The optimal treatment strategy following local recurrence after stereotactic radiosurgery (SRS) remains unclear. While upfront SRS has been extensively studied, few reports focus on outcomes after retreatment. Here, we report the results following a second course of SRS for local recurrence of brain metastases previously treated with SRS.

Using institutional database, patients who received salvage SRS (SRS2) following in-field failure of initial SRS (SRS1) for brain metastases were identified. Radionecrosis and local failure were defined radiographically by MRI following SRS2. The primary endpoint was defined as the time from SRS2 to the date of all-cause death or last follow-up [overall survival (OS)]. The secondary endpoints included local failure-free survival (LFFS) and radionecrosis-free survival, defined as the time from SRS2 to the date of local failure or radionecrosis, or last follow-up, respectively.

Twenty-eight patients with 32 brain metastases were evaluated between years 2004 and 2015. The median interval between SRS1 and SRS2 was 9.7 months. Median OS was 22.0 months. Median LFFS time after SRS2 was 13.6 months. The overall local control rate following SRS2 was 84.4%. The 1- and 2-year local control rates are 88.3% (95% CI, 76.7-100%) and 80.3% (95% CI, 63.5-100%), respectively. The overall rate of radionecrosis following SRS2 was 18.8%. On univariate analysis, higher prescribed isodose line ( = 0.033) and higher gross tumor volume ( = 0.015) at SRS1 were associated with radionecrosis. Although not statistically significant, there was a trend toward lower risk of radionecrosis with interval surgical resection, fractionated SRS, lower total EQD2 (<50 Gy), and lack of concurrent systemic therapy at SRS2.

In select patients, repeat LINAC-based SRS following recurrence remains a reasonable option leading to long-term survival and local control. Radionecrosis approaches 20% for high risk individuals and parallels historic values.