A Multi-institutional Validation Study of a New Prognostic Index (graded Prognostic Assessment, GPA) for Patients with Brain Metastases

A new prognostic index for patients with brain metastases was recently published (Int J Radiat Oncol Biol Phys 70:510-514; 2008), based on an analysis of 1,960 patients in the Radiation Therapy Oncology Group (RTOG) database. This index, the Graded Prognostic Assessment (GPA) uses four criteria (age, KPS, number of brain metastases, and whether extracranial metastases are present or absent) and scores each with a 0, 0.5, or 1.0 value. The patient with the best prognosis would have a GPA of 4.0. The purpose of this study is to validate the GPA by retrospectively correlating the GPA with median survival time (MST) at multiple institutions. A secondary objective was to determine if there is a difference in survival in patients treated with: a) Whole Brain Radiation (WBRT) alone; b) Stereotactic Radiosurgery (SRS) alone; c) WBRT + SRS. A database of 1880 patients treated between July 1985 and August 2007 was created retrospectively from raw data submitted by six institutions. Survival time was calculated from the date treatment began for the brain metastases. The GPA score was calculated for each patient in the group overall (1880), the group without surgery (1598), and the group without salvage therapy or surgery (1201). Salvage therapy was defined as any treatment >2 months from the start of brain metastasis treatment. The GPA was correlated with survival via the log-rank test and the multivariate Cox regression model. These findings were compared with the results from the RTOG database. Survival by initial treatment was assessed. The MST for the RTOG data (n = 1960), the validation data without salvage therapy or surgery (n = 1201), and the validation data overall (n = 1880) was 11.0, 14.3, and 18.2 mo for GPA 3.5-4.0; 8.9, 13.3, and 14.9 mo for GPA 3.0; 3.8, 7.1, and 10.1 mo for GPA 1.5-2.5; and 2.6, 3.0, and 4.3 mo for GPA 0-1.0, respectively. There was a significant difference in survival between each GPA group in the validation data (p < 0.0001). In the non-surgical group (1598), MST for WBRT + SRS, SRS alone, and WBRT alone was 14.8, 8.4, and 2.9 months, respectively. WBRT + SRS was significantly better than SRS alone (p < 0.0001) or WBRT alone (p < 0.0001). 1) The validation data agree with the RTOG data in that the higher the GPA, the longer the survival. 2) There are significant differences in survival between each GPA group (p < 0.0001). 3) The improvement in survival between the RTOG database and the validation database is probably attributable to selection bias and advances in treatment. 4) For the non-surgical group, WBRT + SRS as initial treatment was superior to SRS alone or WBRT alone. 5) This study validates the GPA as a useful prognostic index for future clinical trials and for clinical decision-making regarding which patients warrant aggressive treatment.