Linear accelerator–based stereotactic radiosurgery (SRS) is a treatment option for melanoma patients who have developed brain metastases. Few data are available on treatment of patients with ≥ 5 lesions. We sought to determine the effectiveness of SRS in patients with ≥ 5 melanoma brain metastases.
Jessica Freilich, MD, Nicholas Figura, BS, Kamran Ahmed, MD, Neha Patel, BS, Christian Thomas, BS, Siriporn Sarangkasiri, MS, Prakash Chinnaiyan, MD, Nikhil Rao, MD, Arnold Etame, MD, PhD; Moffitt Cancer Center
Purpose and Objectives: Linear accelerator–based stereotactic radiosurgery (SRS) is a treatment option for melanoma patients who have developed brain metastases. Few data are available on treatment of patients with ≥ 5 lesions. We sought to determine the effectiveness of SRS in patients with ≥ 5 melanoma brain metastases.
Materials and Methods: An analysis of patients with metastatic melanoma treated with SRS to ≥ 5 lesions in one treatment session was performed. Magnetic resonance imaging (MRI) scans were reviewed post-SRS to evaluate local control (LC). Disease progression by imaging was defined by the 2009 Response Evaluation Criteria in Solid Tumors (RECIST) criteria. Survival curves were calculated according to the Kaplan-Meier (KM) method from the date of brain metastases diagnosis or date of SRS. Univariate (UVA) and multivariate analysis (MVA) was performed by the Cox proportional hazards model.
Results: We identified 149 metastatic brain lesions treated in 28 patients. The median age of patients was 60.5 (range: 38–83 yr), and the majority (n = 24, 85.7%) had extracranial metastases. Four patients (14.3%) received previous whole-brain radiation therapy (WBRT), and 11 patients (39.3%) received previous SRS. The median planning target volume (PTV) was 0.34 cm3 (range: 0.01–12.5 cm3). Median follow-up was 6.3 months (range: 1–46 mo). At the time of treatment, 7% of patients were recursive partitioning analysis (RPA) class I, 89% was RPA class II, and 4% was RPA class III. The rate of local failure was 11.4%. KM local control estimates at 6 and 12 months were 91.3% and 82.2%, respectively. PTV volume ≥ 0.34 cm3 was a significant predictor of local failure on UVA (hazard ratio [HR] = 16.1; 95% confidence interval [CI], 3.2–292.6; P < .001) and MVA (HR = 14.8; 95% CI, 3.0–268.5; P < .001). Sixteen (57.4%) patients were noted to undergo distant failure in the brain, with a median time to failure of 3 months (range: 1–15 mo). Nine patients with distant brain failures received WBRT, and seven patients received additional SRS. Median overall survival (OS) was 9.4 and 7.6 months from the date of brain metastases diagnosis and date of SRS, respectively. The KM OS estimates at 6 and 12 months were 57.8% and 28.2%, respectively, from the time of SRS treatment. RPA class was a significant predictor of KM OS estimates from date of treatment (P = .02). Patients who did not receive WBRT after SRS treatment had decreased OS on MVA (HR = 3.5; 95% CI, 1.1–12.0; P = .03), and patients who did not receive WBRT prior to SRS had improved OS (HR = 0.11; 95% CI, 0.02–0.53; P = .007).
Conclusions: SRS to ≥ 5 lesions appears to be effective for selected patients with metastatic melanoma, offering excellent local control. This is particularly important for patients, as new targeted systemic agents are improving outcomes but still have limited efficacy within the central nervous system.