TOMPEI-CMMD Dataset

The original CMMD was published in the Cancer Imaging Archive (TCIA) and consists of 3,728 mammograms from 1,775 patients examined between July 2012 and January 2016. Annotations in this dataset include biopsy-confirmed diagnoses that categorize tumors as benign or malignant; molecular subtypes of tumors are also included for certain patients. The absence of segmentation information in the original dataset limits detailed anatomical analysis because it does not provide data on tumor location, size, shape, and boundary characteristics. The TOMPEI-CMMD dataset, which extends the original CMMD, integrates additional detailed labels and segmentation masks for radiological findings provided by the Department of Diagnostic Imaging at Tohoku University Graduate School of Medicine. Sharing this enhanced dataset is expected to substantially benefit the development of breast cancer computer-aided detection (CAD) systems using deep learning.