A validation study of a new prognostic index for patients with brain metastases: the Graded Prognostic Assessment

Abstract

Object The purpose of this study was to validate a new prognostic index for patients with brain metastases. This index, the Graded Prognostic Assessment (GPA), is based on an analysis of 1960 patients whose data were extracted from the Radiation Therapy Oncology Group (RTOG) database. The GPA is based on 4 criteria: age, Karnofsky Performance Scale score, number of brain metastases, and the presence/absence of extracranial metastases. Each of the 4 criteria is given a score of 0, 0.5, or 1.0, so the patient with best prognosis would have a GPA score of 4.0. Methods Between April 2005 and December 2006, 140 eligible patients with brain metastases were treated at the Gamma Knife Center at the University of Minnesota. The GPA score was calculated for each patient, and the score was then correlated with survival. Survival duration was calculated from the date treatment began for the brain metastases. Eligibility criteria included patients treated with whole-brain radiation therapy, stereotactic radiosurgery, or both. Results The median survival time in months observed in the RTOG and Minnesota data by GPA score was as follows: GPA 3.5–4.0, 11.0 and 21.7; GPA 3.0, 8.9 and 17.5; GPA 1.5–2.5, 3.8 and 5.9; and GPA 0–1.0, 2.6 and 3.0, respectively. Conclusions The University of Minnesota data correlate well with the RTOG data and validate the use of the GPA as an effective prognostic index for patients with brain metastases. Clearly, not all patients with brain metastases have the same prognosis, and treatment decisions should be individualized accordingly. The GPA score does appear to be as prognostic as the RPA and is less subjective (because the RPA requires assessment of whether the primary disease is controlled), more quantitative, and easier to use and remember. A multiinstitutional validation study of the GPA is ongoing.

Poor Outcomes after Whole Brain Radiotherapy in Patients with Brain Metastases: Results from an International Multicentre Cohort Study

“…Due to the lack of explanatory prognostic variables, the reasons for this observation were not clear. The published literature supports diagnosis or primary sitespecific prognostic factors, the most well established being the graded prognostic assessment (GPA) model, originally developed on the basis of four prognostic factors of age, Karnofsky performance status (KPS), number of brain metastases and the presence or absence of extracranial disease [3]. In the most recent update by Sperduto et al [23], a fifth prognostic category based on primary diagnosis was added, with statistically significant differences between brain metastases arising from lung cancer (non-small cell and small cell), melanoma, renal cell carcinoma, breast cancer and gastrointestinal malignancies.…”

“…First, its retrospective nature meant there was limited documentation of known prognostic variables in a cohort of brain metastases patients, including performance status, intra-and extra-cranial disease burden and information on other treatment modalities [1,3] that form the basis of established prognostic models such as the RTOG RPA and GPA categories. Furthermore, there was limited information on neurological symptom status, the response to steroids and other quality of life outcomes.…”

Hepatocellular carcinoma specific graded prognostic assessment can predict outcomes for patients with brain metastases from hepatocellular carcinoma

“…Patients with a GPA of 4.0 would have the best prognosis. The GPA scoring system was established based on a database of 1,960 patients with brain metastases from four RTOG prospective clinical trials, and validated in different patient cohorts [16].…”

A new paradigm in treatment of brain metastases

“…First developed from analysis of 1,960 patients from 5 RTOG randomized trials, the GPA yielded 4 prognostic groups ( Table 2). The GPA was then validated against an institutional database [52]. Since different cancer histologies have varied natural histories and therapeutic options, it suggests that prognostic indices could be improved by tailoring the to specific diagnoses.…”