We applied Topological Data Analysis (TDA) to pre-neoadjuvant chemotherapy (pre-NAC) DCE-MRI datasets from the ISPY-1 trial for NAC treatment of breast cancer (BC). The pairwise topological distance between tumors’ signal enhancement ratio (SER) maps was computed. Hierarchical Agglomerative Clustering (HAC) was applied to cluster topologically similar patients. In combination with clinical and histopathological data using logistic regression models, the predictive performance of MRI topology for pathologic complete response (pCR) was compared to longest diameter (LD) and functional tumor volume (FTV). The preliminary results show that MRI topology may be a more accurate early predictor of BC response to NAC.
This abstract and the presentation materials are available to members only; a login is required.