The accuracy of mammography, sonography and magnetic resonance imaging (MRI) in identifying residual disease after neoadjuvant chemotherapy is evaluated and imaging findings are correlated with pathologic findings. Fifteen patients enrolled in an experimental protocol of preoperative neoadjuvant chemotherapy underwent clinical examination, mammography, sonography and dynamic MRI, performed in this order, before and respectively after 2 and 4 cycles of neoadjuvant chemotherapy. Four radiologists, two for mammography, one for sonography and one for MR, examined the images, blinded to the results of the other examinations. All patients underwent radical or conservative surgery, and imaging findings were compared with pathologic findings. MRI identified 2/15 (13.3.%) clinically complete response (CR), 9/15 (60%) partial response (PR), 3/15 (20%) stable disease (SD) and 1/15 (6.7%) progressive disease. Mammography identified 1/15 (6.7%) clinically CR, 8/15 (53.3%) PR and 4/15 (27%) SD, and was not able to evaluate the disease in 2/15 (13%) cases. Sonography presented the same results as MRI. Therefore, MRI and sonography compared to mammography correctly identified residual disease in 100 vs. 86%. MRI resulted in two false-negative results because of the presence of microfoci of in situ ductal carcinoma (DCIS) and invasive lobular carcinoma (LCI). MRI was superior to mammography in cases of multifocal or multicentric disease (83 vs. 33%). Sonography performed after MRI improves the accuracy in evaluation of uncertain foci of multifocal disease seen on MR images with an increase of diagnostic accuracy from 73 to 84.5%. MRI assesses response to neoadjuvant chemotherapy better than traditional methods of physical examination and mammography.