OBJECTIVEThe results of 3-stage Gamma Knife treatment (3-st-GK-Tx) for relatively large brain metastases have previously been reported for a series of patients in Chiba, Japan (referred to in this study as the C-series). In the current study, the authors reappraised, using a competing risk analysis, the efficacy and safety of 3-st-GK-Tx by comparing their experience with that of the C-series.METHODSThis was a retrospective cohort study. Among 1767 patients undergoing GK radiosurgery for brain metastases at Mito Gamma House during the 2005–2015 period, 78 (34 female, 44 male; mean age 65 years, range 35–86 years) whose largest tumor was > 10 cm3, treated with 3-st-GK-Tx, were studied (referred to in this study as the M-series). The target volumes were covered with a 50% isodose gradient and irradiated with a peripheral dose of 10 Gy at each procedure. The interval between procedures was 2 weeks. Because competing risk analysis had not been employed in the published C-series, the authors reanalyzed the previously published data using this method.RESULTSThe overall median survival time after 3-st-GK-Tx was 8.3 months (95% CI 5.6–12.0 months) in the M-series and 8.6 months (95% CI 5.5–10.6 months) in the C-series (p = 0.41). Actuarial survival rates at the 6th and 12th post–3-st-GK-Tx months were, respectively, 55.1% and 35.2% in the M-series and 62.5% and 26.4% in the C-series (HR 1.175, 95% CI 0.790–1.728, p = 0.42). Cumulative incidences at the 12th post–3-st-GK-Tx, determined by competing risk analyses, of neurological deterioration (14.2% in C-series vs 12.8% in M-series), neurological death (7.2% vs 7.7%), local recurrence (4.8% vs 6.2%), repeat SRS (25.9% vs 18.0%), and SRS-related complications (2.3% vs 5.1%) did not differ significantly between the 2 series.CONCLUSIONSThere were no significant differences in post–3-st-GK-Tx results between the 2 series in terms of overall survival times, neurological death, maintained neurological status, local control, repeat SRS, and SRS-related complications. The previously published results (C-series) are considered to be validated by the M-series results.