Second-Line Chemotherapy with Cisplatin-Ifosfamide in Patients with Ovarian Cancer Previously Treated with Carboplatin-Cyclophosphamide

In an effort to use antineoplastic drug combinations which are active in platinum resistant ovarian cancer or which can induce a second response after a platinum first-line treatment, we conducted a study on 30 ovarian cancer patients previously treated with carboplatin plus cyclophosphamide who were given ifosfamide 5 g/m2 i.v. divided over days 1 to 3 plus mesma combined with cisplatin 100 mg/m2 i.v. divided over days 1 to 3 every 4 weeks as second-line treatment. Eight patients had never entered remission with first-line chemotherapy while 22 patients had tumor recurrence within 6 to 18 months after the end of chemotherapy and their tumors were considered potentially platinum sensitive. Responding patients received 6 courses while palliative treatment for nonresponders was provided. Of the 22 patients with tumor recurrence, 8 patients responded with one partial response (PR) and 7 complete clinical responses (CCR). Two out of the 8 patients with platinum resistant disease demonstrated short lasting PR. Seven patients with CCR underwent second-look operation and in two a pathological CR was documented. Median time to progression was 6 mo (4-12). The median overall survival was 12 mo (4-20). Myelotoxicity despite G-CSF administration was significant with grade 4 leukopenia in 40% and grade 3 thrombocytopenia in 20% of patients. Central nervous system (CNS) toxicity was significant with 30% somnolence, 20% disorientation and an episode of grand-mal epilepsy ascribed to ifosfamide. With a 33% response rate the combination is as effective as new agents employed in relapsed ovarian cancer. Platinum-refractory disease may respond to a lesser degree. The most important determinant of response was the progression-free interval from first-line chemotherapy. Whether patients recurring after carboplatin plus cyclophosphamide have a greater chance to respond to cisplatin plus ifosfamide or vice-versa cannot be supported by the current data and therefore randomized studies should be performed to this end.