Artificial Intelligence–Assisted Colonoscopy for Detecting Polyps, Adenomas, Precancerous Lesions, and Colorectal Cancer
Bibliographic record
Abstract
What Is the Issue? Colorectal cancer is 1 of the most common cancers globally. In Canada, it is estimated that more than 25,000 people will be diagnosed with colorectal cancer in 2024 and that more than 9,000 people will die of it. To reduce the incidence and mortality of colorectal cancer, screening programs across various jurisdictions in Canada have implemented screening strategies involving routine fecal testing and colonoscopy. While colonoscopy is often considered the gold standard for colorectal cancer screening, missed polyps remain a challenge with this modality. In recent years, many artificial intelligence (AI)–enabled polyp detection systems have been developed for use during colonoscopy. A review of the clinical and cost-effectiveness of these systems could help clarify their potential role in clinical practice. What Did We Do? To inform decisions regarding the use of AI-assisted colonoscopy, we conducted a rapid review to identify and summarize evidence that compared the clinical and cost-effectiveness of AI-assisted colonoscopy to conventional colonoscopy and among different AI-assisted colonoscopy systems. We searched key resources, including journal citation databases, and conducted a focused internet search for relevant evidence published since 2019. One reviewer screened articles for inclusion based on predefined criteria, critically appraised the included studies, and narratively summarized the findings. What Did We Find? We found 1 health technology assessment (HTA), 3 systematic reviews (SRs), and 1 randomized controlled trial (RCT) that evaluated the clinical and cost-effectiveness of AI-assisted colonoscopy for detecting polyps, adenomas, precancerous lesions, and colorectal cancer. AI-assisted colonoscopy may improve clinical outcomes compared to conventional colonoscopy, including adenoma detection rates, the number of adenomas detected per procedure, and adenoma miss rates, although it may also lead to longer withdrawal times. However, not all studies included in this report demonstrated statistically significant differences between the groups for each of these outcomes. The included RCT allocated participants to receive AI-assisted colonoscopy with either Deep-GI or CAD EYE, but the authors did not perform statistical testing to compare the outcomes between these 2 treatment groups. Findings from 6 economic evaluations summarized in the HTA suggest that AI-assisted colonoscopy is likely to be cost-effective or dominant — meaning it is less costly and more effective — compared to conventional colonoscopy. We did not find any studies on the relative cost-effectiveness of different AI-assisted colonoscopy systems that met our selection criteria for this review. None of the included studies reported long-term outcomes, such as colorectal cancer incidence and mortality; therefore, the impact of AI-assisted colonoscopy on these outcomes is unknown. What Does This Mean? AI-assisted colonoscopy may improve clinical outcomes and be more cost-effective than conventional colonoscopy for detecting polyps, adenomas, precancerous lesions, and colorectal cancer. The clinical and cost-effectiveness of different types of AI-assisted colonoscopy systems compared to each other is unknown. Clinicians and decision-makers can use the evidence summarized in this review to inform decisions regarding the implementation of AI-assisted colonoscopy.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".