Improving the Algorithms
Experts agree that current and future research will improve the accuracy and quality of AI colonoscopy for all users, leading to new standards and more consistent outcomes in both clinical trials and real-world applications.
Work underway now to improve the algorithms will be an important step in that direction, according to Dr. Christie.
“We need to have enough information to create AI algorithms that allow us to detect early lesions, at least from an imaging standpoint, and we need to improve and increase the sensitivity and the specificity, as well as the predictive value,” she said.
AI can also play a role in health equity, she noted.
“But it’s a double-edged sword, because it depends again on algorithms and machine learning. Perhaps AI can eliminate some of the bias in our clinical decision-making. However, if we don’t train the machine properly with a good, diverse sample of patients and figure out how to integrate some of the social determinants of health that a computer may not otherwise consider, it can create larger disparities and larger biases. AI devices can only be as good and as inclusive as we make them,” Dr. Christie said.
Looking Ahead
Dr. Sharma predicts that “the next slew of studies are going to be on characterization — not just saying there’s an abnormality but distinguishing it further and saying whether the lesion is noncancerous, precancerous, or cancer.”
Other studies will focus on quality improvement of factors, such as withdrawal time and bowel preparation.
In its clinical practice update on AI, the American Gastroenterological Association states, “Eventually, we predict an AI suite of tools for colonoscopy will seem indispensable, as a powerful adjunct to support safe and efficient clinical practice. AI tools that improve colonoscopy quality may become more accepted, and perhaps demanded, by payors, administrators, and possibly even by well-informed patients who want to ensure the highest-quality examination of their colon.”
Dr. Sharma and Dr. Christie disclose no relevant conflicts of interest.
A version of this article appeared on Medscape.com.