Background: This study was designed to evaluate the accuracy of breast magnetic resonance imaging (MRI) and ultrasonography (US) in predicting the extent of breast residual disease after preoperative chemotherapy.
Methods: Patients with stage II-III invasive breast tumors who received preoperative chemotherapy and were imaged with post-treatment MRI were included. Histopathological verification was available for all patients. The longest diameter of residual tumor measured with MRI and US has been compared with the infiltrating residual tumor size at pathologic evaluation.
Results: A total of 108 patients were enrolled: 59 were imaged with both MRI and US (MRI group), and 49 were imaged with US only (non-MRI group). The non-MRI group was enrolled as an external control to avoid possible bias in the selection of patients. In the MRI group, the means of the deltas between MRI residual tumor size and pathologic size and between US and pathologic size were 0.16 cm and -0.06 cm respectively (P = not significant). Overall, a discrepancy limited in the interval from -0.5 cm to +0.5 cm compared with the pathologic size was observed in 54% and 51% of the patients with MRI and US, respectively (P = not significant). The linear correlation between the radiological measurement and pathologic tumor size was r = 0.53 for MRI and r = 0.66 for breast US. In the non-MRI group, the mean of the deltas between US residual tumor size and pathologic size was 0.06 cm, and the linear correlation was r = 0.79.
Conclusions: In this series of patients, MRI and US do not show significant differences in predicting the breast residual infiltrating tumor after preoperative chemotherapy.