AI for the early detection of upper GI cancer from routine blood tests
Research type
Research Study
Full title
Use of artificial intelligence for the early detection of upper GI cancers from routine blood tests
IRAS ID
325429
Contact name
Lesley Smith
Contact email
Sponsor organisation
University of Leeds
Duration of Study in the UK
3 years, 6 months, 1 days
Research summary
Upper GI cancer refers to a group of cancers affecting the oesophagus, stomach, liver, pancreas, and bile ducts. These cancers are often diagnosed at a late stage. This results in poor prognoses for patients. Earlier detection of these cancers may result in more effective treatment and improved outcomes. Late diagnosis occurs due to numerous factors; for some upper GI cancers, such as in the bile ducts, there are rarely symptoms in the early stages of the disease, so the cancer goes undetected. Where upper GI cancers are symptomatic, many of the symptoms (e.g., heartburn, indigestion, weight loss) are not specific to cancer and patients may dismiss or self-treat them with over-the-counter medications.
Patients often receive numerous blood tests during routine care in the years preceding diagnosis. Changes in these blood tests may indicate a patient is at higher risk of upper GI cancer. This study aims to develop an algorithm capable of detecting these changes and identifying patients who are likely to be diagnosed with upper GI cancer in the future. Identifying high-risk patients would increase the chance of diagnosing cancer at an early stage.
The data will be deidentified patient records obtained from Leeds Teaching Hospitals Trust (LTHT). The research participants will be patients who have been referred to the urgent suspected cancer pathway at this trust. These data will include laboratory tests, symptom data, comorbidities, demographics, and cancer diagnoses. We are looking for differences between patients who have upper GI cancer, and those who were not diagnosed with cancer following referral. We expect there will be approximately 260,000 patients (10,000 with upper GI cancer and 250,000 with no cancer) who meet these criteria and will be included in the analysis.REC name
London - Brighton & Sussex Research Ethics Committee
REC reference
23/PR/0619
Date of REC Opinion
21 Jul 2023
REC opinion
Further Information Favourable Opinion