VALIDITY OF THE GRADED PROGNOSTIC ASSESSMENT-DERIVED INDEX TO PREDICT BRAIN-METASTATIC PATIENTS' SURVIVAL AFTER GAMMA KNIFE RADIOSURGERY

Purpose: To appraise whether the graded prognostic assessment (GPA)-derived index is valid for selecting patients with brain metastases for Gamma Knife (GK) radiosurgery. Methods and Materials: A total of 56 consecutive patients in recursive partioning analysis (RPA) Class I (n = 19, 34%) and II (n = 37, 66%) formed the basis of this retrospective study. Their mean age was of 57 years with mean Karnofsky performance score of 77. Primary cancers stemmed mainly from the lungs (59%). A total of 45 patients (80%) harbored multiple tumors. The mean clinical follow-up period was 9 months. Results: Kaplan-Meier analysis demonstrated that the overall median survival time (MST) for the whole series was 11.5 months: 16.5 vs. 6.5 months for RPA class I and II (p = 0.017). Multivariate Cox analysis revealed that female patients and a pre-GK good functional state were favorable prognostic factors. The favorable MST was in patients with a GPA score of 3 to 4 (17 months) followed by a GPA score of 2 to 2.5 (11 months) and GPA score 0 to 1.5 (6.5 months), but without statistical differences (p = 0.413) in between. A modified index (MGPA) is proposed with gender as a cofactor, then there existed a distinct survival differences (p = 0.028) between patients with an MGPA score of 3.5 to 5 (15 months) and with an MGPA score of 0 to 3 (7 months). In addition, the original GPA index failed to imply the difference of MST in patients with lung origin. Conclusions: The GPA-derived index is not applicable to our set of patients for comparing their survival after GK radiosurgery. The gender of the patients is a suggested cofactor to further refine the greater prognostic accuracy of the GPA index. (C) 2010 Elsevier Inc.