Endoscopy · Real-time computer-aided polyp detection (CADe)
GI Genius
Cosmo Intelligent Medical Devices (distributed by Medtronic)
GI Genius is the reference point for AI in colonoscopy — the first device the FDA authorised for the field, cleared by De Novo in April 2021 (DEN200055, first-in-class product code QNP), and the subject of the landmark Repici randomised trial that reported a 30 percent relative rise in adenoma detection. What makes it a two-mark rather than three-mark entry is that the evidence has matured into something more honest than the headline. The benefit is real in general screening and surveillance populations, but a 3,213-patient Spanish randomised trial in a FIT-positive screening programme found no improvement in the detection of advanced neoplasia, and independent meta-analyses report substantial heterogeneity in the size of the CADe effect across trials and operators. The regulatory footprint is broad — a first-in-class FDA De Novo, a string of follow-on 510(k) clearances, and an EU MDR CE mark — but the device's own registers give it a strong, not unqualified, case.
Performance Metrics
Clinical Evidence
GI Genius has one of the strongest randomised evidence bases of any device in the catalogue, and also one of the clearest illustrations of why a single positive trial should not be read as the whole story. The foundational efficacy study is the Repici multicentre randomised trial (Repici A, et al., Gastroenterology 2020;159:512-520, PMID 32371116). In 685 patients, adding GI Genius to high-definition colonoscopy raised the adenoma detection rate from 40.4 percent to 54.8 percent — an absolute increase of about 14 percentage points and a relative increase of roughly 30 percent (rate ratio ~1.30). This is the trial cited in the De Novo record and the basis for the device's clinical reputation. Pooled analyses of the early randomised literature were consistent with it: a meta-analysis of five randomised trials (4,354 patients) reported a pooled adenoma-detection rate ratio around 1.44, though with overlapping authorship across the constituent trials. The counterweight is the FIT-positive screening evidence. A multicentre Spanish randomised trial (Mangas-Sanjuan C, et al., Ann Intern Med 2023;176:1145-1152, PMID 37639723) enrolled 3,213 participants with a positive faecal immunochemical test and found no difference in the detection of advanced colorectal neoplasia between colonoscopy with and without computer-aided detection, even though the AI arm did detect more non-polypoid and smaller lesions. Because advanced-neoplasia yield is the outcome that maps most directly to cancer prevention, this null result matters. Later synthesis has reinforced the nuance rather than resolved it: a 2024 meta-regression (Patel HK, et al., Clin Gastroenterol Hepatol 2024, PMID 38056803) and a related meta-regression analysis (PMID 39924430) both document wide heterogeneity in the CADe effect on adenoma detection across trials, with the operator's own baseline detection rate a large source of that variability — the benefit is smaller, or absent, among already-high detectors. The editorial reading is that GI Genius reliably increases the detection of small and non-polypoid adenomas, but that the downstream benefit in advanced neoplasia — and therefore in the outcomes procurement committees ultimately care about — is population- and operator-dependent and not established in the FIT-positive screening setting.
| Study | Design | n | Sensitivity | Specificity | AUC | Published |
|---|---|---|---|---|---|---|
| Repici A, Badalamenti M, Maselli R, et al. (investigator-conducted RCT) | RCTRCT | 685 | — | — | — | Gastroenterology, 2020;159:512-520; PMID 32371116; ADR 54.8% vs 40.4% (RR ~1.30) |
| Mangas-Sanjuan C, de-Castro L, Cubiella J, et al. (independent, FIT-positive screening RCT) | RCTRCT | 3,213 | — | — | — | Ann Intern Med, 2023;176:1145-1152; PMID 37639723; no difference in advanced-neoplasia detection |
| Patel HK, et al. — meta-regression of CADe randomised trials (independent) | RCTRCT | 4,354 | — | — | — | Clin Gastroenterol Hepatol, 2024; PMID 38056803; wide heterogeneity in the CADe ADR effect |
Clinical Pulse
Verified clinician sentiment
Verified clinician reviews launching soon · Apply to contribute →
Reviews are signed by clinicians verified against their regulator's public register.
Inside the Auris+ Listing
Five more sections complete this device’s Auris+ Listing.
Decision Ledger
ProPro unlocks a private, cross-vendor log of every case you read on this device — what the AI called, what you concluded, and a one-line reason if you overrode it. Ready for the EU AI Act's deployer logging obligations when they land in 2028.
Clinical Evidence Deep Dive
ProPro unlocks the structured clinical-evidence summary — study count, target patient population, and a tabular accuracy-metrics view drawn from peer-reviewed sources.
Peer-Reviewed Publications
ProPro unlocks the curated peer-reviewed publication list with PubMed cross-links — the citation backbone of every editorial verdict.
Post-Market & Regulatory Conditions
ProPro unlocks the post-market surveillance summary, recall record, and the conditions of approval that bound real-world use.
AI Algorithm Version History
ProPro unlocks the chronological record of algorithm version changes — what changed when, drawn from manufacturer changelogs and regulatory filings.
Regulatory Approvals
Safety Record
No GI Genius recalls or FDA safety communications were identified in publicly available sources as of July 2026. The MAUDE adverse-event database and the openFDA recall endpoint could not be queried directly from this environment for product code QNP, so this should be read as "none found in public reporting" rather than an exhaustive vigilance audit. As a detection-only overlay that neither alters the image nor drives resection decisions, the device's harm surface is dominated by two documented behaviours rather than hardware failure: a false-negative frame (no box) is not exclusionary and does not substitute for careful mucosal inspection, and the tool's tendency to flag non-neoplastic polyps can drive unnecessary polypectomy — an overdiagnosis and overtreatment signal noted across the randomised literature.
Intended Use & Indications
GI Genius is a computer-assisted reading tool intended to aid endoscopists in detecting colonic mucosal lesions such as polyps and adenomas in real time during standard white-light colonoscopy. It is an accessory to the procedure: the device draws attention to a region of the field but does not replace the endoscopist's assessment, does not characterise or diagnose a lesion, and does not substitute for histopathological sampling. The cleared US indication is a detection-and-localisation aid (CADe), distinct from a characterisation (CADx) claim — GI Genius marks where to look, it does not say what a lesion is. In every jurisdiction the final decision to inspect, resect or biopsy remains the endoscopist's, and a negative frame — no box — is not a clearance that a segment is lesion-free.