Mahmood Lab aims to develop machine learning, data fusion, and medical image analysis methods for objective diagnosis, prognosis, and biomarker discovery. We are interested in developing automated and objective mechanisms for reducing interobserver and intraobserver variability in cancer diagnosis using artificial intelligence as an assistive tool for pathologists. The lab also focuses on the development of new algorithms and methods to identify clinically relevant morphologic phenotypes and biomarkers associated with response to specific therapeutic agents. We develop multimodal fusion algorithms for combining information from multiple imaging modalities, familial and patient histories and multi-omics data to make more precise diagnostic, prognostic and therapeutic determinations. We are affiliated with the Harvard Data Science Initiative; the Harvard Bioinformatics and Integrative Genomics (BIG) program; the Cancer Data Science Program at the Dana-Farber Cancer Institute and the Cancer Program at the Broad Institute of Harvard and MIT.
Congratulations to our superstar Anurag Vaidya on successfully defending his PhD @ Harvard-MIT HST! Anurag worked on some very exciting studies spanning multimodal outcome prediction (SurvPath, CVPR 2024), auditing demographic bias in computational pathology (Nature Medicine, 2024), and multiple foundation models including MADELEINE (ECCV 2025), THREADS (Nature Cancer, 2026, to appear), and KRONOS.
New paper to appear at ICLR 2026: Mixture of Mini Experts: Overcoming the Linear Layer Bottleneck in Multiple Instance Learning, see article here, and code here. Also, see new pre-print on enhancing Pathology foundation models using spatial transcriptomics using SEAL, see article here and code here.
Congratulations to our superstar postdocs Andrew Song and Guillaume Jaume for starting their faculty appointments and research labs at the MD Anderson Cancer Center and University of Lausanne. We are thrilled to celebrate this major milestone and look forward to all they will accomplish.
Modella AI the first start-up to spin off from our lab and led by our former PhD students Richard Chen (Harvard BIG) and Max Lu (MIT EECS) has been acquired by AstraZeneca. Read the press release here and here.
Honored to deliver three invited addresses in early 2026: Nathan Kauffman Lecture at USCAP 2026, a Plenary Talk at AACR 2026, and the Keynote at ISBI 2026, see more about our speaking schedule here.
Faisal has been appointed as the Director of the newly established MGB Center for AI Research. Further updates on this exciting endeavor to follow later in 2026.
TITAN, our multimodal whole slide foundation models is published in Nature Medicine, access the article here, and download the models here.
Excited to share the renewal of our R35 MIRA Outstanding Investigator Award for the next five years, Multimodal and Generative AI for Pathology.
Mahmood Lab together with Le Lab from MGH and other collaborators received an S10 award from the NIH to substantially expand our GPU compute cluster for ongoing computational pathology, multimodal data integration, and healthcare AI research.
Congratulations to our superstar PhD students, Cristina Almagro-Pérez for receiving the prestigious Rafael del Pino Foundation Fellowship, and Tong Ding for passing his qualifying exam! We’re also excited to welcome HMS AIM PhD student Caiwei Tian to the lab.
Introducing KRONOS - our new foundation model for spatial proteomics, read the pre-print, download the model, check out our code and tutorials, read our blog posts from Sizun and Shaban. More to follow soon!
Our lab has received an ARPA-H award from the ADAPT program to develop multimodal foundation models, and autonomous AI Agents to identify cancer resistance traits, predict treatment response, and discover new biomarkers. Read more about the awards here, and about the program in general here.
Comment on benchmarking in machine learning for biomedicine published in Nature Medicine.
We are excited to announce the release of two major internally developed open libraries for large scale batch processing of pathology WSIs and foundation model benchmarking: TRIDENT and Patho-Bench. Read our blog about these tools.
PathChat DX a clinical grade version of PathChat developed by the lab and published in Nature last June has received FDA Breakthrough Device Designation, becoming one of the first generative AI tools for pathology to receive the designation.
New pre-print, TITAN - A multimodal whole slide foundation model for computational pathology. See pre-print, and download model.
HEST libraryand dataset accepted for publication at NeurIPS 2024. See codehere and download data here.
