SIOPEL 2 was a pilot study designed to test the efficacy and toxicity of two chemotherapy (CT) regimens, one for patients with hepatoblastoma (HB) confined to the liver and involving no more than three hepatic sectors ('standard-risk (SR) HB'), and one for those with HB extending into all four sectors and/or with lung metastases or intra-abdominal extra hepatic spread 'high-risk (HR) HB'. SR-HB patients were treated with four courses of cisplatin (CDDP), at a dose of 80 mg/m(2) every 14 days, delayed surgery, and then two more similar CDDP courses. HR-HB patients were given CDDP alternating every 14 days with carboplatin (CARBO), 500 mg/m(2), and doxorubicin (DOXO), 60 mg/m(2). Two courses of CARBO/DOXO and one of CDDP were given postoperatively. Between October 1995 and May 1998, 77 SR-HB (10 of whom were actually treated with the HR protocol) and 58 HR-HB patients were registered and all 135 could be evaluated. Response rates for the entire SR-HB and HR-HB groups were 90% (95% CI 80-96%) and 78% (95% CI 65-87%), and resection rates were 97% (95% CI 87-99%) and 67% (95% CI 54-79%) including several children undergoing liver transplantation. For SR-HB patients, 3-year overall and progression-free survivals were 91% (+/-7%) and 89% (+/-7%) and for the HR-HB group 53% (+/-13%) and 48% (+/-13%), respectively. The short-term toxicity of these regimens was acceptable, with no toxic deaths. A treatment strategy based on CDDP monotherapy and surgery thus appears effective in SR-HB but, despite CT intensification, only half of the HR-HB patients are long-term survivors. For SR-HB patients, the efficacy of CDDP monotherapy and the CDDP/DOXO ('PLADO') combination are now being compared in a prospective randomised trial (SIOPEL 3).