De-Escalation of Radiation Therapy in Patients with cT1-T2 (< 3 cm) N0 HER2+ Breast Cancer Treated With Neoadjuvant Systemic Therapy With Pathologic Complete Response at the Time of Breast Conserving Surgery

Abstract

The excellent survival and locoregional outcomes from the APT trial in predominantly stage I HER2+ breast cancer (BC) treated with upfront surgery (∼75% breast conservation surgery [BCS]) and adjuvant paclitaxel/trastuzumab has raised optimism that radiation (RT) may be de-escalated in this setting. Our initial analysis using the National Cancer Database (NCDB) instead demonstrated an association between RT omission and worse overall survival (OS) in these patients. Patients with HER2+ disease that receive neoadjuvant systemic therapy (NST) demonstrate great outcomes for patients with a pathologic complete response (pCR). This suggests that pCR may be used to de-escalate RT in early-stage HER2+ patients treated with NST+BCS. We used the NCDB to test the hypothesis that RT omission results in equivalent OS in cT1-2 (< 3 cm) N0 HER2+ patients treated with NST+BCS with pCR.We identified patients with cT1-T2 (< 3 cm) N0 HER2+ BC treated with anti-HER2 based NST and BCS +/- RT from 2013-2015. We identified factors associated with RT omission using logistic regression analysis. We evaluated OS in the subset of patients with pCR stratified by receipt of RT. We used a propensity score matched (PSM) analysis to compare OS between the groups. Patients were matched on clinical and demographic factors. The effect of RT omission was evaluated with a Cox proportional hazards model with robust standard errors to account for clustering in matched pairs.We identified 4,842 patients (4,505 RT; 337 no RT): 31.7% ER-/PR-; 48.5% with pCR. On multivariate analysis, age≥70 (OR = 2.88, P < 0.0001), Black race (OR = 1.43, P = 0.03), and use of single-agent cytotoxic chemotherapy (OR = 1.47, P = 0.02) were associated with RT omission. There was a trend towards RT omission in patients with pCR (OR = 1.21, P = 0.11). OS data was available in 1,581 (1,471 RT; 110 no RT) of the 2,348 patients with pCR. Median follow-up was 27.1 months (IQR = 19.7-37.0 months) with 14 deaths (3 RT; 11 RT omission). The 2-year OS was significantly worse for patients with RT omission (96.9% vs. 99.7%, P = 0.02). The PSM cohort consisted of 103 (of a possible 110) pairs of patients with 6 deaths (3 RT; 3 RT omission) and median follow-up 25.6 months (IQR, 17.9-35.9 months). In the PSM cohort, RT omission was not significantly associated with an increased risk of death (2-year OS 96.7% vs. 98.5%; HR = 1.12, 95% CI 0.23-5.41, P = 0.89).We did not find an association between RT omission and increased risk of death in patients with cT1-T2 (< 3 cm) N0 HER2+ BC treated with NST+BCS with pCR. Patients that did not receive RT tended to be older, of Black race, and to have received single-agent cytotoxic chemotherapy. These data support prospective evaluation of RT omission in this patient population in a randomized controlled trial.