Perceptions and Expectations
Emerging evidence suggests that endoscopists’ perceptions and expectations may affect assessments of AI’s potential benefits in practice, Dr. Sharma noted.
“Someone might say, ‘I’m a trained physician. Why do I need a machine to help me?’ That can create a situation in which the endoscopist is constantly challenging the device, trying to overrule it or not give it credit.”
Others might perceive that the AI device will definitely help and therefore not look as carefully themselves for adenomas.
A study at The University of Texas MD Anderson Cancer Center in Houston in which activation of the AI system was at the discretion of the endoscopist found that real-time CADe did not improve adenoma detection among endoscopists with high baseline detection rates.
However, despite its availability, AI-assisted colonoscopy was activated in only half of the cases, and multiple concerns were raised by staff and endoscopists in a postprocedural survey. In particular, endoscopists were concerned that the system would result in too many false-positive signals (82.4%), was too distracting (58.8%), and prolonged procedure time (47.1%).
The authors of the Stanford study that found no benefit with CADe in routine practice noted, “Most concerning would be if, inadvertently, CADe use was accompanied by a simultaneous unconscious degradation in the quality of mucosal exposure, possibly due to a false sense of comfort that CADe would ensure a high-quality examination.”
“We’re trying to evaluate some of these interactions between endoscopists and AI devices both pragmatically in practice as well as in clinical trials,” Dr. Sharma said. “Much depends on the context of how you approach and present the devices. We tell physicians that this is an assist device, not something you’re competing against and not something that’s here to replace you. This is something which may make your lives easier, so try it out.”
Are Less Experienced Endoscopists Helped More?
It seems intuitive that less experienced endoscopists would be helped by AI, and indeed, some recent studies confirm this.
A small randomized controlled trial in Japan, presented during the Presidential Plenary at the American Society for Gastrointestinal Endoscopy (ASGE) annual meeting in May 2023, showed that a CADe system was “particularly useful” for beginning endoscopists, who had lower adenoma miss rates with the device vs a white light control device.
Another randomized controlled trial in Japan found that CADe use was associated with an increased overall ADR among endoscopists in training.
But experienced endoscopists probably can benefit as well, noted Jennifer Christie, MD, Division Director, Gastroenterology and Hepatology at the University of Colorado School of Medicine Anschutz Medical Campus in Aurora.
“We know that these AI devices can be useful in training our fellows to detect certain lesions in the colon,” she said. “However, they’re also helpful for many very seasoned practitioners, as an adjunctive tool to help in terms of diagnosis.”
Some studies reflect that dual benefit.
The AID-2 study, designed specifically to look at whether experience had an effect on AI findings during colonoscopy, was conducted among nonexpert endoscopists (lifetime volume of less than 2000 colonoscopies). The researchers, including Dr. Sharma, found that CADe increased the ADR by 22% compared with the control group.
An earlier study, AID-1 , used a similar design but was conducted among experienced endoscopists. In AID-1, the ADR was also significantly higher in the CADe group (54.8%) compared with the control group (40.4%), and adenomas detected per colonoscopy were significantly higher in the CADe group (mean, 1.07) than in the control group (mean, 0.71).
A multivariate post hoc analysis that pooled results from both AID-1 and AID-2 showed that use of CADe and colonoscopy indication, but not the level of examiner experience, were associated with ADR differences. This led the researchers to conclude, “Experience appears to play a minor role as a determining factor for ADR.”
Similarly, a 2023 study from China looked at the mean number of adenomas detected per colonoscopy according to the endoscopist’s experience. All rates were significantly higher in AI-assisted colonoscopies compared with conventional non-AI colonoscopy: overall ADR, 39.9% vs 32.4%; advanced ADR, 6.6% vs 4.9%; ADR of expert endoscopists, 42.3% vs 32.8%; ADR of nonexpert endoscopists, 37.5% vs 32.1%; and adenomas per colonoscopy, 0.59 vs 0.45, respectively.
The authors concluded that “AI-assisted colonoscopy improved overall ADR, advanced ADR, and ADR of both expert and nonexpert attending endoscopists.”