Abstract

Objective: The diagnosis of breast cancer (BC) lesions is a major factor influencing the treatment of neoadjuvant chemotherapy (NAC). This study evaluated the accuracy of Mammography (MG), ultrasound (US), and magnetic resonance imaging (MRI) in detecting BC lesions during NAC treatment for BC. Methods: Our study selected 226 cases of BC patients who were treated and discharged from our hospital between 2021 and 2023 as subjects. The primary diagnosis of these subjects was invasive ductal carcinoma of the breast. They all received NAC and surgical treatment during their hospitalization. Immunohistochemistry was used to determine the molecular type of BC in the subjects. MG, US, and MRI were employed to assess tumor size preoperatively. ROC curves and Pearson correlation analysis were used to evaluate the clinical diagnostic accuracy and significance of the differences. Results: In comparison with pathology, the three imaging modalities showed significant differences (p<0.001) in detecting tumor size in all tumor groups, HR-/HER2+ groups, and HR+/HER2+ groups. Regarding accu-racy, the highest accuracy rates were observed in the HR+/HER2- group, with respective percentages of 86% (nMG=86/108), 85% (nUS=92/108), and 83% (nMRI=89/108). In the HR+/HER2+ group, overestimation of tumor size was more common with MRI and US (48%, 27%). In the TNBC group, underestimation of tumor size was more common with MG (19%). For clinical diagnostic efficacy, ROC curves confirmed that the AUC values of all curves were greater than 0.5. Moreover, the sensitivity of MRI and US was similar to and higher than that of MG detection. The correlation between pathological tumor size and MRI or US was highest in the HR+/HER2+ group (r=0.767, r=0.792). Conclusion: The accuracy and relevance of MG, US, and MRI vary for patients with different molecular types. Compared to MG and US, MRI has a higher correlation with pathology and greater accuracy for various types of patients. Therefore, preoperative MRI examination is more important.