Background Accurate evaluation of response following chemotherapy treatment is essential for surgical decision making in patients with breast cancer.Modalities that have been used to monitor response to neo-adjuvant chemotherapy (NAC) include physical examination (PE),ultrasound (US),and magnetic resonance imaging (MRI).The purpose of this study was to evaluate the accuracy of PE,US,and MRI in predicting the response to NAC in patients with breast cancer.Methods According to the response evaluation criteria in solid tumors guidelines,the largest unidimensional measurement of the tumor diameter evaluated by PE,US,and MRI before and after NAC was classified into four grades,including clinical complete response,clinical partial response,clinical progressive disease,clinical stable disease,and compared with the final histopathological examination.Results Of the 64 patients who received NAC,the pathologic complete response (pCR) was shown in 13 of 64 patients (20％).The sensitivity of PE,US,and MRI in predicting the major pathologic response was 73％,75％,and 80％,respectively,and the specificity was 45％,50％,and 50％ respectively.For predicting a pCR,the sensitivity of PE,US,and MRI was 46％,46％,and 39％,respectively,and the specificity was 65％,98％,and 92％ respectively.Conclusions Compared with final pathologic findings,all these three clinical and imaging modalities tended to obviously underestimate the pCR rate.A more appropriate,universal,and practical standard by clinical and imaging modalities in predicting the response to neo-adjuvant chemotherapy in vivo is essential.