On the basis of its efficacy against ovarian carcinoma and its safe peritoneal administration, cisplatin administered by the intraperitoneal route was studied in a phase II multicentric trial. 34 patients with good performance status and residual disease less than 1 cm were treated with a 90 mg/m2 dose (60 mg/m2 at first cycle), administered in the abdominal cavity every 3 weeks for at least four cycles. In case of haematological or renal toxicity, intravenous sodium thiosulphate was perfused simultaneously with intraperitoneal cisplatin with protective intent. 25 patients were evaluable for response: 3 patients had pathological complete response and 1 patient had a microscopic disease (16% response rate in evaluable patients). Systemic toxicity was mild, and sodium thiosulphate clearly protected against leucopenia (6 patients) and renal toxicity (8 patients). Local side-effects were evaluable in 34 patients with 2 cases of infectious peritonitis, 1 of wound infection and 2 of haemorrhage. Of the 147 evaluable chemotherapy cycles, nine resulted in partial and one in total inflow obstruction, for which 4 patients needed surgical procedures for catheter-related complications, and 1 patient died of acute abdominal complications after such a procedure. We conclude that 90 mg/m2 intraperitoneal cisplatin has activity in pretreated patients with minimal residual disease, and that thiosulphate protects against haematological and renal toxicities. Only a randomised study can demonstrate a true benefit, which will have to be balanced with the toxicity of intraperitoneal drug administration.