Single-Isocenter Multitarget Stereotactic Radiosurgery is Associated with Less Long-Term Radionecrosis Compared with Gamma Knife for Multiple Brain Metastasis

Abstract

Radiosurgery and linear accelerator-based stereotactic radiosurgery (SRS) are both effective and offer excellent local control in the management of multiple brain metastases. When compared to conventional SRS, single-isocenter multitarget (SIMT) linear accelerator-based SRS offers superior treatment time and patient convenience, at the theoretical expense of dose conformity. To date, the clinical efficacy of SIMT and GK SRS has not been compared. We studied overall survival (OS) and radionecrosis (RN) in a multi-institutional cohort of patients treated with SIMT technique at our institution and GK at a collaborating institution. We identified patients treated with either SIMT or GK who were treated to ≥ 2 lesions and had available follow up. To account for confounders, cox proportional hazards and 1:1 nearest neighbor propensity score matching (caliper=0.1) were performed on the basis of sex, year of treatment, primary tumor site, number of lesions, tumor volume, prior whole brain radiation therapy, and use of concurrent immunotherapy. We used Kaplan-Meier curves to estimate freedom from RN using the matched data. We identified 361 patients who were treated in 523 courses to a total 2605 lesions (SIMT: 1014, GK: 1591). Median follow up was 13.1 months. There was no difference in OS between SIMT and GK (hazard ratio [HR] = 1.01; 95% confidence interval [CI] 0.77 - 1.33; p = 0.92). There was no significant difference in freedom from any grade RN (HR = 1.45; 95% CI 0.78 - 2.70; p = 0.24) or grade 2 or higher (grade 2+) RN (HR = 1.59; 95% CI 0.78 - 3.23; p = 0.20). Actuarial 6-, 12-, and 18-month freedom from any grade RN were 93.8%, 91.7%, and 89.5% for SIMT and 93.8%, 90.0%, and 84.9% for GK, respectively. Six-, 12-, and 18-month freedom from grade 2+ RN were 94.8%, 92.7%, and 90.4% for SIMT and 94.8%, 90.8%, 85.0% for GK, respectively. Crude rates of lesional any grade RN were 1.8% and 2.8% for SIMT and GK, respectively. On multivariate analysis, treatment with GK was associated with worse freedom from any grade RN (HR = 2.75; 95% CI 1.33 - 5.71; p = 0.006) and grade 2+ RN (HR = 3.18; 95% CI 1.33 - 7.61; p = 0.009). After propensity score matching, there were 140 treatment courses in each group. Standardized mean differences were <0.2 for all variables. SIMT was associated with better freedom any grade (log rank p = 0.04) and grade 2+ RN (log rank p = 0.03). In this multi-institutional study, we found no difference in OS or rates of RN between patients treated with GK or SIMT SRS. After adjusting for confounders, we found SIMT was associated with better freedom from RN. While these results should be interpreted with caution due to the potential for institutional selection bias, it appears that SIMT SRS has at least comparable rates of RN when compared to GK. These findings should be validated in an independent cohort.