and 3. Mesna was administered at a dose of 640 mg/m2 for three doses each day. The first dose was mixed in the same solution as the ifosfamide. The second and third doses were mixed in 180 ml soft drink or juice and administered orally 2 hours and 8 hours after the start of the ifosfamide infusion unless the patient was vomiting, in which case the doses were mixed in 250 ml 5% dextrose or sodium chloride and administered intravenously 4 hours and 8 hours after the start of the ifosfamide infusion. Vinorelbine was administered intravenously for 6 to 10 minutes days 1, 2, and 3 either before or after ifosfamide. Antiemetics were left to the discretion of the treating physician, but ondansetron at a dose of 32 mg intravenously was suggested. Fluids were adjusted to ensure a daily fluid intake of at least 2 l. Therapy was administered either in the inpatient or the outpatient setting, as appropriate for the individual patient. All patients received subcutaneous granulocyte colony-stimulating factor 5 μg/kg/d days 5 to 11, or continuing until the absolute neutrophil count was greater than or equal to 4,000/μl. Prophylactic stool softeners were recommended for all patients. Cycles were repeated every 21 days.
The initial doses were ifosfamide 1.6 g/m2/d × 3 days and vinorelbine 30 mg/m2/d × 3 days. These proved too toxic, and starting doses were sequentially reduced to a final combination of 960 mg/m2/d × 3 of ifosfamide and 20 mg/m2/d × 3 of vinorelbine.
The successive dose levels used can be seen in Table 1. Subsequent doses of both ifosfamide and navelbine in an individual patient were reduced by 25% for platelet count nadir less than or equal to 20,000/μl, neutropenic fever, or if treatment was delayed for failure to recover counts to an absolute neutrophil count of greater than or equal to 1,500/μl and platelet count of greater than or equal to 100,000/μl by day 21.
Patient Monitoring and Response Evaluation
Serum chemistries were repeated each cycle. A complete blood count with platelet and differential count was repeated twice weekly. Toxicities were graded according to the Cancer and Leukemia Group B Expanded Common Toxicity Criteria. Table 2 shows the Cancer and Leukemia Group B Expanded Common Toxicity Criteria for sensory neuropathy. Peripheral sensory neuropathy was assessed with a brief questionnaire that was supposed to be administered at baseline and at least every other cycle. This included five questions selected from the Functional Assessment of Cancer Therapy/Gynecologic Oncology Group neurotoxicity assessment (Table 3). 18 Each was scored from 0 to 3 (not at all to very much). Any numbness or tingling was considered at least grade I neuropathy. A score of 2 on at least two of the questions about numbness and tingling or 3 on at least one was considered at least grade II neuropathy. Any difficulty buttoning was considered a grade III neuropathy (functional impairment) although this would often be only grade II in the revised Common Toxicity Criteria version 2.0 19 (which were issued after completion of this trial). A sum of the scores on all five questions was used to crudely assess progression of neuropathy during vinorelbine therapy.
Tumor measurements were repeated after every two cycles. Complete response was defined as normalization of the CA-125 along with complete disappearance of all detectable tumor for at least 6 weeks. A duration of 6 weeks rather than 4 weeks was chosen to avoid extra radiologic imaging for response confirmation. Partial response was defined as reduction by at least 50% of the products of the longest perpendicular diameters of the most easily measured or largest tumor mass (indicator lesion), without growth of any lesion or appearance of new lesions. These conditions had to exist for at least 6 weeks. Progression was defined as increase by more than 25% of the product of the perpendicular diameters of the indicator lesion, or the appearance of new metastatic lesions. Stable disease was defined as a change in the product of the perpendicular diameters of the indicator lesion by less than required for progression or response with no appearance of new lesions. A CA-125 response in patients with evaluable disease was defined according to the criteria proposed by Rustin et al. as either a serial decrease over three samples to less than 25% of the first sample or a 50% decrease after two samples confirmed by a fourth sample. 20
A two-stage design was used. Fourteen patients were treated in the first stage. If no major responses (complete response or partial response) were seen among the first 14 patients (95% confidence that the true response rate was no higher than 20%), then the study would close. For this purpose, patients with evaluable disease were considered stable unless they had a complete response. If one or more responses were observed, a total of 30 patients would be treated. This would allow for a 95% confidence interval of ± 13% if the true response rate were 20%.
