Due to copyright issues, an electronic copy of the thesis must be ordered from the faculty. For the faculty to have time to process the order, the order must be received by the faculty at the latest 2 days before the public defence. Orders received later than 2 days before the defence will not be processed. After the public defence, please address any inquiries regarding the thesis to the candidate.
Trial Lecture – time and place
See Trial Lecture.
Adjudication committee
- First opponent: Professor Alexander Meining, University Hospital Würzburg,
- Second opponent: Consultant specialist Noriko Suzuki, St Mark’s National Bowel Hospital,
- Third member and chair of the evaluation committee: Professor Marius Rehn, University of Oslo
Chair of the Defence
Associate professor Vendel Kristensen, University of Oslo
Principal Supervisor
Professor Yuichi Mori, University of Oslo
Summary
Artificial intelligence in colonoscopy – balancing benefits and harms
The aim of this thesis was to evaluate the clinical and economic consequences of artificial intelligence in colonoscopy to inform the scientific fields of health economics and gastroenterology.
Specifically, the thesis examined whether computer-aided detection (CADe) of polyps such as adenomas in the colorectum can reduce colorectal cancer incidence and mortality through improved adenoma detection rate (ADR), and whether computer-assisted diagnosis (CADx) can support optical diagnosis of diminutive rectosigmoid polyps to reduce unnecessary polypectomy and histopathology without compromising follow-up recommendations.
The thesis comprises of three studies: 1) a 10-year microsimulation modelling study of CADe in primary screening colonoscopy and FIT-positive colonoscopy in simulated European cohorts aged 60 to 69 years; 2) a microsimulation-based cost-effectiveness analysis estimating the cost per colorectal cancer prevented with CADe; and 3) a comparative cost analysis of CADx using EndoBRAIN International cohort data from 1,134 patients with 1,716 polyps from centres in Japan, the United Kingdom, and Norway.
CADe was associated with small absolute reductions in estimated colorectal cancer incidence and mortality, no measurable change in adverse events, and an increased probability of follow-up colonoscopy. The additional cost per colorectal cancer prevented was substantial and was driven largely by follow-up procedures. Optical diagnosis reduced colonoscopy-related costs, whereas the incremental benefit of CADx over endoscopist assessment alone was limited, although follow-up recommendations were not compromised. Overall, the thesis shows that implementation of AI in colonoscopy should be guided by careful assessment of clinical benefit, uncertainty, local capacity, and cost.
Additional information
Contact the research support staff.