Abstract PS1-29: Predicting axillary pathologic response to neoadjuvant chemotherapy for node-positive breast cancer: Clinical predictive model by using MRI and ultrasound

Abstract

Abstract Purpose: Neoadjuvant chemotherapy (NAC) has been shown to eradicate axillary lymph node metastasis in approximately 40% of patients. Sentinel lymph node biopsy (SLNB) could be an alternative surgical approach for these patients to avoid morbidity from axillary lymph node dissection (ALND). However, high false-negative rates of SLNB for initially node-positive patients were reported in previous trials. The aim of this study was to evaluate clinicopathological factors and imaging characteristics by MRI and ultrasound (US) as predictors of axillary pathologic complete response (ypN0) after NAC, which enable to identify candidates for SLNB in patients with clinically node-positive disease. Patients and methods: We identified 153 patients with clinically node-positive breast cancer who received NAC from May 2009 to December 2019. Clinicopathological data including age, clinical T/N status, nuclear grade, hormone receptor (HR) and HER2 status were collected. All patients underwent MRI and US before and after NAC. Patients were judged to be node-positive when they have cytologically-proven nodal disease by fine-needle aspiration (FNA) or suspicious lymph nodes by diagnostic imaging. Lymph nodes with cortical thickness (>3.5mm), loss of fatty hilum or round shape (short-axis/long-axis ratio > 0.5) were defined as suspicious lymph nodes. All imaging data were evaluated at baseline and after NAC. To develop a predictive model for ypN0, the association between ypN0 status and clinicopathological and imaging characteristics were assessed by multivariate logistic regression analysis. The area under the receiver operating characteristic (ROC) curve was used to evaluate discrimination by the model. Results: The median age was 55.0 (range: 22-79) years and the mean tumor size was 3.86 ± 2.04 cm. Pretherapeutic lymph node status was assessed by FNA in 88 (57.5%) patients. Of 153 patients, 80 (52.3%) patients had luminal, 39 (25.5%) had HER2-positive, and 34 (22.2%) had triple negative disease. Sequential anthracycline and taxane were administered for 138 (90.2%) patients, and 37 (94.9%) patients with HER2-positive-disease received concomitant anti-HER2 agents preoperatively. Overall, 62 (40.5%) patients achieved ypN0. Independent predictors of ypN0 status were breast complete response by MRI (odds ratio [OR]: 9.01, p<0.001), clinical stage N1 (OR: 6.64 vs. cN2-3, p=0.009), absence of lymphadenopathy after NAC (OR: 6.09, p<0.001), HR negativity (OR: 3.10, p=0.02) and HER2 positivity (OR: 2.87, p=0.04). In a model using these predictors, the area under the ROC curve was 0.870 (95% confidence interval: 0.814-0.925, p<0.001). Sensitivity, specificity, positive predictive value and negative predictive value of the model were 72.6%, 86.8%, 78.9% and 82.3%, respectively. After a median follow-up of 49.7 months, 5-year disease-free survival (DFS) was significantly higher in patients who achieved ypN0 than patients with residual axillary disease (88.7% vs. 76.9%, p=0.046). Among 62 patients who achieved ypN0 after NAC, ALND was omitted in 24 (38.7%) patients and irradiation to regional lymph nodes was performed in 8 (33.3%) out of 24 patients. Five-year DFS was comparable between patients with or without ALND (86.8% vs. 91.7%, p=0.815). Conclusions: Our predictive model based on imaging characteristics by MRI and US could help to identify good candidates for omission of ALND after NAC in patients with initially node-positive breast cancer. Citation Format: Akiko Matsumoto, Miki Yamada, Daisuke Tsukahara, Yasuko Umemoto, Hiromitsu Jinno. Predicting axillary pathologic response to neoadjuvant chemotherapy for node-positive breast cancer: Clinical predictive model by using MRI and ultrasound [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PS1-29.