Use of the Radiation Therapy Oncology Group recursive partitioning analysis classification system and predictors of survival in 19 women with brain metastases from ovarian carcinoma

Abstract

Brain metastases are an uncommon complication in women with primary ovarian carcinoma; thus, little is known about whether the Radiation Therapy Oncology Group (RTOG) recursive partitioning analysis (RPA) prognostic classification system is valid in this patient population. From September 1985 to June 2002, 19 patients with brain metastases resulting from primary ovarian carcinoma underwent treatment at the Cleveland Clinic Foundation. The medical records of these patients were retrospectively reviewed. At the time of data analysis, all 19 women had died. The median age at diagnosis of primary ovarian carcinoma and brain metastasis was 51 and 54 years of age, respectively. Fifteen patients presented with a Karnofsky performance status (KPS) of 70 or higher. Seven patients had a single brain lesion and 12 had multiple lesions. All RTOG RPA prognostic classes were represented, with median survivals of 24.7, 8.9, and 2.6 months for Classes I, II, and III, respectively (P = 0.31). Patients who underwent surgical resection survived longer than those who did not (33.7 vs. 7.4 mos). The presence of multiple lesions was adversely related to survival on multivariate analysis (P = 0.03). Primary control was an important predictor of survival on multivariate analysis as well (P = 0.01) and was achieved in 15 of the 19 women. This is the first study to support the prognostic usefulness of the RTOG RPA classification for ovarian carcinoma patients with metastasis to the brain. The number of metastatic intracranial lesions should be included when determining the prognosis.