The role of MRI radiomic features in prediction of tumor response to NAC in breast cancer.

Pre-surgery NeoAdjuvant Chemotherapy (NAC) plays a key role in patients with Locally Advanced Breast Cancer (LABC). NAC can reduce tumor size, downstage tumors, increase the patient's surgical options, reduce metastasis, and improve the efficacy of follow-up treatment options after surgery. However, the pathological Complete Response (pCR) of NAC for Breast Cancer (BC) ranges between 10 and $50%$ due to the heterogeneity of tumors. Radiomics, $i.e.$, the extraction of quantitative features from the image and their correlation with the outcome, can be employed to optimize BC treatments. Specifically, MRI-based radiomics may help in the identification of pCR-related biomarkers that can be used in the prediction of pCR to NAC. This would result in a reduction of the delay in effective treatment of patients with non-responding or progressive tumors, and in avoiding side effects of unnecessary treatment. Several radiomics features have already been employed for this scope, leading to the construction of predictive models. However, the scenario is still vast, because there are lots of variables concerning radiomics features, statistical analysis, and underlying biology to be considered.