Stereotactic radiosurgery for glioblastoma multiforme: report of a prospective study evaluating prognostic factors and analyzing long-term survival advantage.

Abstract

Prospective evaluation of the toxicity and efficacy of radiosurgery with external beam radiotherapy in the management of newly diagnosed glioblastoma. From 5/89 to 12/92, 31 out of 51 patients with glioblastoma multiforme underwent radiosurgery, in addition to 54 Gy in 1.8 Gy/fraction following biopsy (n = 12) or resection (n = 19). Eligibility required supratentorial glioblastoma, tumor not > 4 cm in > 1 axis, age > 18 years, and location > 1 cm from optic chiasm. Patient characteristics were: age 20-78 years (median = 57); 22 male, 9 female; Karnofsky score 20-90 (m = 70), and volume of 2.3-59.7 c.c. (m = 17.4). Eighteen patients were treated with 1 collimator, 5 with 2, 7 with 3, and 1 with 4; peripheral isodoses were 40-90% (m = 72.5) and minimum and maximum tumor dose ranges were 10-20 (m = 12) and 15-35 Gy (m = 18.75). Patients were followed clinically and radiographically every 8-12 weeks to analyze survival, quality of life, and toxicity. With a follow-up of 12-171 weeks, 8 out of 31 (26%) patients are alive. Median survival is 42 weeks. Twelve and 24-month actuarial survival are 38 and 28%. Comparison of the 2-year survival with previous Radiation Therapy Oncology Group patients was carried out using a nonparametric recursive partitioning technique and the observed vs. expected values are 28 vs. 9.7% (p < 0.05). Extent of resection and performance status were associated with improved survival in a multivariate analysis. No significant acute toxicity was encountered. Four patients (13%) developed clinically significant necrosis verified by biopsy or positron emission tomography scan at 9-59 weeks after radiosurgery. The improvement in median survival in broadly selected glioblastoma patients treated with radiosurgery is difficult to determine, but the 2-year survival may be superior. Future randomized trials of radiosurgery are recommended, and ad hoc use of this modality should be discouraged.