Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Video

Updated: Jun 30, 2026

Enhancing Tumor Content through Tumor Macrodissection

Enhancing Tumor Content through Tumor Macrodissection

Published on: February 12, 2022

Dissecting and directing pathology foundation models.

Chanwoo Kim, Jakub Kaczmarzyk, Deepika Savant

Biorxiv : the Preprint Server for Biology

|June 29, 2026

Summary

This summary is machine-generated.

Related Concept Videos

Fixation and Sectioning

Fixation and Sectioning

Two basic types of preparation are used to visualize specimens with a light microscope: wet mounts and fixed specimens.
The simplest type of preparation is the wet mount, in which the specimen is placed in a drop of liquid on the slide. A liquid specimen can be directly deposited on the slide using a dropper. Solid specimens, such as skin scraping, can be placed on the slide before adding a drop of liquid to prepare the wet mount. Sometimes the liquid is simply water, but stains are often added...

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

Designing DNA With Tunable Regulatory Activity Using Discrete Diffusion.

bioRxiv : the preprint server for biology·2026

Same author

Genetic background shapes AI-predicted variant effects.

bioRxiv : the preprint server for biology·2026

Same author

Predicting non-coding variant effects with AlphaGenome.

Cell research·2026

Same author

Classifying Liver Cytology Using the WHO Reporting System: Institutional Experience and Malignancy Risk Assessment.

Diagnostic cytopathology·2026

Same author

Myofibroblasts Induce Neuroplasticity to Promote Pancreatic Inflammation and Cancer Progression.

Cancer discovery·2026

Same author

Small Bowel Neuroendocrine Tumors: An Update.

Archives of pathology & laboratory medicine·2025

Same journal

Layered social competition coordinates reproductive hierarchy formation in ants.

bioRxiv : the preprint server for biology·2026

Same journal

Combination epigenetic-targeted therapy increases the immunogenicity of poorly immunogenic sarcomas.

bioRxiv : the preprint server for biology·2026

Same journal

Loss of LanC-like proteins delays post-injury regeneration of aging skeletal muscles.

bioRxiv : the preprint server for biology·2026

Same journal

Integrative Transfer Network: Deep Transfer Learning Across Populations and Prediction Targets.

bioRxiv : the preprint server for biology·2026

Same journal

Confidence-supported label-free metabolic imaging with FPhaS phase autofluorescence microscopy.

bioRxiv : the preprint server for biology·2026

Same journal

Sequence-encoded autoinhibition couples mRNA decapping activity to phase separation.

bioRxiv : the preprint server for biology·2026

See all related articles

PICASSO makes foundation models (FMs) in digital pathology interpretable by decomposing image embeddings into visual concepts. This framework enhances trustworthiness, enables discovery of new biomarkers, and facilitates controlled manipulation of models for research.

Area of Science:

Computational Pathology
Artificial Intelligence in Medicine
Digital Pathology

Background:

Foundation models (FMs) are crucial in digital pathology, encoding histology images into embeddings for diagnostics and outcome prediction.
The 'black box' nature of these embeddings limits clinical trust and scientific discovery potential.
Existing FM embeddings lack transparency, hindering their full clinical translation and utility.

Purpose of the Study:

To introduce PICASSO (Pathology Image Concept Atlas built via SparSe dictiOnary learning), a framework for interpretable and controllable pathology FMs.
To decompose FM embeddings into human-interpretable visual concepts.
To enable auditing, discovery, and controlled manipulation of pathology FM representations.

Main Methods:

More Related Videos

Automated Dissection Protocol for Tumor Enrichment in Low Tumor Content Tissues

Automated Dissection Protocol for Tumor Enrichment in Low Tumor Content Tissues

Published on: March 29, 2021

Related Experiment Videos

Last Updated: Jun 30, 2026

Enhancing Tumor Content through Tumor Macrodissection

Enhancing Tumor Content through Tumor Macrodissection

Published on: February 12, 2022

Automated Dissection Protocol for Tumor Enrichment in Low Tumor Content Tissues

Automated Dissection Protocol for Tumor Enrichment in Low Tumor Content Tissues

Published on: March 29, 2021

Developed PICASSO, a framework utilizing sparse autoencoder for decomposing FM embeddings into visual concepts.

Trained PICASSO on over 120 million tissue patches across 32 cancer types to create a pan-cancer atlas of histomorphological concepts.

Integrated PICASSO-derived concepts with spatial transcriptomics and explored concept manipulation for counterfactual analysis.

Main Results:

PICASSO successfully decomposes FM embeddings into interpretable concepts, creating the first pan-cancer atlas of histomorphological concepts.
Demonstrated PICASSO's utility in auditing clinical models, discovering novel biomarkers (e.g., hobnailing morphology for EGFR mutations), and linking morphology to gene expression.
Showcased PICASSO's ability to mitigate technical artifacts and enable controlled manipulation of embeddings for therapeutic analysis, such as survival outcome prediction.

Conclusions:

PICASSO provides a principled framework to transform opaque pathology FMs into interpretable platforms for mechanistic insight and discovery.
The framework enhances trustworthiness and utility of FMs in digital pathology for both clinical validation and scientific exploration.
PICASSO facilitates the discovery of new biological insights and enables precise control over FM representations for advanced research applications.