Definition of Platinum-resistant
Platinum resistant was defined as a less than partial response at the end of the most recent treatment with a platinum compound with less than 6 months until relapse after treatment with a platinum compound.
Thirty patients were entered between April 1995 and May 1997. Their characteristics are summarized in Table 4. They received a total of 110 courses of chemotherapy, with the median number of courses per patient being 3 (range: 1–9). Fourteen patients received four or more cycles, and seven received only one cycle. Median survival from time of study entry for all patients was 9 months.
Hematologic toxicity after cycle 1 is shown in Table 5. Neutrophil toxicity was severe despite the routine use of granulocyte colony-stimulating factor, and high rates of neutropenic fever and two septic deaths at the early dose levels caused us to progressively reduce the starting doses of both drugs. The final starting doses were ifosfamide 960 mg/m2/d × 3 days and vinorelbine 20 mg/m2/d × 3 days. The 10 patients treated at these doses experienced no neutropenic fevers at any time during their therapy, although grade IV absolute neutrophil count remained common.
Nineteen patients completed at least 2 questionnaires, and 13 completed 3 or more. Six patients failed to complete a baseline neuropathy assessment, and two patients completed no neurotoxicity questionnaires. Baseline neuropathy for patients completing the questionnaire was grade 0 = 6 patients, grade I = 8 patients, grade II = 1 patient, and grade III = 7 patients. Four patients had a worsening grade of neuropathy during therapy: three who had been grade 0 developed grade I neuropathy, and one who had been grade I became grade III. Considering just those 11 patients with a repeat neurotoxicity assessment after four or more cycles, only 2 had any worsening of neuropathy. One progressed from grade 0 to grade I, and the other from grade I to grade III. The latter patient noted mild difficulty buttoning for the first time after four cycles of therapy, although she did not describe any worsening of the baseline mild numbness and tingling in her hands.
When the sum of the scores on the five questions was evaluated, six patients had improvement (reduction in sum score), and five had worsening. Of the seven patients who had a baseline grade III neuropathy, four completed at least two questionnaires. Two had improvement, one was stable, and one progressed from a sum score of 5 to a sum score of 6 after nine cycles of therapy.
Although the protocol recommended prophylactic stool softeners, compliance was not documented. Four patients developed grade III or IV constipation.
Other Nonhematologic Toxicity
Nonhematologic toxicities apparently were not dose related, and are summarized in Table 6. Two patients experienced central nervous system toxicity attributed to ifosfamide. One had dramatic mental status changes (grade III), and the other had agitation and sleeplessness (grade II). Allergic reactions included a drug fever that recurred on rechallenge with single-agent ifosfamide, facial erythema and swelling that recurred on rechallenge with the drug combination, and a rash that resolved and recurred on rechallenge with the drug combination. The grade III bony pain was related to granulocyte colony-stimulating factor, and improved with a 50% dose reduction of that agent. One other patient had grade II back pain attributed to granulocyte colony-stimulating factor injection.
There was only one partial response in measurable disease among 30 patients treated in this trial (response rate of 3%). This occurred at dose level −2 in a patient with platinum-sensitive disease who had previously been exposed to cyclophosphamide. Five of the 23 patients with measurable disease were not evaluable for response because they discontinued therapy after one cycle as a result of toxicity. Two of the seven patients with evaluable disease had CA-125 responses. Both were treated at dose level −3, had platinum-refractory disease, and were cyclophosphamide naïve. Three patients with measurable disease also had CA-125 responses; all had stable disease by radiologic examination. The patient with a partial response did not have an elevated CA-125.
Neutropenic fever developed in 3 of 13 chemotherapy-naïve patients with lung cancer (23%) on the recommended phase II dose of the phase I study from which the initial doses of our treatment regimen were taken. 14 The cycle 1 rates of neutropenic fever in our study were not too dissimilar (5 of 17, or 29%), particularly considering that all of our patients were pretreated. However, the fact that two patients died of neutropenic sepsis seemed unacceptable for a palliative therapy. At the final dose level, this regimen was hematologically well tolerated. However, the low level of activity was disappointing. Some responses may have been obscured by the high levels of toxicity observed in the early part of the study, which rendered several patients inevaluable for response. Nevertheless, even taking that into account, our response rate was not superior to the 15% to 30% observed with single-agent vinorelbine at doses that do not require growth factor support. This is probably related to the heavy level of pretreatment in our patient population (median of three prior regimens). It is also possible that the schedule of navelbine used on this trial (daily × 3 every 3 weeks) is not as active as the more usual weekly schedule. However, Gershenson et al. 11 studied a similar schedule in patients with a maximum of two prior regimens, and observed an overall response rate of 30%.
