Central nervous system failure in melanoma patients: results of a randomised, multicentre phase 3 study of temozolomide- and dacarbazine- based regimens.

<bold>Background: </bold>This study compared the central nervous system (CNS) metastasis incidence between a temozolomide- and a dacarbazine-based regimen in untreated stage IV melanoma patients.<bold>Methods: </bold>A total of 150 patients were randomly assigned to receive either oral temozolomide (200 mg m(-2) per day; days 1-5) or intravenous dacarbazine (800 mg m(-2); day 1), in combination with intravenous cisplatin (75 mg m(-2); day 1) and subcutaneous interleukin-2 (3 MU twice daily; days 9-18), every 28 days (CTI and CDI).<bold>Results: </bold>A total of 149 patients were eligible for an intention-to-treat analysis (CTI: n=74, CDI: n=75). The 1-year cumulative CNS incidence failure was 20.6% for CTI and 31.1% for CDI (P=0.22). In all 24 patients in CTI (32%) and 34 (45%) in CDI developed CNS metastases; 31 patients died of early systemic progression, before CNS evaluation. Median survival time was 8.4 months in the CTI and 8.7 in the CDI arm; in patients with CNS metastases the median survival time was 13.5 months in the CTI and 11.5 in the CDI arm. No difference in toxicity was observed between the two arms.<bold>Conclusion: </bold>The incidence of CNS failures in metastatic melanoma was not significantly reduced and the clinical course was not modified substituting a dacarbazine-based regimen with a temozolomide-based regimen. Patients who developed CNS metastases did not have a worse prognosis than patients progressing in other sites and should not be excluded from new investigational studies. [ABSTRACT FROM AUTHOR]