Fig. 1From: Quantitative nuclear histomorphometric features are predictive of Oncotype DX risk categories in ductal carcinoma in situ: preliminary findingsIllustration of the overall workflow: (1) Regions of DCIS were annotated on the whole slide image (WSI) by an experienced breast pathologist. (2) Nuclei were segmented from the annotated tumor region via a deep learning model [16]. (3) Nuclear histomorphometric features were extracted. (4) The features were then evaluated in their ability to distinguish different ODx risk categories via supervised classification and unsupervised clustering approachesBack to article page