Effect of Margin Status on 10-Year Local Recurrence and Survival Outcomes in a Large Population-Based Analysis of Women Treated With Breast-Conserving Therapy

Abstract

A 2014 SS0/ASTRO consensus statement suggests “no ink on tumor” is a sufficient surgical margin for invasive breast cancer treated with breast conserving surgery (BCS). Whether long-term outcomes in cases with close margin <2mm are inferior to cases with clear margins >2mm remains controversial. This study evaluates 10-year outcomes by margin status in a population-based cohort of women treated with BCS and adjuvant radiotherapy (RT). Subjects were 10,863 women with pT1-T3, any N, M0 invasive breast cancer referred from July 1, 2001 to Dec. 31, 2011, an era in which re-excision was generally used for close or positive margins. Data on margin location was available in a subset of 2381 cases referred from Jan. 1, 2010 to Dec. 31, 2011. All women underwent BCS and adjuvant whole breast ± boost RT (boost RT generally used for close or positive final margins or age < 50). Ten-year local recurrence (LR), and breast cancer-specific survival (BCSS) were examined using Kaplan-Meier (KM) and competing risk analysis in cohorts with negative (>2mm; n=9241, 85%), close (≤2mm; n=1310, 12%), or positive (tumor touching ink; n=312, 3%) margins. Competing risk multivariable analysis (MVA) was performed using Fine and Gray modeling. Matched-pair analysis compared LR and BCSS between close/positive cases matched to negative margin controls. Median follow-up was 8 years. Systemic therapy was used in 87% of patients. Boost RT was used in 34.1%, 76.9% and 79.5% of patients with negative, close, and positive margins, respectively. In the negative, close, and positive margin cohorts, 10-year KM outcomes were: LR 2.8%, 4.3%, and 4.0% and BCSS 93.7%, 91.5%, and 87.4%, respectively. On competing risk MVA, close margins were not significantly associated with either increased LR (HR 1.25, 95% CI 0.79-1.97, p=0.35) or reduced BCSS (HR 1.25, 95% CI 0.98-1.58, p=0.07) relative to negative margins. Age <45, grade 3, ≥4 positive nodes, and no systemic therapy were factors associated with both increased LR and reduced BCSS. On MVA, use of boost RT was not associated with LR (p=0.92) or BCSS (p=0.30). On matched-pair analysis, close margin cases had similar LR (p=0.11) and BCSS (p=0.10) compared to negative margin controls. Among the 554 cases with close or positive margins from 2010-2011 in whom data on margin location were available, 57% were anterior or posterior, so these findings may not be broadly generalizable to cohorts in which the majority of close margins are potentially re-resectable. In patients treated with breast-conserving therapy for invasive breast cancer, close margins were not associated with increased LR or reduced BCSS compared to negative margins. This population-based analysis corroborates the SSO/ASTRO consensus statement.