We did not see evidence that ifosfamide is a more effective salvage regimen in patients who have not received prior cyclophosphamide therapy. The low doses of ifosfamide used could have decreased its effectiveness. However, our results are similar to those reported by Markman et al., 21 who tested single-agent ifosfamide in 21 pretreated, cyclophosphamide-naïve patients with ovarian cancer, and achieved an objective response in measurable disease in only one patient.
Interestingly, three of our patients with measurable disease who had CA-125 responses had stable disease by radiologic measurement. Overall response rates using World Health Organization criteria and CA-125 response rates have been noted to be similar both for first-line 20 and second-line chemotherapy. 22 Possibly our findings are due to chance, given the small sample size of our trial, or possibly CA-125 alone is a less reliable surrogate for response rate in heavily pretreated patients. It might still be an equally good indicator of the potential activity of our drug combination in a less heavily pretreated population.
Although all our patients had been pretreated with paclitaxel, and eight had significant neuropathy before entry in this study, we did not observe the dramatic peripheral neurotoxicity that has been reported in paclitaxel-pretreated patients. Fazeny et al. 16 reported results of a trial of vinorelbine at a dose of 30 mg/m2 given every other week to 14 patients with advanced breast cancer, all whom had been pretreated with paclitaxel. In four cases, after only one to three doses of vinorelbine, grade III neuropathy was observed. Three of the patients had sensory ataxia and bilateral foot drop that made them unable to walk unaided. Baseline neuropathy of these patients was grade I in three cases and grade II in one case. The authors concluded that patients with prior paclitaxel therapy should not receive vinorelbine. Parimoo et al. 17 reported on five patients treated with a combination of paclitaxel 150 mg/m2/every 2 weeks and vinorelbine 20 mg/m2 to 30 mg/m2 every 2 weeks. All had prior paclitaxel therapy, and grade I to II baseline sensory neuropathy. Three had ovarian cancer. Grade III to IV sensorimotor neuropathy developed in all of these patients, with four requiring a wheelchair, and two having near-total paresis of toe flexors and extensors. The authors concluded that patients with preexisting neuropathy from prior paclitaxel should be excluded from vinorelbine therapy.
Table 7 reviews the degree of peripheral neuropathy reported by the four published trials of single-agent vinorelbine in patients with ovarian cancer. Because many patients with ovarian cancer would have been exposed to cisplatin as well as paclitaxel, their risk for vinorelbine-induced neurotoxicity might be even higher than that of patients with breast cancer. However, dramatic peripheral neuropathy appears to be uncommon. Obviously, grading of neurotoxicity, like grading of much nonhematologic toxicity, is quite subjective. Detection of grade I to II toxicities will depend on how vigorously they are sought. We looked fairly carefully, and did not see evidence of unacceptable neurotoxicity. Parimoo et al. were giving a combination of vinorelbine plus fairly dose-intense paclitaxel to their patients, and this may represent a greater insult than vinorelbine alone. Two of the patients of Fazeny et al. in whom neuropathy developed had received more than 3,000 mg/m2 of paclitaxel, which is well in excess of the 810 to 1,050 mg/m2 that would represent six cycles of chemotherapy for a typical patient with ovarian cancer. Neurotoxicity is a particularly distressing side effect because of its long duration and effect on quality of life. Clearly, vinorelbine has the potential to cause occasional significant neurotoxicity. However, we believe that vinorelbine can be used with reasonable safety in patients who have had prior paclitaxel therapy as long as their baseline symptoms are mild, and they are monitored for progression of peripheral neuropathy. This may be of particular importance in patients with breast cancer, who are now often given adjuvant paclitaxel, and for whom vinorelbine represents active salvage chemotherapy.
Acknowledgment: The authors thank Jan Marie McEvilly for data management support.
1. McGuire WP, Ozols RF. Chemotherapy of advanced ovarian cancer
. Semin Oncol 1998; 25: 340–8.
2. Kudelka AP, Tresukosol D, Edwards CL, et al. Phase II study of intravenous topotecan as a 5-day infusion for refractory epithelial ovarian carcinoma. J Clin Oncol 1996; 14: 1552–7.
