Patients with melanoma brain metastases (MBM) have poor prognosis, albeit advances in locoregional and systemic treatments. The melanoma-specific Graded Prognostic Assessment (GPA) effectively stratifies survival for patients with MBM. Nevertheless, lactate dehydrogenase (LDH), a well known prognostic factor for patients with melanoma, is not represented in the GPA scores and might add prognostic information for patients with MBM. In this study, 150 consecutive patients with MBM were retrospectively analyzed with the aim of evaluating independent prognostic factors for MBM patients, including LDH. Furthermore, we implemented a disease-specific prognostic score and estimated survival according to treatment modalities. On the basis of multivariable Cox regression analyses, six prognostic factors (age, BRAF status, number of MBM, number of extracranial metastatic sites, performance status, and LDH level) resulted statistically significant in terms of survival and were combined in a prognostic score to stratify patients in distinct prognostic groups ( P < 0.0001). Among treatment modalities, a multimodal approach with stereotactic radiosurgery or neurosurgery associated with systemic therapy showed the best outcome (median overall survival: 12.32 months, 95% confidence interval, 7.92-25.30). This is the first study to demonstrate that LDH has independent prognostic value for patients with MBM and might be used to improve prognostic stratification, albeit external validation is mandatory. Survival of patients with MBM is affected by both disease-specific risk factors and treatment modalities, with locoregional treatments associated with better outcomes.
