Mahmood Lab
Part of Harvard Medical School
A computational pathology lab at Harvard Medical School and Brigham and Women's Hospital, building open models for cancer diagnosis and prognosis.
Models (9)
Vision-omics finetuning that aligns pathology foundation models with spatial transcriptomics so morphology features predict local gene expression.
Attention-based multiple instance learning heads for whole-slide pathology, pretrained on a 108-way pan-cancer slide classification task.
KRONOS
Mahmood Lab / Brigham and Women's Hospital / Harvard Medical School / Broad Institute / Dana-Farber Cancer Institute / Beth Israel Deaconess Medical Center / Stanford University / The Ohio State University / University of Tübingen
Released June 3, 2025
Spatial proteomics foundation model, marker-aware and panel-agnostic, pretrained on 47 million multiplexed tissue-imaging patches from 175 markers.
TITAN
Mahmood Lab / Brigham and Women's Hospital / Helmholtz Munich / University of Tokyo
Released November 29, 2024
Slide-level pathology foundation model turning whole-slide images into reusable embeddings for classification, retrieval, and report generation.
Slide-level pathology foundation model that learns whole-slide embeddings by aligning multiple stains of the same tissue during pretraining.
PathChat
Mahmood Lab / Brigham and Women's Hospital / Harvard Medical School / Massachusetts General Hospital / The Ohio State University
Released July 10, 2024
Multimodal vision-language copilot for pathology that answers open-ended questions about histology images and reasons about differential diagnoses.
DeepLabV3 segmentation model that separates tissue from glass background in H&E and IHC whole-slide images, as used by the HEST-Library.
Computational pathology foundation model (ViT-L/16, DINOv2) pretrained on over 100 million H&E tiles from more than 100,000 whole-slide images.
Histopathology vision-language foundation model pretrained on 1.17 million image-caption pairs with contrastive and captioning objectives.