3. Rose PG, Blessing JA, Mayer AR, et al. Prolonged oral etoposide as second-line therapy for platinum-resistant and platinum-sensitive ovarian carcinoma: a Gynecologic Oncology Group study. J Clin Oncol 1998; 16: 405–10.
4. Muggia FM, Hainsworth JD, Jeffers S, et al. Phase II study of liposomal doxorubicin in refractory ovarian cancer
: antitumor activity and toxicity modification by liposomal encapsulation. J Clin Oncol 1997; 15: 987–93.
5. Markman M, Hakes T, Reichman B, et al. Ifosfamide
and mesna in previously treated advanced epithelial ovarian cancer
: activity in platinum-resistant disease. J Clin Oncol 1992; 10: 243–8.
6. Shapiro JD, Millward MJ, Rischin D, et al. Activity of gemcitabine in patients with advanced ovarian cancer
: responses seen following platinum and paclitaxel. Gynecol Oncol 1996; 63: 89–93.
7. Sutton GP, Blessing JA, Homesley HD, et al. Phase II trial of ifosfamide
and mesna in advanced ovarian carcinoma: a Gynecologic Oncology Group study. J Clin Oncol 1989; 7: 1672–6.
8. Morris M, Brader KR, Levenback, et al. Phase II study of vinorelbine
in advanced and recurrent squamous cell carcinoma of the cervix. J Clin Oncol 1998; 16: 1094–8.
9. Jones S, Winer E, Vogel C, et al. Randomized comparison of vinorelbine
and melphalan in anthracycline-refractory advanced breast cancer. J Clin Oncol 1995; 13: 2567–74.
10. George MJ, Heron JF, Kerbrat P, et al. Navelbine in advanced ovarian epithelial cancer: a study of the French oncology centers. Semin Oncol 1989; 16 (suppl 4): 30–2.
11. Gershenson DM, Burke TW, Morris M, et al. A phase I study of a daily × 3 schedule of intravenous vinorelbine
for refractory epithelial ovarian cancer
. Gynecol Oncol 1998; 70: 404–9.
12. Bajetta E, Di Leo A, Biganzoli L, et al. Phase II study of vinorelbine
in patients with pretreated advanced ovarian cancer
: activity in platinum-resistant disease. J Clin Oncol 1996; 14: 2546–51.
13. Burger RA, DiSaia PJ, Roberts JA, et al. Phase II trial of vinorelbine
in recurrent and progressive epithelial ovarian cancer
. Gynecol Oncol 1999; 72: 148–53.
14. Masters GA, Hoffman PC, Hsieh A, et al. Phase I study of vinorelbine
in advanced non-small-cell lung cancer. J Clin Oncol 1997; 15: 884–92.
15. Dittrich C, Zifko U, Fazeny B, et al. Vinorelbine
after paclitaxel in breast cancer: cross resistance and cumulative neurotoxicity
? Ann Oncol 1994; 5: 473–4.
16. Fazeny B, Zifko U, Meryn S, et al. Vinorelbine
in patients with advanced breast cancer pretreated with paclitaxel-a phase II study. Cancer Chemother Pharmacol 1996; 150–6.
17. Parimoo D, Jeffers S, Muggia FM. Severe neurotoxicity
-paclitaxel combinations. J Natl Cancer Inst 1996; 88: 1079–80.
18. Walczak JR, Almadrones L, Brady M. Assessment of peripheral neuropathy and physical/role function in a phase III randomized chemotherapy trial in the gynecologic oncology group (GOG). Oncol Nurs Forum 1993; 20: 358.
19. http. //184.108.40.206/ctc/plsql/ctc000w$.startup, accessed 7-1-99.
20. Rustin GJS, Nelstrop AE, McClean P, et al. Defining response of ovarian carcinoma to initial chemotherapy according to serum CA 125. J Clin Oncol 1996; 14: 1545–51.
21. Markman M, Kennedy A, Sutton G, et al. Phase II trial of single agent ifosfamide
/mesna in patients with platinum/paclitaxel refractory ovarian cancer
who have not previously been treated with an alkylating agent. Gynecol Oncol 1998; 70: 272–4.
22. Rustin GJS, Nelstrop AE, Crawford M, et al. Phase II trial of oral altretamine for relapsed ovarian carcinoma: evaluation of defining response by serum CA125. J Clin Oncol 1997; 15: 172–6.