Longitudinal DCE-MRI-Based Quantification of Tumor Morphological Complexity for Predicting Treatment Response in Patients with Breast Cancer

Motivation: Currently, there is a lack of standardized morphological biomarkers for Predicting pathologic complete response (pCR) to neoadjuvant chemotherapy (NAC) in breast cancer patients in breast cancer patients. Goal(s): To evaluate whether fractal dimensions (FDs) derived from longitudinal DCE-MRI can improve pCR prediction in breast cancer patients undergoing NAC. Approach: A total of 232 patients received DCE-MRI before and after two NAC cycles. Variables were assessed using logistic regression and linear mixed-effects models. Results: A model combining clinicopathologic variables and FDs showed good performance for predicting pCR to NAC with an AUC of 0.832. Impact: A model combining clinicopathologic variables and longitudinal fractal dimensions from DCE-MRI effectively predicts pathologic complete response to NAC in breast cancer, offering a valuable tool to enhance treatment decision-making and personalized care.