To assess the rate of tumor response and activity of carboplatin in stabilizing the growth of progressive low-grade gliomas.

Eligible patients received carboplatin 560 mg/m(2) intravenously every 4 weeks for 1 year after maximum tumor response or until disease progression or unacceptable toxicity.

Between October 1993 and October 2000, 81 children (median age, 79 months; range, 6 to 204) were enrolled onto this study. Patients received a median of 11 cycles of carboplatin (range, one to 29). Median follow-up from the time of enrollment was 55 months (range, 10 to 93). The overall objective response (complete response [CR] + partial response [PR] + minor response [MR]) and disease stabilization (CR + PR + stable disease + MR) rates to carboplatin treatment were 28% (95% confidence interval [CI], 18% to 38%) and 85% (95% CI, 74% to 93%), respectively. Eleven and 14 patients suffered progressive disease on study and after stopping therapy, respectively. Toxicity was predominantly myelosuppression and included grade 3/4 neutropenia in 56 patients and grade 3/4 thrombocytopenia in 40 patients. The 3-year failure-free survival (FFS) and overall survival (OS) for all patients were 64% (95% CI, 54% to 76%) and 84% (95% CI, 76% to 93%), respectively. Patients with diencephalic tumors had inferior FFS and OS compared with those with tumor at other sites (38% v 74% for FFS, P =.011; 54% v 91% for OS, P =.004). Neurofibromatosis type 1 patients with progressive low-grade glioma had a significantly better OS (95% v 80%; P =.052).

Carboplatin, in the schedule used in this study, produced disease stabilization or improvement in a majority of children with progressive low-grade glioma, with manageable toxicity. Improved treatment strategies are particularly required for patients with diencephalic tumors.