Pathologic complete remission after neoadjuvant chemotherapy has a role in guiding the management of breast cancer. The present meta-analysis examined the accuracy of contrast-enhanced magnetic resonance imaging (CEMRI) and diffusion-weighted magnetic resonance imaging (DW-MRI) in detecting the response to neoadjuvant chemotherapy and compared CE-MRI with ultrasonography, mammography, and positron emission tomography/ computed tomography (PET/CT). Medical subject heading terms and related keywords were searched to generate a compilation of eligible studies. The pooled sensitivity, specificity, diagnostic odds ratio, area under summary receiver operating characteristic curve (AUC), and Youden index (Q* index) were used to estimate the diagnostic efficacy of CE-MRI, DW-MRI, ultrasonography, mammography, and PET/CT. A total of 54 studies of CE-MRI and 8 studies of DW-MRI were included. The overall AUC and the Q* index values for CE-MRI and DW-MRI were 0.88 and 0.94 and 0.80 and 0.85, respectively. According to the summary receiver operating characteristic curves, CE-MRI resulted in a higher AUC value and Q* index compared with ultrasonography and mammography but had values similar to those of DW-MRI and PET/CT. CE-MRI accurately assessed pathologic complete remission in specificity, and PET/CT and DW-MRI accurately assessed pathologic complete remission in sensitivity. The present meta-analysis indicates that CE-MRI has high specificity and DW-MRI has high sensitivity in predicting pathologic complete remission after neoadjuvant chemotherapy. CE-MRI is more accurate than ultrasonography or mammography. Additionally, PET/CT is valuable for predicting pathologic complete remission. CE-MRI, combined with PET/CT or DW-MRI, might allow for a more precise assessment of pathologic complete remission.

Introduction

Neoadjuvant chemotherapy (NAC), also called termed or primary chemotherapy, was first described in patients with locally advanced breast cancer in 1978.1 It plays a well-established role in the management of breast cancer.2 One of the advantages of NAC is to downstage the primary tumor and increase the likelihood of successful breast-conserving surgery, avoiding mastectomy.3 Published clinical trials have shown that breast cancer patients with pathologic complete remission (pCR) after NAC have a significantly better prognosis than those without pCR.4-9 The accurate assessment of the response to NAC before surgery is crucial in breast cancer management. If pCR cases could be distinguished perfectly from non-pCR cases, additional surgical management would be avoided. Ultrasonography, mammography, and contrast-enhanced magnetic resonance imaging (CE-MRI) have been widely applied to evaluate and predict the pathologic response in patients with breast cancer who received NAC. However, positron emission tomography/computed tomography (PET/CT) has not been widely used.10-12 CE-MRI has been proposed to be valuable in predicting the responsiveness of breast cancer after NAC.13 Also, diffusionweighted MRI (DW-MRI) has been considered a potential tool to predict the response to chemotherapy.14,15 Therefore, we focused on the performance of MRI (including CE-MRI and DW-MRI) in predicting for pCR after NAC. Despite the value of CE-MRI in detecting the response to NAC, the routine use of CE-MRI remains controversial because of the cost, lengthy imaging time, use of contrast injection, and relatively low sensitivity in detecting pCR.16 Skaane et al17 reported that ultrasonography and mammography were the primary tools for evaluating the response to NAC; however, the accuracy was limited. Most published studies have evaluated MRI and other imaging techniques in isolation, not in the same cohort.