BMIRDS Datasets
Datasets
- Dartmouth Lung Cancer Histology Dataset
- MHIST: A Minimalist Histopathology Image Analysis Dataset
- Dartmouth Kidney Cancer Histology Dataset
- Dartmouth Breast Cancer Recurrence Risk Dataset
Dartmouth Lung Cancer Histology Dataset
Dartmouth Lung Cancer Histology Dataset comprises 143 hematoxylin and eosin (H&E)-stained formalin-fixed paraffin-embedded (FFPE) whole-slide images of lung adenocarcinoma from the Department of Pathology and Laboratory Medicine at Dartmouth-Hitchcock Medical Center (DHMC). This dataset contains whole-slide images in .tif image format, which were scanned by an Aperio AT2 whole-slide scanner at 20x or 40x magnification and converted to Generic tiled Pyramidal TIFF format.
MHIST: A Minimalist Histopathology Image Analysis Dataset
MHIST comprises 3,152 hematoxylin and eosin (H&E)-stained Formalin Fixed Paraffin-Embedded (FFPE) fixed-size images (224 by 224 pixels) of colorectal polyps from the Department of Pathology and Laboratory Medicine at Dartmouth-Hitchcock Medical Center (DHMC). Classes in the dataset indicate the predominant histological pattern of each image. The MHIST Classification task focuses on the clinically important binary distinction between Hyperplastic Polyps (HPs) and Sessile Serrated Adenomas (SSAs), a challenging problem with considerable inter-pathologist variability.
Dartmouth Kidney Cancer Histology Dataset
Dartmouth Kidney Cancer Histology Dataset comprises 563 hematoxylin and eosin (H&E)-stained formalin-fixed paraffin-embedded (FFPE) whole-slide images of renal cell carcinoma (RCC) from the Department of Pathology and Laboratory Medicine at Dartmouth-Hitchcock Medical Center (DHMC). In addition to surgical resection slides, this dataset includes biopsy slides that were used as an extended test set in our work.
Dartmouth Breast Cancer Recurrence Risk Dataset
The Dartmouth Breast Cancer Recurrence Risk Dataset comprises 990 hematoxylin and eosin (H&E)-stained, formalin-fixed paraffin-embedded (FFPE) whole-slide images (WSIs) along with corresponding recurrence risk and clinicopathologic information, including Oncotype DX Breast Recurrence Score®, patient age, tumor size, tumor grade, histologic type, ER status, PR status, and HER2 status for breast cancer cases from Dartmouth Health. This dataset was used to develop and evaluate a multi-model approach for predicting low- and high-risk breast cancer recurrence based on histology and clinicopathologic information, as reported in “A multi-model approach integrating whole-slide imaging and clinicopathologic features to predict breast cancer recurrence risk”.
