AI System developed at the UB to be tested in a clinical trial in UK

Capsule endoscopy is already being used in many countries for lower gastrointestinal disease investigations to deliver a range of benefits to patients and clinicians.
Capsule endoscopy is already being used in many countries for lower gastrointestinal disease investigations to deliver a range of benefits to patients and clinicians.
Research
(09/03/2022)

The National Institute of Health Research (NHIR) has funded a £1.3bn (1,54M€) clinical trial to validate an artificial intelligence system for detecting bowel polyps using an endoscopic capsule. The system, based on a deep learning model developed at the UB, is able to reduce the time needed to review images by up to 80% and it improves the accuracy of experts.

The project titled Capsule  Endoscopy delivery at SCalleges through enhanced AI analysis (CESCAIL) has begun to recruit patients this November. The project is led by the University Hospitals Coventry & Warwickshire and the company CorporateHealth International. It has a length of one and a half years.


 

Capsule endoscopy is already being used in many countries for lower gastrointestinal disease investigations to deliver a range of benefits to patients and clinicians.
Capsule endoscopy is already being used in many countries for lower gastrointestinal disease investigations to deliver a range of benefits to patients and clinicians.
Research
09/03/2022

The National Institute of Health Research (NHIR) has funded a £1.3bn (1,54M€) clinical trial to validate an artificial intelligence system for detecting bowel polyps using an endoscopic capsule. The system, based on a deep learning model developed at the UB, is able to reduce the time needed to review images by up to 80% and it improves the accuracy of experts.

The project titled Capsule  Endoscopy delivery at SCalleges through enhanced AI analysis (CESCAIL) has begun to recruit patients this November. The project is led by the University Hospitals Coventry & Warwickshire and the company CorporateHealth International. It has a length of one and a half years.


 

CESCAIL, one of the winners of the NIHR Artificial Intelligence in Health and Care Award will provide vital steps on the journey towards eliminating bowel cancer, the third most common cancer in the world1, which is potentially curable and even preventable if detected early enough.

Capsule endoscopy is already being used in many countries for lower gastrointestinal disease investigations to deliver a range of benefits to patients and clinicians, especially by adding urgently needed capacity to systems which were already strained before the pandemic intensified the problem. 

However, accurately analysing colon capsule endoscopy videos is a lengthy task for clinicians especially on complicated cases; time which can be better spent with patients. Artificial intelligence (AI) using Machine Learning to train Neural Networks can reduce that time by helping clinicians during analysis.

Machine learning algorithms, built through collaboration between CHI and the researchers from the Faculty of Mathematics and Computer Science from the University of Barcelona, accurately identify video images with potential signs of cancer or other abnormalities and highlights these, making video analysis faster and more efficient. 

ʻDeep learning, a machine learning model inspired by the human brain, has the ability to understand and find complex image patterns. Weʼve built an accurate model able to analyse a complete capsule video in few minutes. Thanks to AI, we improve the expertsʼ accuracy and reduce the required time for video screeningʼ Professor Santi Segui, University of Barcelona.

Further information

 

The video shows how the AI system works: