Single versus Multifraction Stereotactic Radiosurgery for Large Brain Metastases: An International Meta-analysis of 24 Trials.

Abstract

Multifraction (MF) stereotactic radiosurgery (SRS) purportedly reduces radionecrosis risk over single-fraction (SF) SRS in the treatment of large brain metastases. The purpose of the current work is to compare local control (LC) and radionecrosis rates of SF-SRS and MF-SRS in the definitive (SF-SRS_D and MF-SRS_D) and postoperative (SF-SRS_P and MF-SRS_P) settings. Population, Intervention, Control, Outcomes, Study Design/Preferred Reporting Items for Systematic Reviews and Meta-analyses and Meta-analysis of Observational Studies in Epidemiology guidelines were used to select articles in which patients had "large" brain metastases (Group A: 4-14 cm³, or about 2-3 cm in diameter; Group B: >14 cm³, or about >3 cm in diameter); 1-year LC and/or rates of radionecrosis were reported; radiosurgery was administered definitively or postoperatively. Random effects meta-analyses using fractionation scheme and size as covariates were conducted. Meta-regression and Wald-type tests were used to determine the effect of increasing tumor size and fractionation on the summary estimate, where the null hypothesis was rejected for P < .05. Twenty-four studies were included, published between 2008 and 2017, with 1887 brain metastases. LC random effects estimate at 1 year was 77.6% for Group A/SF-SRS_D and 92.9% for Group A/MF-SRS_D (P = .18). LC random effects estimate at 1 year was 77.1% for Group B/SF-SRS_D and 79.2% for Group B/MF-SRS_D (P = .76). LC random effects estimate at 1 year was 62.4% for Group B/SF-SRS_P and 85.7% for Group B/MF-SRS_P (P = .13). Radionecrosis incidence random effects estimate was 23.1% for Group A/SF-SRS_D and 7.3% for Group A/MF-SRS_D (P = .003). Radionecrosis incidence random effects estimate was 11.7% for Group B/SF-SRS_D and 6.5% for Group B/MF-SRS_D (P = .29). Radionecrosis incidence random effects estimate was 7.3% for Group B/SF-SRS_P and 7.5% for Group B/MF-SRS_P (P = .85). Metaregression assessing 1-year LC and radionecrosis as a continuous function of increasing tumor volume was not statistically significant. Treatment for large brain metastases with MF-SRS regimens may offer a relative reduction of radionecrosis while maintaining or improving relative rates of 1-year LC compared with SF-SRS. These findings are hypothesis-generating and require validation by ongoing and planned prospective clinical trials.