No significant difference in ADR with computer-aided detection vs. standard colonoscopy
Click Here to Manage Email Alerts
Key takeaways:
- The mean adenoma detection rate was 36.8% with the computer-aided polyp detection system and 33.8% with standard colonoscopy.
- The polyp detection rate was 50.8% vs. 50.5%, respectively.
A computer-aided polyp detection system did not improve adenoma or polyp detection among colorectal cancer screening candidates in an outpatient setting in Germany, according to data presented at UEG Week.
“As we all know, randomized clinical trials examining the effects of [computer-aided polyp detection (CADe)] systems regarding ADR were all positive in the beginning,” Alexander Hann, Dr. med., deputy head of gastroenterology at the University Hospital Würzburg, said. “But, as you probably also know, we also had two randomized controlled trials that presented no benefit of CADe. During that time, we decided to develop our own CADe system that was publicly funded.”
Using more than 500,000 manually labeled images from six centers, Hann and colleagues developed the EndoMind CADe system and compared it with traditional colonoscopy among 928 adults at five outpatient facilities from November 2021 to November 2022. Individuals with an indication for CRC screening, post-polypectomy surveillance or a positive fecal immunochemical test were eligible for participation, and all endoscopists had at least 10 years of experience.
The primary outcome of interest was adenoma detection rate, while secondary outcomes included polyp detection rate and withdrawal time.
According to results, 457 participants underwent CADe and 471 underwent traditional colonoscopy, with an indication for CRC screening higher than 85% in both groups.
Researchers reported no significant differences in outcomes between groups, with a mean ADR of 36.8% in the CADe group vs. 33.8% in the traditional colonoscopy group, and a polyp detection rate of 50.8% vs. 50.5%, respectively.
Further, median withdrawal time was 7.5 minutes vs. 7.4 minutes, respectively, and median time without interventions was 6.9 minutes for both groups.
“There is some conflicting data about the influence of CADe on ADR, and we already saw some studies here that underline this conflicting data,” Hann said. “In our study, we couldn’t present a significant difference of CADe on adenoma detection rate.”
Collapse
Hann A, et al. Screening modified by artificial intelligence in endoscopy (SMARTIE): A multicenter randomized clinical trial. Presented at: UEG Week; Oct. 14-17, 2023; Copenhagen, Denmark (hybrid meeting).
Read more about
Click Here to Manage Email Alerts