English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Recent breakthroughs in self-supervised learning have enabled the use oflarge unlabeled datasets to train visual foundation models that can generalizeto a variety of downstream tasks. While this training paradigm is well suitedfor the medical domain where annotations are scarce, large-scale pre-trainingin the medical domain, and in particular pathology, has not been extensivelystudied. Previous work in self-supervised learning in pathology has leveragedsmaller datasets for both pre-training and evaluating downstream performance.The aim of this project is to train the largest academic foundation model andbenchmark the most prominent self-supervised learning algorithms bypre-training and evaluating downstream performance on large clinical pathologydatasets. We collected the largest pathology dataset to date, consisting ofover 3 billion images from over 423 thousand microscopy slides. We comparedpre-training of visual transformer models using the masked autoencoder (MAE)and DINO algorithms. We evaluated performance on six clinically relevant tasksfrom three anatomic sites and two institutions: breast cancer detection,inflammatory bowel disease detection, breast cancer estrogen receptorprediction, lung adenocarcinoma EGFR mutation prediction, and lung cancerimmunotherapy response prediction. Our results demonstrate that pre-training onpathology data is beneficial for downstream performance compared topre-training on natural images. Additionally, the DINO algorithm achievedbetter generalization performance across all tasks tested. The presentedresults signify a phase change in computational pathology research, paving theway into a new era of more performant models based on large-scale, parallelpre-training at the billion-image scale.

Computational Pathology at Health System Scale -- Self-Supervised Foundation Models from Three Billion Images