Predictive Model for the Pathological Complete Response of Axillary Lymph Nodes After Neoadjuvant Therapy for Breast Cancer

Abstract Background: Accurately assessing the efficacy of neoadjuvant therapy (NAT) for breast cancer may predict patient outcomes. Axillary pathological complete response (apCR) after NAT is considered to indicate survival benefit, allowing initially cN+ patients to avoid axillary lymph node dissection (ALND) by turning ALNs (axillary lymph node) negative. Therefore, there has been an increasing interest in accurately predicting apCR post-NAT. Here, we integrated the clinicopathological characteristics of patients with breast cancer and positive ipsilateral ALNs who underwent NAT and surgery. Statistical methods were used to screen for predictors of apCR and design a model to predict apCR rate post-NAT, thereby guiding surgeons’ decision-making and improving patients’ quality of life.Methods: Female patients (n=157) with locally advanced breast cancer and positive ALNs, diagnosed by a pathologist and treated via NAT and surgery between January 2016 and June 2020 at our hospital, were enrolled. Patients received a complete and standardized neoadjuvant regimen. Their clinicopathological characteristics were retrospectively analyzed, the factors of apCR post-NAT analyzed by logistic regression, and a nomogram used to construct a predictive model for apCR post-NAT.Results: Sixty-two patients (39.5%) achieved apCR. Univariate analysis demonstrated that apCR was significantly associated with pathological grading, molecular typing, stroma-tumor ratio (STR), clinical N staging, and response to chemotherapy (all P <0.05). A multivariate binary logistic regression equation was established, and the variables were entered into the model via forward stepwise selection; results indicated that the possibility of apCR post-NAT was higher in these cases. A predictive model for apCR post-NAT—consisting of four independent predictors: pathological grading of the aspiration specimen, STR, clinical N staging, and response to chemotherapy—was constructed using statistically significant variables in the logistic regression equation. The model had good accuracy and showed good clinical utility when both ROC curves and internal validation were performed.Conclusion: The predictive model combined clinical and pathological features; the pathological grading of the aspiration specimen, STR, clinical N staging, and response to chemotherapy were found to be independent predictors of apCR post-NAT. The model had good applicability regarding apCR prediction post-NAT, contributing to the option of individualized de-escalation of axillary surgery.