Background: Assessment of cancer- and host-related prognostic factors has a long tradition in patients with brain metastases. In continuation of large-scale studies performed by the Radiation Therapy Oncology Group (RTOG) in the United States, the 4-tiered diagnosis-specific graded prognostic assessment (DS-GPA) score has been developed. It stratifies patients with common primary tumours metastasizing to the brain (malignant melanoma, lung, breast, kidney and gastrointestinal cancers) into subgroups with different prognoses. However, many patients in the DS-GPA study were treated with surgical resection or radiosurgery (SRS). The present multi-institutional analysis examined for the first time whether DS-GPA is a valid score in European patients managed in routine clinical practice.
Material/Methods: This was a retrospective analysis of 412 patients with primary malignant melanoma, lung, breast, kidney or gastrointestinal cancers. Survival was evaluated in uni- and multivariate tests.
Results: DS-GPA significantly predicted survival and outperformed initial GPA, a score that is not diagnosis-specific. Median survival by DS-GPA strata (all 412 patients) was 2.7, 3.6, 7.0 and 11.3 months in the 4 groups with 0–1, 1.5–2, 2.5-3 and 3.5–4 points, respectively. The previously published survival data (median 7.2 months for all patients) could not be replicated in this cohort (median 3.6 months).
Conclusions: DS-GPA is a valid prognostic score that might improve shared decision making as well as patient stratification in prospective clinical trials.

Keywords: Proportional Hazards Models, Physician's Practice Patterns, Multivariate Analysis, Europe, Brain Neoplasms - therapy, Aged, 80 and over, Reproducibility of Results, young adult