Linac radiosurgery versus whole brain radiotherapy for brain metastases. A survival comparison based on the RTOG recursive partitioning analysis

Background and purpose: for patients with inoperable brain metastases, whole brain radiotherapy (WBRT) has been the standard treatment for decades. Radiosurgery is an effective alternative strategy, but has failed to show a substantial survival benefit so far. The prognostic factors derived from the RTOG recursive partitioning analysis (RPA) provide a framework that allows a nonrandomized comparison of the two modalities. Patients and methods: from 1991 to 1998, 117 patients with one to three previously untreated cerebral metastases underwent single-dose linac radiosurgery (median dose 20 Gy) without adjuvant WBRT. After radiosurgery, 26/117 patients (22%) had salvage WBRT, radiosurgery or neurosurgical resection of recurrent (4/117) and/or new (24/117) metastases. Survival of these patients was compared to a historical group of 138 patients with one to three lesions treated by WBRT (30-36 Gy/3-Gy fractions) from 1978 to 1991; only nine of these patients (7%) had salvage WBRT. All patients were classified into the three RPA prognostic classes based on age, performance score, and presence of extracranial tumor manifestations. Results: in RPA class I (Karnofsky performance score ≥ 70, primary tumor controlled, no other metastases, age < 65 years), radiosurgery resulted in a median survival of 25.4 months (n = 23, confidence interval [CI] 5.8-45.0) which was significantly longer than for WBRT (n = 9, 4.7 months, CI 3.8-5.5; p < 0.0001). In RPA class III (Karnofsky performance score < 70), no significant difference in survival between radiosurgery (n = 20, 4.2 months, CI 3.2-5.3) and WBRT (n = 68, 2.5 months, CI 2.2-2.8) was found. In RPA class II (all other patients), radiosurgery produced a small, but significant survival advantage (radiosurgery: n = 74, 5.9 months, CI 3.2-8.5, WBRT: n = 61, 4.1 months, CI 3.4-4.9; p < 0.04). Conclusion: radiosurgery in patients with one to three cerebral metastases results in a substantial survival benefit only in younger patients with a low systemic tumor burden when compared to WBRT alone. It cannot be excluded that this effect is partially caused by the available salvage options after radiosurgery. (orig.)