Carboplatin in childhood medulloblastoma/PNET: feasibility of an in vivo sensitivity test in an "up-front" study.

Sixteen patients with high risk MB/PNET at diagnosis were included in a pilot study employing carboplatin (CBDCA) as a single drug prior to conventional therapy. The main goal of the study was to identify in a short-term trial a significant response that would predict further response to CBDCA in the single patient. Exploration of CBDCA activity was focused on response after the first course as compared to the response following the second course. A course consisted of CBDCA 600 mg/m2 on days 1 and 2 administered in a 1 h infusion to be repeated 3-4 weeks later. After two cycles we observed 1 CR and 9 PR, that is a 62% response rate. The first course resulted in 5 PR, 5 MR, 5 SD, and 1 PD; after the subsequent course in all responding patients, response persisted or improved whereas in no patient with SD any improvement was observed. The correlation of response to the first course with response to the second course was statistically significant (P = 0.0009). The main toxicity of the single course was hematologic and consisted of rapidly reversible grade 3-4 neutropenia and thrombocytopenia in 94% of patients. Pharmacokinetic studies showed a very limited interpatient variability of both Cmax 57.6 +/- 9.9 micrograms/ml) and AUC (15.3 +/- 1.5 mg/ml.min) of free CBDCA, which eliminates an important variable in the evaluation of response. In conclusion, this "in vivo test" appears effective, reasonably safe, and reproducible in identifying patients likely to benefit from CBCDA: after a period of time as short as 3-4 weeks following the first course, multidrug chemotherapy including CBDCA may be employed in the responding patients, whereas an alternative regimen would be indicated in the non-responding patients.