Neoadjuvant Chemotherapy Response Evaluation in Breast Cancer: Diagnostic Performance of MRI Combined with Ultrasound Across Molecular Subtypes

Background: This study aimed to evaluate the efficacy of magnetic resonance imaging (MRI) combined with ultrasound (US) in predicting pathological complete response (pCR) after neoadjuvant chemotherapy (NACT) in breast cancer and to assess clinicopathological and radiological factors influencing radiology–pathology concordance. Methods: A retrospective single-center study included 95 patients with locally advanced breast cancer who received NACT between 2018 and 2021. Baseline and post-NACT MRI and US were assessed according to RECIST 1.1, integrating both modalities. Surgical pathology served as the reference standard. Diagnostic performance was calculated overall and by molecular subtype, with group comparisons using chi-square or Fisher’s exact tests with Bonferroni correction. Results: pCR was achieved in 35 patients (36.8%), differing significantly among molecular subtypes (p < 0.001), with the highest rate in HER2-positive tumors (88.2%). MRI + US predicted pCR with 83% sensitivity, 83% specificity, 74% PPV, 89% NPV, and 83% accuracy. NPV varied significantly across subtypes (global p = 0.008), lowest in HER2-positive tumors (33.3%), significantly lower than Luminal A and TNBC (both 100%, p = 0.012). US alone showed lower performance (accuracy 72.6%). MRI-estimated tumor size correlated moderately with pathology (rho = 0.63, p < 0.001), versus weaker correlation for US (rho = 0.49, p < 0.001). MRI estimated residual size within 10 mm of pathology in 71.5% of cases. Conclusions: MRI combined with US is accurate for predicting pCR after NACT. However, its lower NPV in HER2-positive tumors limits reliability in excluding residual disease, underscoring the need for histopathological confirmation before modifying surgical planning.