Determinants of survival in patients with brain metastases from cutaneous melanoma

BACKGROUND: This retrospective study aimed to identify prognostic factors in patients with brain metastases from cutaneous melanoma. METHODS: In all, 265 patients under regular screening according to valid national surveillance guidelines were included in the study. Kaplan-Meier analyses were performed to estimate and to compare overall survival. Cox modeling was used to identify independent determinants of the overall survival, which were used in explorative classification and regression tree analysis to define meaningful prognostic groups. RESULTS: In total, 55.5% of our patients presented with two or less brain metastases, 82.6% had concurrent extracranial metastasis and 64% were asymptomatic and diagnosed during surveillance scans. In all, 36.7% were candidates for local treatment (neurosurgery or stereotactic radiosurgery (SRS)). The median overall survival of the entire collective was 5.0 months (95% confidence interval: 4.3-5.7). Favourable independent prognostic factors were: normal pre-treatment level of serum lactate dehydrogenase (P < 0.001), administered therapy (neurosurgery or SRS vs other, P = 0.002), number of brain metastases (single vs multiple, P = 0.032) and presence of bone metastasis (false vs true, P = 0.044). Three prognostic groups with significantly different overall survival were identified. Candidates for local treatment (group I) had the longer median survival (9 months). Remaining patients could be further classified in two groups on the basis of serum lactate dehydrogenase. CONCLUSION: Applied treatment and serum lactate dehydrogenase levels were independent predictors of survival of patients with brain metastases from cutaneous melanoma. Patients receiving local therapy have overall survival comparable with general stage IV melanoma patients. British Journal of Cancer (2010) 102, 1213-1218. doi: 10.1038/sj.bjc.6605622 www.bjcancer.com Published online 6 April 2010 (C) 2010 Cancer Research UK