Anonymised Dataset for Predicting DCIS Grade Using Mammographic and Sonographic Features

This dataset contains anonymised raw data for 130 patients diagnosed with pure ductal carcinoma in situ (DCIS) at XXX between 2010 and 2020. The dataset includes 79 imaging features extracted from digital mammography and breast ultrasound, along with demographic information, breast density, lesion detectability, and DCIS histological grade.These data were used to develop and validate a multinomial logistic regression–based nomogram to predict DCIS grade (low, intermediate, or high) using a dual-modality feature selection pipeline (Boruta and Recursive Feature Elimination).All personal identifiers have been removed in accordance with ethics committee approval (MREC ID: 202046-8456).<br>The dataset supports the findings reported in the manuscript entitled <b>“Nomogram-Based Method to Predict DCIS Grade from Mammography and Sonography”</b>, submitted to PLOS ONE.<br>