Improving diagnostic yields of FIT using AI & ML (version 1.0)
Research type
Research Study
Full title
Improving diagnostic yields of the Faecal Immunochemical Test using Artificial Intelligence and Machine Learning
IRAS ID
292082
Contact name
Mitch Grigoriu
Contact email
Sponsor organisation
Advanced Expert Systems Ltd.
Clinicaltrials.gov Identifier
BCSPRAC_0256(ODR1819_258), BCSP Research Advisory Committee
Duration of Study in the UK
1 years, 0 months, 0 days
Research summary
Title: Improving diagnostic yields of the Faecal Immunochemical Test using Artificial Intelligence and Machine Learning (Version 1.0). \nCondition/area under study: Bowel Cancer.\nConsortium undertaking the Study: University Hospitals Coventry & Warwick (UHCW)and Advanced Expert Systems Ltd.(AES).\nSites where the study will be conducted: One NHS site at UHCW and one site at AES.\nStudy duration: 1 year\nThe Bowel Cancer Screening Programme (BCSP) invites people aged over 60 years to submit stool samples every two years for the screening test known as Faecal Immunochemical Test (FIT). This measures the amount of altered blood in the stool. FIT samples are analysed against a benchmark (a pre-set maximum limit of blood). Patients whose tests exceed this limit are invited to undergo a colonoscopy to examine the lower bowel for causes of bleeding such as cancer or precancerous growths (polyps). Many colonoscopies are performed each year, but as many as 45% show entirely normal results. Colonoscopy is an invasive procedure that can have serious side effects such as gastrointestinal bleeding, perforation, etc. and is costly to the NHS. \nIn order to reduce unnecessary colonoscopies, a consortium comprising of NHS clinicians and researchers from the University Hospital of Coventry and Warwick (UHCW) and Advanced Expert Systems Ltd (AES) propose the development of an AI system – “ColonSys” - to predict potential cases of cancer or polyps using FIT results and other patient-specific data to enhance the ability of NHS BCSP hubs to identify bowel cancer more effectively. AES have completed feasibility studies (funded by Innovate UK) that have determined which clinical and nonclinical variables will feed into the predictive model, with framework options for how the new system could be embedded within NHS BCSP hubs.\nThis project uses only de-identified data from FIT Pilot Study 2014. \n \nResearch findings would be shared with the National Screening Committee and representatives in England, Wales and Scotland.\n
REC name
North of Scotland Research Ethics Committee 1
REC reference
20/NS/0144
Date of REC Opinion
9 Dec 2020
REC opinion
Favourable Opinion