Mature Local Control and Radiation Necroses Outcomes Specific to Brain Metastases Treated with Salvage Stereotactic Radiosurgery for Radiosurgical Failures

Abstract

<h3>Purpose/Objective(s)</h3> The management of locally recurrent brain metastases is a clinical challenge. A second course of radiosurgery may be delivered as the primary salvage treatment or as postoperative radiosurgery to the surgical cavity if the recurrence resected. We report our institutional experience of repeat radiosurgery for locally recurrent brain metastases, with a focus on local control and radionecrosis (RN) outcomes. <h3>Materials/Methods</h3> We report a single institution retrospective cohort study specific to patients with recurrent brain metastases treated with repeat radiosurgery. Clinical and dosimetric details were collected. Overall survival was estimated using the Kaplan Meier method. Local recurrence and RN rates were estimated using the Aalen-Johnson method, with death from any cause as the competing risk for both endpoints. Univariable competing risk regression using Fine and Gray's methods were performed, and subsequent multivariable (MVA) regression based on a priori variable selection generated final adjusted models. <h3>Results</h3> We identified 96 patients and 130 brain metastases re-treated with radiosurgery (either single fraction or hypofractionation between July 2010 to April 2020. The most common primary sites were lung (N = 44, 33.8%), breast, (N = 38, 29.2%), and melanoma (N = 26, 20%). More than half of the retreated lesions were postoperative cavities (N = 68, 52.3%). Re-treatment was most commonly delivered with 25 Gy in 5 fractions (N = 77, 59.2%) followed by 27.5 Gy in 5 fractions (N = 23, 17.7%). The mean BED₁₀ was 39.71 Gy (standard deviation [SD] 6.16) with a mean treatment volume of 14.94 cm³ (SD, 20.15). Nearly half (N = 64, 49.2%) of lesions were treated in the absence of systemic therapy, while 43.8% (N = 57) and 6.9% (N = 9) were treated while receiving targeted therapy/immunotherapy and chemotherapy, respectively. The median time between treatment courses was 13.7 months (IQR 10.88). With a median follow up of 34.9 months (interquartile range [IQR] 25-41.9), the 1 and 2-year overall survival rates were 59.8% (95% confidence interval [95% CI] 50.4-70.9) and 36.2% (95% CI, 27.2-48.2), respectively. The risk of local failure at 1 and 2 years was 26.2% (95%CI, 18.5-34) and 28% (95%CI, 20.1-35.9), respectively. MVA identified increasing target volume (hazard ratio [HR] 1.03; 95% CI, 1.02-1.04), and a shorter time-interval between radiosurgery courses (HR 0.29; 95% CI, 0.19-0.45) to be associated with worse local control. The crude incidence of RN was 18.5% (N = 24), of which 8 events (6.15%) were symptomatic. The 1 and 2-year rates of RN were 17.6% (95% CI, 10.9-34) and 19.3% (95%CI, 12.3-35.9), respectively. <h3>Conclusion</h3> Repeat radiosurgery for locally recurrent brain metastases results in good local control with a moderate risk of RN. Larger target volume and a shorter timeframe between radiosurgery treatments are associated with worse control.