Axillary Pathologic Complete Response to Neoadjuvant Chemotherapy in Clinically Node-Positive Breast Cancer Patients: A Predictive Model Integrating the Imaging Characteristics of Ultrasound Restaging with Known Clinicopathologic Characteristics

Abstract

The goal of this study was to evaluate various clinicopathologic and imaging characteristics as independent predictors of axillary pathologic complete response (pCR) to neoadjuvant chemotherapy (NAC) and to determine the added value of a model that integrates imaging characteristics of ultrasound (US) restaging with known clinicopathologic characteristics. A total of 227 clinically node-positive breast cancer patients underwent axillary US after NAC (termed US restaging) before surgery. We constructed a clinicopathologic model with independent predictors of clinicopathologic characteristics in multivariate analyses. A combined model was created by integrating imaging characteristics with clinicopathologic characteristics. The predictive values of the models were compared using the area under the receiver operating characteristic curve. Of the 227 patients, 106 (46.7%) achieved axillary pCR. Multivariate analysis revealed that higher histologic grades (odds ratio [OR] = 4.21 and 10.11 for moderate and high grade, respectively), negative hormonal receptor status (OR = 2.88), smaller (≤1.5 cm) residual tumor size (OR = 2.83), absence of fatty hilum loss (OR = 14.06) and absence of eccentric cortical thickening of the axillary lymph node (OR = 4.42) were independently associated with the axillary pCR (all p values < 0.05). Integrating the imaging characteristics of the US restaging significantly increased the predictive capability of the model that applied only the clinicopathologic characteristics (c-index, 0.783 vs. 0.657; p < 0.001). Imaging characteristics of the US restaging were independently associated with axillary pCR after NAC and they significantly improved the predictive capability of the model that used only the clinicopathologic characteristics.