Digital Pathology & Spatial AI
Deep learning on digitised pathology and spatial molecular data, to detect, grade and predict molecular features.
What this centre does
Pathology slides contain far more information than the human eye can extract. This centre learns to detect and grade disease, to predict molecular features directly from tissue images, and to map where different cell types sit in relation to each other within the tumour.
Methods
Whole-slide image analysis; weakly supervised and multiple-instance learning; morphology-to-molecular prediction; spatial transcriptomics analysis; nuclei and tissue segmentation.
Data inputs
Digitised slides from partner pathology laboratories; matched molecular and outcome data; public slide repositories.
Outputs
Research-use models for grading and molecular prediction; spatial atlases; publications.
Funding route
AI-in-health grants; digital pathology collaborations; industry partnerships.
What has to be true first
A slide-digitisation route through a partner laboratory, and a data-sharing agreement with a defined consent basis.
Themes this centre carries
Work with this centre
The research centres collaborate with funders, sponsors, hospitals and academic groups. Tell us what you are working on and the enquiry is routed to the centre lead.
Contact the research office