External Validation of a Prognostic Score for Patients with Brain Metastases: Extended Diagnosis-Specific Graded Prognostic Assessment

• Published in: Oncology Research and Treatment
• Main Authors: Nieder, Carsten, Hess, Sebastian, Lewitzki, Victor
• Format: Article in Journal/Newspaper
• Language: English
• Published: S. Karger AG 2020
• Subjects: Norway; North Norway
• Online Access: http://dx.doi.org/10.1159/000506954; https://www.karger.com/Article/Pdf/506954

Purpose: The aim of our study was the external validation of an extended variant of the four-tiered diagnosis-specific graded prognostic assessment (DS-GPA) that includes more information about extracranial disease burden and blood test results, and predicts survival of patients with brain metastases. The extracranial DS-GPA (EC-GPA) includes serum albumin, lactate dehydrogenase, and number of extracranial organs involved. Originally, the score was developed in Germany. Patients and Methods: A retrospective analysis of 236 patients with brain metastases treated with primary whole-brain radiotherapy in North-Norway was performed (independent external validation cohort). Results: The four-tiered EC-GPA score showed good discrimination between all prognostic groups (log-rank test p < 0.05 for all pairwise comparisons). One-year survival was 0, 11, 30, and 100%, respectively. Median survival was 0.7 months (95% CI, 0.5–0.9) in the worst prognostic group, with a hazard ratio for death of 44.31 (95% CI, 5.78–339.50) compared to the best group. In the German database, the corresponding HR was 31.64 (median survival 0.4 months). The remaining hazard ratios in this validation study were 7.13 and 12.10, compared with 4.84 and 9.26 in the score development study. Conclusions: This study provides an independent validation of the EC-GPA, which was the best prognostic model for defining patients who did not benefit from radiation therapy of brain metastases in terms of overall survival in the original German study. The proposed modification of the established DS-GPA should undergo further validation in multi-institutional databases.