Awesome Digital and Computational Pathology

This curated list of useful resources is supported by:

Contents

• Software
  • Assistant
  • Image Analysis
  • Image IO
  • Machine Learning
  • Model
  • Foundation Model
  • Platform
  • Viewer
  • Viewer (Free)
• Data
  • Challenges
  • Datasets
  • References
• Publications
  • Papers

Software

Assistant

• Digital Pathology Assistant - Specify your requirements in plain english and I'll provide PathML and Python code for your use-case.

Image Analysis

• HistomicsTK - Toolkit for the analysis of digital pathology images.
• HistoQC - Quality control tools for digital pathology.
• PathProfiler - Quality assessment of histopathology WSI cohorts.
• PyHIST - Histological image segmentation tool.
• pyslide - Digital pathology WSI analysis toolbox.
• TIA Toolbox - Computational pathology toolbox that provides an end-to-end API for pathology image analysis.

Image IO

• Bio-Formats - Java software tool for reading and writing microscopy image using standardized, open formats.
• compay-syntax - Tissue mask and tiling pipeline.
• cuCIM - NVIDIA's accelerated computer vision and image processing software library for multidimensional images.
• libvips - A fast image processing library with low memory needs.
• OpenSlide - Provides a simple C interface with Python bindings to read WSIs in multiple formats.
• svg2svs - Generate checkerboard and build multi-layer pyramidal SVS files from SVG images.
• tifffile - Read and write TIFF-like files using in bioimaging.
• WholeSlideData - Batch iterator that enables fast, efficient and easy patch sampling.

Machine Learning

• DLUP - Deep learning utilities for pathology.
• eva - Evaluation framework for oncology foundation models.
• histocartography - Library designed to facilitate the development of graph-based computational pathology pipelines.
• nuclei.io - Human-in-the-loop active learning framework for pathology image analysis.
• PathML - Tools for computational pathology.
• Slideflow - Python package that provides a unified API for building and testing deep learning models for histopathology.

Model

• ACMIL - WSI classification.
• BEPH - BEiT-based model pre-training on WSIs.
• Cell-DETR - Attention-based transformers for instance segmentation of cells in microstructures.
• CellViT - Vision transformers for precise cell segmentation and classification.
• Cerberus - Multi-task learning enables simultaneous histology image segmentation and classification.
• CLAM - Data-efficient and weakly supervised computational pathology on WSI.
• DeepLIIF - Deep-learning inferred multiplex immunofluorescence for immunohistochemical image quantification.
• DiffInfinite - Large mask-image synthesis via parallel random patch diffusion in histopathology.
• DMMN - Deep Multi-Magnification Network for multi-class tissue segmentation of WSI.
• DT-MIL - Deformable transformer for multi-instance learning on histopathological image.
• FrOoDo - Framework for out of distribution detection.
• HistoGPT - Generating highly accurate histopathology reports from whole slide images.
• HistoSegNet - Semantic segmentation of histological tissue type in WSIs.
• HoVer-Net - Simultaneous segmentation and classification of nuclei in multi-tissue histology images.
• LongViT - Vision Transformer that can process gigapixel images in an end-to-end manner.
• MCAT - Multimodal co-attention transformer for survival prediction in gigapixel WSIs.
• MMP - Multimodal prototyping for cancer survival prediction.
• MSINet - Deep learning model for the prediction of microsatellite instability in colorectal cancer.
• PANTHER - Morphological prototyping for unsupervised slide representation learning in computational pathology.
• Patch-GCN - WSI are 2D point clouds: Context-aware survival prediction using patch-based graph convolutional networks.
• RSP - Self-supervised driven consistency training for annotation efficient histopathology image analysis.
• SparseConvMIL - Sparse convolutional context-aware multiple instance learning for WSI classification.
• StainGAN - Stain style transfer for digital histological images.
• stainlib - Augmentation & normalization of H&E images.
• StainTools - Tools for tissue image stain normalisation and augmentation.
• StarDist - Object detection with star-convex shapes.
• TANGLE - Transcriptomics-guided slide representation learning in computational pathology.
• TCGA segmentation - Weakly supervised multiple instance learning histopathological tumor segmentation.
• torchstain - Stain normalization transformations.
• TransMIL - Transformer based correlated multiple instance learning for WSI classification.

Foundation Model

• CONCH - Vision-language foundation model for computational pathology.
• Hibou - A family of foundational vision transformers for pathology.
• HIPT - Scaling vision transformers to gigapixel images via hierarchical self-supervised learning.
• H-optimus - Foundation model for histology.
• PathDino - Rotation-agnostic image representation learning for digital pathology.
• Phikon - Scaling self-supervised learning for histopathology with masked image modeling.
• Prov-GigaPath - A whole-slide foundation model for digital pathology from real-world data.
• ROAM - A transformer-based weakly supervised computational pathology method for clinical-grade diagnosis and molecular state revelation of gliomas.
• TransPath - Transformer-based unsupervised contrastive learning for histopathological image classification.
• UNI - General-purpose foundation model for computational pathology.
• VIM4Path - Self-supervised vision mamba for WSIs.
• Virchow - Self-supervised vision transformer pretrained using 1.5M WSIs.

Platform

• Digital Slide Archive - Provides the ability to store, manage, visualize and annotate large imaging datasets.

