Abstract
BACKGROUND: Accurate assessment of pathological complete response (pCR) to neoadjuvant chemotherapy (NAC) in breast cancer (BC) is crucial for mitigating chemotherapy-related toxicity in patients who do not respond to the treatment. Conventional ultrasound (US) has become a pivotal method for evaluating treatment response due to its cost-effectiveness, convenience, and absence of ionizing radiation. The objective of this study was to develop a model combining US and clinicopathological characteristics at baseline, as well as US features after one cycle of NAC, to predict the pCR to NAC in BC. METHODS: This retrospective study included 74 patients with invasive BC who underwent NAC from January 2022 to December 2023. Data from US and clinicopathological characteristics before NAC (pre-NAC) and US features after one cycle of NAC were collected from all patients. Univariate and multivariate analyses were used to screen the factors independently associated with pCR and to develop the prediction model. Receiver operating characteristic (ROC) curve analysis was performed, and the area under the curve (AUC), sensitivity, and specificity were calculated to assess the predictive efficiency. RESULTS: Four characteristics, including human epidermal growth factor receptor 2 (HER2)-positive [odds ratio (OR) 9.265; 95% confidence interval (CI): 1.617-53.095, P=0.012] and absence of posterior feature or posterior acoustic enhancement of the breast mass on the US pre-NAC (OR 9.435; 95% CI: 1.585-56.180, P=0.014), the maximum diameter reduction measured with the US (OR 1.081; 95% CI: 1.009-1.157, P=0.026), and the angular or spiculated margin of the breast lesion with the US after one cycle of NAC (OR 9.475; 95% CI: 1.247-71.969, P=0.030), were screened as independent predictors. The AUC, sensitivity, and specificity of the prediction model were 0.912, 90.0%, and 79.6%, respectively. CONCLUSIONS: US and clinicopathological characteristics at baseline and the US features after one cycle of NAC helped predict pCR for BC. The prediction model may enable early evaluation of the efficacy of treatment strategies and guide less invasive surgical options or personalized post-treatment plans.