Can current prognostic scores reliably guide treatment decisions in patients with brain metastases from malignant melanoma?

PURPOSE We evaluated the performance of the new 4-tiered melanoma-specific graded prognostic assessment (GPA) score and the previously published general GPA score in patients with brain metastases from malignant melanoma managed with different approaches including best supportive care. MATERIALS AND METHODS Retrospective analysis of 51 patients. Compared with the original analysis of the melanoma-specific GPA score, these patients were more representative of the general population of patients with brain metastases from this disease. RESULTS The present data confirmed that both scores identify patients with favorable prognosis who might be candidates for focal treatments. However, survival in the 2 unfavorable prognostic subgroups defined by the melanoma-specific GPA was not significantly different. Median survival in the melanoma-specific GPA classes was 3.1, 3.7, 7.5, and 12.7 months. Karnofsky performance status (KPS) and serum lactatdehydrogenase (LDH) level significantly predicted survival. CONCLUSION In order to select the right patient to the right treatment and avoid overtreatment and suboptimal resource utilization in patients with very limited survival, improved prognostic tools are needed. The melanoma-specific GPA does not include extracranial disease extent or surrogate markers such as LDH. We suggest that a combination of KPS <70 and elevated LDH might better predict short survival than any of the GPA scores. This hypothesis should be confirmed in larger studies.