Does resection of cavity shave margins impact survival in breast cancer patients?

513 Background: Randomized controlled trials have shown that resection of cavity shave margins (CSM) halves the margin positivity and re-excision rate in breast cancer (bc) patients (pts), but the impact on overall (OS) and disease-free survival (DFS) remains unclear. Methods: Between Oct 21, 2011 and Nov 25, 2013, the SHAVE trial enrolled 235 pts in a single center study. Between Jul 28, 2016 and Apr 13, 2018, SHAVE2 enrolled 396 pts across 9 other US sites. Both studies were similar in their study criteria and methods. Pts ≥ 18 years of age, with planned partial mastectomy (PM) for stage 0-3 bc were eligible. Surgeons performed their standard PM with excision of selective margins as needed based on gross evaluation and specimen radiograph. Pts were then randomized intraoperatively to either “shave” or “no shave” arms. In the former, surgeons were instructed to take additional CSM from ≥4 faces of the cavity; in the latter, to close. Pts were followed for a median of 60.4 months (mo). Results: Across both studies (n = 631), the median pt age was 64 (range; 29-94), and the median invasive tumor size was 1 cm (range; 0-8 cm). Forty pts had no further disease at the time of PM either due to all disease being removed in the core biopsy or due to neoadjuvant chemotherapy. Before randomization, margin positivity was similar between the “shave” and “no shave” arms (37.8% vs. 34.8%, p = 0.457). After randomization, the “shave” group had a significantly lower positive margin rate (14.0% vs. 34.8%, p < 0.001) and re-excision rate (9.2% vs. 24.1%, p < 0.001). In the “shave” arm, 13.3% of those with negative margins prior to randomization had cancer found in the CSM. Pts in the “shave” arm were less likely to have positive final margins (after re-excision where performed; 6.3% vs. 15.8%, p < 0.001). Both groups had similar rates of radiation therapy (XRT) (82.8% vs. 81.8%, p = 0.831). Local recurrence rates were nearly identical at a median follow-up of 60.4 mo (1.3% vs. 1.6%, p = 1.000). 5-year DFS (91.9% vs. 92.6%, p = 0.820) and OS were not significantly different between groups (93.5% vs. 93.5%, p = 0.841). On Cox proportional hazard modelling, including randomization arm, final margin (after re-excision), and use of XRT, only the use of XRT significantly improved DFS (5 yr DFS 86.3% vs. 94.4%, HR: 0.47; 95% CI: 0.256-0.865, p = 0.015). Conclusions: While CSM reduced margin positivity and re-excision rates, it does not have an impact on DFS or OS at a median of 60.4 mo of surveillance; only the use of XRT significantly improved DFS. Although resection of CSM removed additional disease in 13.3% of margin-negative pts and was associated with a lower final positive margin rate, this did not translate into lower local recurrence rates. These results suggest that XRT may effectively manage small-volume residual disease that remains after PM. Clinical trial information: NCT01452399 ; NCT02772731 .