Exemplar Project:

Genomic biomarkers are critical tools in personalised medicine. Biomarker status helps to inform clinical trials recruitment, predict prognosis, treatment response, and disease progression. Biomarker data exists in electronic patient records (EPR), but largely in unstructured format not suitable for large scale computational analysis.

In order to allow for cohort identification at scale for clinical and research purposes, biomarker status and additional genomic information should be extracted and stored in a structured manner. This would benefit both clinical delivery and research in the NHS through such approaches as: reducing barriers to clinical trials recruitment, optimising treatment plans, reducing adverse drug reactions, and stratifying patient by risk to improve population healthcare delivery.   

Project Overview  

This project will focus on developing an automated AI-driven method of extracting genomic biomarker data from EPRs, which would unleash its potential for clinical care and research. The extracted genomic biomarker annotations will be linked to existing clinical datasets in NHS Trusts where this tool is deployed. 

This project will be aligned with ongoing NHS initiatives such as the Secure Data Environment programme, and leverages the investment and existing data infrastructure in the Sub National Secure Data Environment for London.  

National benefit

Once the NLP model is developed, it will be open source which means other NHS Trusts using the appropriate open-source informatics platforms can implement the model and annotate genomic biomarkers. This amplifies the use of the models across the NHS to enhance patient outcomes. 

For more information, contact the team:

Alexander Deng, Clinical Lead – alexander.deng@nhs.net

Sophie Ratkai, Lead Clinical Scientist for Genomics AI – sophie.ratkai@gstt.nhs.uk