Are prognostic indices for brain metastases of melanoma still valid in the stereotactic era?

<bold>Background: </bold>Malignant melanoma brain metastases (MBM) are the third most common cause for brain metastases (BM). Historically Whole-brain radiotherapy (WBRT) was considered the goldstandard of treatment even though melanoma cells are regarded as very radioresistant. Therapeutic possibilities have fundamentally changed since the availability of stereotactic radiotherapy (SRT), where it is possible to apply high ablative doses in a very precise manner. In this work we analyze prognostic factors of overall survival (OS) after SRT in patients with MBM and evaluate the applicability of popular prognostic indices that mainly stem from the WBRT-era.<bold>Materials and Methods: </bold>This work is a retrospective analysis of OS of 80 malignant melanoma (MM) patients who received SRT for intracranial melanoma metastases between 2004 and 2014 who had not received prior treatment for MBM in terms of surgery or WBRT. Potential prognostic factors were analyzed using univariable and multivariable analysis. Existing prognostic scores [Graded Prognostic Assessment (GPA), Diagnosis-Specific-GPA (DS-GPA), Golden Grading System (GGS) and RADES] were calculated and tested using log-rank analysis.<bold>Results: </bold>Eighty patients, respectively 177 brain metastases, were irradiated. The median survival time from radiation was 7.06 months. Overall, GGS, GPA and DS-GPA were significant predictors of survival. The MM-specific index DS-GPA showed the best p-value but did not show adequate division when looking at the two intermediate risk subgroups. RADES did not show any statistically significant prognostic value. In univariable as well as in multivariable analyses a higher Karnofsky-Index, a single BM, and non nodular melanoma (NM) histology were positive predictors of survival.<bold>Conclusion: </bold>The existing prognostic scores do not seem to ideally fit for this special group of patients. Our results indicate that the histologic subtype of MM could add to the prognostic value of specialized future indices. [ABSTRACT FROM AUTHOR]