Viewer

• ASAP - Desktop application for visualizing, annotating and automatically analyzing WSIs.
• Cytomine - Collaborative analysis of WSIs.
• DigiPathAI - Tool to visualize gigantic pathology images and use AI to segment cancer cells and present as an overlay.
• HistomicsUI - Web interface to visualize WSI and manage annotations.
• slim - Interoperable web-based slide microscopy viewer and annotation tool.
• QuickAnnotator - Model assisted tool for rapid annotation of WSIs.
• QuPath - Java application that enables researchers and pathologists to visualize, analyze and annotate WSIs.

Viewer (Free)

• Aperio ImageScope - Freely downloadable software for viewing WSIs. Windows only.
• PathPresenter - A complete enterprise workflow platform built by pathologists.

Data

Challenges

• ACDC - Automatic Cancer Detection and Classification of lung histopathology.
• ACROBAT - AutomatiC Registration Of Breast cAncer Tissue.
• ANHIR - Automatic Non-rigid Histological Image Registration.
• BACH - BreAst Cancer Histology images.
• BCI - Breast Cancer Immunohistochemical image generation.
• BreastPathQ - Quantitative biomarkers for the determination of cancer cellularity.
• CAMELYON16 - Cancer metastasis detection in lymph node.
• CAMELYON17 - Building on CAMELYON16 by moving from slide level analysis to patient level analysis.
• CellSeg - Cell segmentation in multi-modality high-resolution microscopy images.
• CoNIC - Colon Nuclei Identification and Counting.
• DigestPath 2019 - Digestive-system pathological detection and segmentation.
• ENDO-AID - Endometrial carcinoma detection in pipelle biopsies.
• Gleason 2019 - Automatic Gleason grading of prostate cancer in digital histopathology.
• HER2 Scoring Contest - Automated HER2 scoring algorithms in WSI of breast cancer tissues.
• HEROHE - Predicting HER2 status in breast cancer from H&E.
• KPIs - Kidney Pathology Image segmentation.
• LEOPARD - LEarning biOchemical Prostate cAncer Recurrence from histopathology sliDes.
• LYSTO - LYmphocytes aSsessment hackathOn in immunohistochemically stained tissue sections.
• LYON19 - LYmphocyte detectiON in IHC stained specimens.
• MIDOG 2021 - MItosis DOmain Generalization on tissue preparation and image acquisition.
• MIDOG 2022 - MItosis DOmain Generalization on tissue types.
• MITOS-ATYPIA-14 - Detection of mitosis and evaluation of nuclear atypia score.
• MoNuSAC - Multi-Organ NUclei Segmentation And Classification.
• MoNuSeg - Multi-Organ NUclei Segmentation.
• PAIP2019 - Liver cancer segmentation.
• PAIP2020 - Classification and segmentation of microsatellite instability (MSI) in colorectal cancer.
• PAIP2021 - Perineural invasion in multiple organ cancer.
• PAIP2023 - Tumor cellularity prediction in pancreatic cancer and colon cancer.
• PANDA - Prostate cANcer graDe Assessment.
• SegPC - Segmentation of multiple myeloma in Plasma Cells.
• TIGER - Fully automated assessment of tumor-infiltrating lymphocytes (TILs) in H&E breast cancer slides.
• TUPAC16 - TUmor Proliferation Assessment.
• WSSS4LUAD - Weakly-supervised tissue semantic segmentation for lung adenocarcinoma.

Datasets

• ARCH - Multiple instance captioning.
• BCNB - Early Breast Cancer Core-Needle Biopsy WSI dataset.
• BCSS - Breast Cancer Semantic Segmentation.
• BRACS - BReAst Carcinoma Subtyping.
• CRC - 100,000 histological images of human colorectal cancer and healthy tissue.
• CryoNuSeg - Nuclei segmentation of cryosectioned H&E-stained histological images.
• H2T - Handcrafted Histological Transformer for unsupervised representation of WSIs.
• HEST - Bringing spatial transcriptomics and histopathology together.
• LC25000 - Lung and colon cancer histopathological image dataset.
• LyNSeC - Lymphoma nuclear segmentation and classification dataset.
• MHIST - Minimalist histopathology image analysis dataset.
• NuInsSeg - A fully annotated dataset for nuclei instance segmentation in H&E-stained histological images.
• NuCLS - A scalable crowdsourcing approach & dataset for nucleus classification, localization and segmentation in breast cancer.
• OCELOT - Overlapped cell on tissue dataset for histopathology.
• PanNuke - Dataset for nuclei instance segmentation and classification.
• PCAM - PatchCamelyon provides a new benchmark for machine learning models akin to CIFAR-10 and MNIST.
• TCGA - The Cancer Genome Atlas is a landmark cancer genomics program that molecularly characterized over 20,000 primary cancer and matched normal samples spanning 33 cancer types.
• UNITOPATHO - A labeled histopathological dataset for colorectal polyps classification and adenoma dysplasia grading.
• UNMaSk - Unmasking the immune microecology of ductal carcinoma in situ.

References

• ADP - Atlas of digital pathology for deep learning.
• Cytomine Collection - Open access to high quality WSI in several formats.
• DICOM WSI Standard - DICOM WSI document.
• Jerad Gardner, MD - Popular YouTube channel for educational videos by a pathologist.
• WebPathology - Visual survey of surgical pathology.

Publications

Papers

• chen2022self - Self-Supervised Vision Transformers Learn Visual Concepts in Histopathology.
• kang2022benchmarking - Benchmarking Self-Supervised Learning on Diverse Pathology Datasets.
• wolflein2023good - A Good Feature Extractor Is All You Need for Weakly Supervised Pathology Slide Classification.
• vaidya2024demographic - Demographic bias in misdiagnosis by computational pathology models.