To assess the cost-effectiveness of computer-aided detection (CADe) for screening colonoscopy and to quantify, through value-of-information (VOI) analysis, which uncertainties most critically affect decision-making and should be prioritized for further research.

We adapted a Markov model of colorectal cancer (CRC) natural history to compare three strategies for average-risk 45-year-olds: no screening, 10-yearly colonoscopy, and 10-yearly CADe-assisted colonoscopy up to age 75, from a U.S. payer perspective. The model incorporated screening, post-polypectomy surveillance, and CRC treatment, using published data on adenoma detection rate (ADR), CRC risks, costs, and utilities. Outcomes were CRC incidence, CRC mortality, quality-adjusted life-years (QALYs), costs, and net monetary benefit (NMB) at a willingness-to-pay (WTP) of $100,000/QALY. Probabilistic sensitivity analysis (10,000 simulations) informed the expected value of perfect information (EVPI) and expected value of partial perfect information (EVPPI) for CADe performance, CRC natural history, and key cost parameters. Scenario analyses explored different magnitudes of CADe-driven ADR improvement, alternative background CRC incidence, and a non-linear relationship between ADR and CRC risk (risk plateau above 40% ADR).

Both screening strategies were cost-effective vs no screening. Compared with standard colonoscopy, CADe-assisted colonoscopy further reduced lifetime CRC incidence and mortality, yielding 75 additional QALYs per 10,000 individuals at an incremental cost of $388 per person (ICER $44,074/QALY). CADe was optimal in 71% of simulations at the $100,000/QALY WTP. Per-patient EVPI was $100, corresponding to a population EVPI of $5.8 billion over a 10-year technology lifetime. The largest EVPPI contributions came from CADe-related ADR effects, CRC natural history, and polypectomy costs. In conservative (4% ADR gain) and ADR-threshold scenarios, CADe remained potentially cost-effective but EVPI exceeded $8 billion.

CADe-assisted colonoscopy is likely to be cost-effective for CRC screening, yet substantial decision uncertainty persists. VOI analysis shows that high-value future research should focus on real-world CADe effects on ADR, CRC natural history, and the shape of the ADR–CRC relationship, to better support guideline and reimbursement decisions on CADe implementation.