A Dataset of Annotated Histopathological Images for Tumor Cellularity Assessment in Breast Cancer

The dataset consists of 2 H&E whole slide images, captured at a magnification of 40×, which overall contains tens of thousands of tumor nuclei from patients with breast cancer. The annotations, which consist of tumor and non-tumor nuclei, were provided by an expert pathologist, using the QuPath software. In detail, within the indicated ROIs, 1419 tumor-nuclei and 605 non-tumor-nuclei were labeled for a total of 2024 nuclei (respectively 762 and 381 from the first WSI, 657 and 224 from the second WSI). The institutional Ethic Committee of the IRCCS Istituto Tumori Giovanni Paolo II approved the study -- Prot n. 8/CE (February 2, 2021)