A modified melanoma-molGPA scoring model: assessment of survival after and efficacy of different radiotherapy modalities in patients with melanoma brain metastases

Abstract

Abstract Purpose Patients with malignant melanoma brain metastases (MBMs) have poor prognoses. For MBMs, the Melanoma-molGPA is the most widely used predictive score, but its predictive value remains uncertain in patients fully treated with radiotherapy. We identified MBMs prognostic factors and modified the prognostic scoring model. Methods We retrospectively analyzed patients diagnosed with MBMs between December 2010 and November 2021 for prognostic factors influencing overall survival (OS) by univariate and multivariate analyses. Nomogram plots were based on Cox regression modeling. We evaluated overall survival (OS) using Kaplan–Meier survival curves and log-rank tests. Results The median OS (mOS) was 7.9 months. On multivariate analysis, BRAF mutation status (p < 0.001), number of brain metastases (BM) (p < 0.001), presence of liver metastases (p < 0.001), brain metastases with a midline shift (p = 0.003), Karnofsky Performance Score (p = 0.02), and lymphocyte-to-monocyte ratio (p < 0.0001) were independent OS predictors. These were incorporated into a modified risk-stratification model. Overall, whole-brain radiotherapy (WBRT) did not significantly affect mOS (mOS, 6.89 vs. 8.83 months; p = 0.07). After risk stratification using our model, WBRT resulted in no significant survival benefit in the low-risk group (mOS 10.07 vs. 13.1 months; p = 0.71) but significantly worse prognosis in the high-risk group (mOS, 2.37 vs. 6.92 months; p = 0.026). Conclusion We propose a modified model that accurately distinguishes the prognosis of patients with MBMs and guides decision-making for radiotherapy. Based on this novel model, WBRT should be cautiously selected for high-risk patients.