The aim of this study was to review the experience of the Institut Gustave Roussy in 163 patients with malignant mesothelioma over a 9-year period. Data from seven consecutive prospective trials, four of chemo-immunotherapy and three of chemotherapy were reviewed. The rationale, methods and results of these trials are summarised and discussed. 98 patients were included in four phase II trials of chemo-immunotherapy whose common denominator was a combination of cisplatin and alpha-interferon. The response rate ranged from 15% to 40%. High-dose weekly cisplatin combined with alpha-interferon yielded the highest response rate but the toxicity of this regimen was considered unacceptable. Neither higher doses of alpha-interferon or the addition of mitomycin C or interleukin-2 to the regimen were able to enhance the activity of this combination. 18 patients were included in a paclitaxel-cisplatin phase II trial. The response rate was only 6% (95% confidence interval (CI): 0-24) and toxicity was also significant. This regimen was, therefore, considered ineffective. Of 17 patients with mesothelioma included in a phase I trial that combined raltitrexed and oxaliplatin, 6 (35%) obtained a partial response. Responses were seen even in cisplatin-refractory mesothelioma. Preliminary results of a subsequent ongoing phase II trial using raltitrexed (3 mg/m(2)) and oxaliplatin (130 mg/m(2)) have confirmed this promising activity with a 30% (9/30) response rate (95% CI: 15-49). The tolerance of this outpatient regimen is acceptable (no significant haematological toxicity and no alopecia) and compares favourably with that of our previous regimens. The final results concerning response and survival are required to confirm the efficacy of this combination. The preliminary results of two studies suggest promising activity with the combination of raltitrexed-oxaliplatin in malignant mesothelioma. The efficacy/toxicity ratio of this combination compares favourably with that of our previous chemotherapy and chemo-immunotherapy regimens.