EVALUATION OF ARTIFICIAL INTELLIGENCE ENABLED COMPUTER AIDED DETECTION ASSISTANCE IN DETECTING COLON POLYPS IN THE COMMUNITY (AI-SEE): A MULTICENTER RANDOMIZED CLINICAL TRIAL

Background: Crohn’s disease (CD) is a chronic inflammatory intestinal disorder, for which the majority of patients need to undergo surgery. Following ileocolonic resection, most patients experience recurrent CD in the neoterminal ileum. Endoscopic lesions usually precede symptoms and predict the severity of the disease course. No treatments have been approved for recurrence-prevention CD. REPREVIO is a prospective placebo-controlled randomized trial investigating the preventive effect of vedolizumab, an anti-integrin antibody, on recurrence of CD.
Methods: Following ileocolonic resection, patients were randomized to treatment with intravenous vedolizumab (300 mg at week 0,8,16 and 24) or PLC (1:1) at 12 sites in the Netherlands, France, Italy and Spain. Patients and investigators were blinded to treatment assignment that was initiated within 4 weeks of ileocolonic resection with anastomosis. Twenty-six weeks after randomization patients underwent ileocolonoscopy for assessment of recurrent CD lesions. Video recordings were centrally scored using the modified Rutgeerts’ score (i0-i4) by 2 readers with adjudication in case of disagreement. The primary endpoint was endoscopic recurrence (ER) of Crohn’s disease according to Rutgeerts classification categories (non-parametric analysis of distributions between groups); secondary endpoints were the proportion of patients with ER >i2a, endoscopic remission (i0) and clinical recurrence (CDAI increase >70 points between baseline and week 24). Data analysis was based on intent-to-treat with patients dropping out or not having final endoscopy counted as failure (i4). Adverse events were recorded.
Results: Of 95 pts screened, 80 were randomized. Seventy-six patients reached week 26 (2 drop-outs for withdrawal of consent, 1 for perforation and 1 for cancer). Baseline patient characteristics were comparable as shown in table 1. Using a non-parametric analysis, patients in the treatment group had a 77.8% (95% CI 66.4 to 86.29%) chance of having a better Rutgeerts’ score than control patients (p <0.0001). Based upon the dichotomized analysis of high risk recurrence versus low risk recurrence (i0-2A versus i2B-i4), 77% of patients who received VDZ had i0-2A versus 38% for PLC, yielding an absolute difference of 39% in favor of VDZ (p=0.0004)(figure 1). Endoscopic remission (i0) was observed in 18/43 pts for VDZ (42%) versus 1/37 pt for PLC (3%)(p<0.001). Clinical recurrence (CDAI increase >70 pts between baseline and week 24) occurred in 9 pts on VDZ and 8 pts on PLC. No new safety signals were observed.
Conclusion: Treatment with VDZ started immediately after ileocolonic resection is highly effective for the prevention of postoperative recurrence of CD.

Background: There has been increasing interest in artificial intelligence in gastroenterology. To reduce miss rate during colonoscopy, there has been significant exploration in computer aided detection (CADe) devices. In this study, we evaluate the use of CADe in colonoscopy in community based, non-academic practices.
Methods: Between September 28, 2020 and September 24, 2021, a randomized controlled trial (AI-SEE) was performed evaluating the impact of CADe on polyp detection in 4 private practice endoscopy centers in the USA. Patients were block randomized to undergoing colonoscopy with or without CADe. Primary outcomes measured were adenomas per colonoscopy (APC) and adenomas per extraction (APE; the percentage of polyps removed that are adenomas). Secondary endpoints included serrated polyps per colonoscopy, non-adenomatous, non-serrated polyps per colonoscopy, adenoma and serrated polyp detection rate, and procedural time.
Results: A total of 769 patients were enrolled (387 with CADe), with similar patient demographics between the two groups. There was no significant difference in adenomas per colonoscopy in the CADe and non-CADe groups (0.73 vs 0.67, p=0.496). While use of CADe did not improve identification of serrated polyps per colonoscopy (0.08 vs 0.08, p=0.965), use of CADe increased identification of non-adenomatous, non-serrated polyps per colonoscopy (0.90 vs 0.51, p<0.0001), resulting in a lower APE in the CADe group (Table 1). Adenoma detection rate (35.9 vs 37.2%, p=0.774) and serrated polyp detection rate (6.5 vs 6.3%, p=1.000) were similar in the CADe and non-CADe group (Table 2). Mean withdrawal time was longer in the CAD compared to non-CADe group (11.7 vs 10.7 minutes, p=0.003). However, when no polyps were identified, there was similar mean withdrawal time (9.1 vs 8.8 minutes, p=0.288). There were no adverse events.
Conclusions:
To our knowledge, this is the first prospective multicenter study to assess the clinical impact of real-time CADe in primarily non-academic centers. In contrast to prior studies, use of CADe did not result in a statistically significant difference in the number of adenomas detected. Additional studies are needed to better understand why some endoscopists derive substantial benefits from CADe and others do not.