Increased use of cancer screening, improved imaging, and diagnostic intervention techniques has led to the diagnosis of smaller cancers, including breast cancer. Most breast cancer patients receive systemic therapy, and some treatments are given before surgery, such as neoadjuvant therapy, even in an operable setting. Improved neoadjuvant chemotherapy has increased rates of pathological complete response; however, surgery is still required to determine complete tumor remission. Inadequate preoperative evaluations after neoadjuvant therapy can result in excessive surgical stress. Clinical imaging tests such as ultrasound and magnetic resonance imaging of the breast are often performed with neoadjuvant therapy. These clinical imaging techniques, in addition to measuring tumor size, have made it possible to evaluate certain functional aspects of the tumors. Herein, we review the current state of clinical imaging research focused on predicting neoadjuvant chemotherapy response in breast cancer. We also discuss the upfront prediction of treatment response before and during neoadjuvant therapy and the later prediction of pathological residual tumors, including pathological complete response, using ultrasound and magnetic resonance imaging. Upfront prediction can help decision-making and develop new treatment strategies. Predicting the localization of microscopic residual tumors may contribute to disease management without surgery, using radiation or other local treatments. Further larger studies on the prediction of neoadjuvant therapy response using clinical imaging could improve clinical practice and patient benefits.
Keywords: Breast cancer; artificial intelligence (AI); elastography; imaging biomarker; neoadjuvant chemotherapy; prediction.