Skip to content

Pathology AI

Self-learning pathology AI for toxicology and clinical trial teams.

TileBio Precedent is a preclinical toxicology suite. It detects abnormal tissue, searches by morphology and measures how dose changes morphology. TileBio Motif is for exploratory biomarker discovery in clinical trial slides. Both run on our patented tissue language model, which learns from unlabelled slides.

Setup
No labelsYour slides are indexed as they stand
Method
PatentedAnd published in Nature Communications
Whole-slide images
2M+Available for model training and validation
Deployment
On siteOn your own hardware in Docker, or hosted in the TileBio cloud
Haematoxylin and eosin stained tissue at high magnification

Technology

A model that learns tissue without labels

Our tissue language model learns from slides that nobody has annotated. It groups tissue into recurring patterns, then reads each slide as an arrangement of those patterns. A pathologist can open any pattern and see the tiles in it.

Partners

Data partnerships with the NHS

TileBio has a data partnership with NHS Greater Glasgow & Clyde and the NHS West of Scotland Innovation Hub. The method is patented and published in Nature Communications.

  • NHS Greater Glasgow & Clyde
  • NHS West of Scotland Innovation Hub

TileBio has agreed exploratory collaborative studies with Terrain Life Science, and is in discussions with pharmaceutical companies on toxicology and discovery pilots.

Run it on your own slides

Tell us about your study or cohort and the question you want answered.