A validation study of graded prognostic assessment index for lung cancer patients with brain metastases.

Abstract

2070 Background: The Graded Prognostic Assessment (GPA) is a commonly used index based on 5 prior randomized trials performed by the RTOG (Protocols 7916, 8528, 8905, 9104, and 9508). The purpose of this study was to validate GPA index in a recent cohort of lung cancer patients with brain metastases (LBM) at a larger tertiary care center. Methods: Cleveland Clinic Brain Tumor and Neuro-Oncology Center's database was used to identify LBM treated in the recent era (2000-12). A proportional hazards model was used to assess overall survival (OS), measured from the date of diagnosis of brain metastases to death or last follow-up. Results: 490 LBM (250 males) median age 61 years (range 35-86) were included. Histology included small cell (64, 13%) adenocarcinoma (289, 59%), squamous (53, 11%), large cell (26, 5%) and unknown (55, 11%). The median number of brain metastases was 1 (range, 1-47). Karnofsky Performance Scale (KPS) was 90-100 in 187 (41%), 70-80 in 238 (52%) and <70 in 33 (7%). Extracranial Metastases was present in 327 patients (67%). OS was 14.1 months (95% C.I. 12.3-15.7). GPA for lung cancer is derived from KPS, the number of brain mets, the presence/absence of extracranial metastases, and age. GPA was 0-1.0 in 83 patients, 1.5-2 in 211, 2.5-3 in 135 and 3.5-4 in 18 patients. Although overall GPA was prognostic for survival (p<.0001), not all of the factors used to derive it were significant (Table). Factors noted to be prognostic for survival included primary controlled (p=<.0001), squamous cell histology (p=.0001), age (p=.0002), KPS (p=.0002). The new index divided the patients into unfavorable, intermediate, favorable that was prognostic for survival in this cohort (p<.0001). Conclusions: New index developed based on a revised set of independent prognostic factors (primary controlled, squamous cell histology, age and performance status) is proposed. [Table: see text]