Local tumor progression treated with Gamma Knife radiosurgery: Differences between patients with 2–4 versus 5–10 brain metastases based on an update of a multi-institutional prospective observational study (JLGK0901)

OBJECTIVE The Japanese Leksell Gamma Knife (JLGK)0901 study proved the efficacy of Gamma Knife radiosurgery (GKRS) in patients with 5–10 brain metastases (BMs) as compared to those with 2–4, showing noninferiority in overall survival and other secondary endpoints. However, the difference in local tumor progression between patients with 2–4 and those with 5–10 BMs has not been sufficiently examined for this data set. Thus, the authors reappraised this issue, employing the updated JLGK0901 data set with detailed observation via enhanced MRI. They applied sophisticated statistical methods to analyze the data. METHODS This was a prospective observational study of 1194 patients harboring 1–10 BMs treated with GKRS alone. Patients were categorized into groups A (single BM, 455 cases), B (2–4 BMs, 531 cases), and C (5–10 BMs, 208 cases). Local tumor progression was defined as a 20% increase in the maximum diameter of the enhanced lesion as compared to its smallest documented maximum diameter on enhanced MRI. The authors compared cumulative incidence differences determined by competing risk analysis and also conducted propensity score matching. RESULTS Local tumor progression was observed in 212 patients (17.8% overall, groups A/B/C: 93/89/30 patients). Cumulative incidences of local tumor progression in groups A, B, and C were 15.2%, 10.6%, and 8.7% at 1 year after GKRS; 20.1%, 16.9%, and 13.5% at 3 years; and 21.4%, 17.4%, and not available at 5 years, respectively. There were no significant differences in local tumor progression between groups B and C. Local tumor progression was classified as tumor recurrence in 139 patients (groups A/B/C: 68/53/18 patients), radiation necrosis in 67 (24/31/12), and mixed/ undetermined lesions in 6 (1/5/0). There were no significant differences in tumor recurrence or radiation necrosis between groups B and C. Multivariate analysis using the Fine-Gray proportional hazards model revealed age < 65 years, neurological symptoms, tumor volume ≥ 1 cm³, and prescription dose < 22 Gy to be significant poor prognostic factors for local tumor progression. In the subset of 558 case-matched patients (186 in each group), there were no significant differences between groups B and C in local tumor progression, nor in tumor recurrence or radiation necrosis. CONCLUSIONS Local tumor progression incidences did not differ between groups B and C. This study proved that tumor progression after GKRS without whole-brain radiation therapy for patients with 5–10 BMs was satisfactorily treated with the doses prescribed according to the JLGK0901 study protocol and that results were not inferior to those in patients with a single or 2–4 BMs.