Led by Professor Jens Rittscher
AI-assisted colon imaging
The accuracy of NHS colonoscopies – tests to check inside the bowels using a flexible tube with a small camera – is variable, which can lead to differences in outcomes for patients, including missed cancers.
In the previous BRC, we used AI-assisted 3D models of Barrett’s oesophagus, developed from endoscopy videos. We are now extending this technique to the colon, which is more complicated to access, to get real-time feedback when assessing ulcerative colitis, familial adenomatous polyposis (a hereditary condition characterised by multiple polyps lining the colon) and cancer of the colon. This work is led by Professor James East and Professor Rittscher
Improving diagnosis of HFpEF
Heart failure with preserved ejection fraction (HFpEF) is a condition that results in hospitalisation and mortality, but accurate diagnosis remains difficult.
We are applying artificial intelligence (AI) to a large-scale NHS dataset of echocardiograms – ultrasound scans that look at the heart and nearby blood vessels – taken from the EchoVision study. This has allowed Professor Paul Leeson’s team to develop automated methods to grade diastolic function (the ‘relaxation’ phase when the heart is pumping blood) and so simplify the diagnosis and management of HFpEF.
We are now testing and refining these methods across multiple NHS hospitals and in specific high-risk patient groups.
We are conducting studies to assess how cost-effective the imaging technologies we are developing are, so decisions can be made on whether they should be adopted in the NHS. We are also evaluating the ongoing implementation of imaging networks between NHS Trusts across England, which aim to maximise the use of existing capacity, improve access to specialist opinion and increase efficiencies and economies of scale. These studies will be led by Professor Apostolos Tsiachristas and Professor Philip Clarke, both of the Health Economics Research Centre.