Project aims to diagnose bowel cancer with AI

Bowel cancer is currently detected by inserting an endoscope into a patient’s colon which then travels around the large bowel allowing doctors to check for cancer.

Clinical Capsule Endoscopy (CCE) utilises a ‘smart pill’ containing cameras which, once swallowed by a patient, records images of the intestines as it passes through. The diagnostic procedure is currently being rolled out across the NHS in Scotland through the ScotCap programme.

At present, images captured by the capsules are reviewed by doctors, but AI could safely and ethically speed up the process, make it more cost-effective and increase its use.

To this end, a consortium of 12 European partners has received a £6m Horizon Europe grant to work towards eliminating the current medical, technical, and economic barriers to the adoption of AI-supported Image Analysis in Large Bowel Camera Capsule Endoscopy (AICE).

This includes demonstrating that AI algorithms are at least as effective as humans at analysing endoscopy images and identifying where further investigation is required.

The four-year AICE project involves Strathclyde University, NHS Highland and Islands, The Data Lab, and the Strathclyde-hosted Digital Health and Care Innovation Centre.

In a statement Professor Roma Maguire, Strathclyde University, said: “Capsule endoscopy, particularly AI-assisted, has huge potential to improve the early diagnosis of bowel cancer, but for such an approach to be adopted it has to be acceptable to patients.

“We are developing digital tools that will help us to understand patient outcomes of capsule endoscopy and crucially patients’ experience of, and feelings about, this procedure.”

The international collaboration is being led by the Centre for Clinical Implementation of Capsule Endoscopy at Odense University Hospital, Denmark.

The multidisciplinary team will cover aspects such as validation and development of algorithms, the creation of a clinical support platform, clinical indications and guidelines, patient engagement, cost-efficiency, ethical considerations and future implementation strategies.

The expected benefits include earlier initiation of treatment, less advanced staged cancers, fewer complications related to the diagnostic procedure, better patient acceptability and compliance and a significant reduction in costs from diagnostics and treatment. It also has potential to reduce the capacity pressures NHS health boards across the UK are experiencing.

Strathclyde is leading on work to develop and validate Patient Reported Experience Measures for capsule endoscopy, as well as the design and prototyping of a digital patient engagement tool to support participation in the procedure, adherence, and report access.  

Researchers will also develop nine AI algorithms which will be tested and validated in clinical settings in Denmark and by NHS Highland.

Professor George Crooks OBE, CEO, Digital Health & Care Innovation Centre, said: “Harnessing innovative technology to support the day-to-day management of patients is becoming increasingly recognised as one way that the NHS can deal with the significant challenges that it faces today. To secure high quality, safe and effective service delivery, there is a vital need to engage all the relevant expertise from across the clinical, research and technology communities to secure not only the latest technical innovations but also the clinical service models that serve the needs of patients and clinicians in equal measure.”

The other partners involved in the project include The Maersk Mc-Kinney Moller Institute, Lund University, Umeå University, UiT- The Arctic University of Norway, National & Kapodistrian University of Athens, and Universitat Autònoma de Barcelona.