Local Control after Resection and Adjuvant Radiosurgery Compared to Radiosurgery Alone for Brain Metastasis: Exploratory Analysis of Alliance NCCTG N107C

The phase III Alliance NCCTG N107C trial validated post-operative stereotactic radiosurgery (SRS) as an alternative to whole brain radiation (WBRT) for resected brain metastases (BM). No randomized trials have assessed the value of adding neurosurgery to radiosurgery (NS+SRS) for BM. This exploratory analysis evaluates the local control (LC) of BM after radiosurgery with and without surgical intervention within Alliance NCCTG N107C. NCCTG N107C randomized patients with 1-4 BM after resection of 1 lesion to either WBRT or post-operative SRS. Remaining intact lesions were treated by SRS regardless of randomization. We performed an unplanned exploratory analysis to compare LC of BM after SRS alone vs NS+SRS. Patients were analyzed per protocol. The primary endpoint was local recurrence (LR) of treated BM defined as time from randomization to event. LC was compared calculating cumulative incidence of LR with competing risk regression to adjust for prognostic factors and competing risk of death. Covariates include treatment, primary histology, size of intact BM or tumor cavity, number of BM, and total dose. Calculated hazard ratios for LR used discrete intervals of 0-3, >3-6, >6-9, and > 9 months (mo) to compare outcomes after NS+SRS or SRS alone. There were 184 evaluable patients. After excluding those receiving WBRT, 86 were included for analysis with a total 112 BM. Median follow-up was 12.8 mo. SRS alone was used for 23% (26 lesions) while 77% (86 lesions) were treated with NS+SRS. Median BM size was 0.6 cm (0.4-2.3 cm) and 2.8 cm (0.7-4.4 cm), respectively. Median SRS dose was 24 Gy (range: 15-24 Gy) for SRS alone and 17 Gy (range: 12-20 Gy) for NS+SRS. One-year cumulative LR incidence was 38.5% vs 50.6% (p = 0.52). Median time to LR was 9.3 versus 8.9 mo (p = 0.53). There was no difference in LR with death as a competing risk (50.0% vs. 38.1%; p = 0.53). LR by treatment did not vary with time as covariate at 3, 6, 9, and > 9 months (Table 1). When BM size was stratified as greater than and equal to or less than the median size within treatment group, there was no difference in cumulative 12-mo LR incidence. LR was higher for larger lesions within each treatment group but this did not reach significance (63.6% vs 49% and 43.6 vs 32.7%, respectively; p = 0.72). In this exploratory analysis of phase III data, there was no difference in local control for lesions receiving NS+SRS compared to SRS alone. Differences in baseline patient characteristics between arms limit this analysis and warrant prospective study. Support: U10CA180821, U10CA180882; U10CA180863; U10CA180868; NCT00377156Abstract 3743; TableUnadjusted risk for LR by modality with time as covariateTime (months)NS+SRS vs. SRS (ref) HRp-value0-30.410.253-61.130.836-91.080.93>90.680.58 Open table in a new tab