The use of AI to aid diagnosis has been heralded increasingly often as an imminent evolution in the world of gastroenterology (1, 2). Thanks to an Italian study, in April the FDA has approved for the first time a tool for colonoscopy that helps doctors detect suspicious lesions in real time whilst performing the examination (3, 4).
Colonoscopy is considered the best tool for the prevention and diagnostic confirmation of colorectal cancer, as it allows us to identify colon polyps or other small changes in the intestinal mucosa that could develop into a tumour. Some polyps, however, 'escape' the diagnosis or cannot be correctly classified by colonoscopy alone (5).
This is where the recently approved artificial intelligence endoscopic device, called GI Genius, can help. By generating markers, accompanied by low-volume sounds and superimposing them on the video from the endoscope camera upon detection of a potential lesion, it suggests to the operator and specialists that further assessment may be required, such as a more thorough visual evaluation, tissue sampling and testing, removal, or ablation (4).
Following approval by the EMA in 2020, which attributed the CE mark necessary for commercialisation in Europe (6), the FDA evaluated the safety and efficacy of the device based on an all-Italian multicentre study on 685 individuals, which showed that the use of artificial intelligence applied to traditional colonoscopy was able to identify laboratory-confirmed adenomas or carcinomas with 14% increased diagnostic accuracy (i.e. absolute increase in adenoma detection rate) (4). This novel tool does not, however, replace the specialist or laboratory tests, and does not suggest how to manage suspected polyps. It remains up to the doctor to decide whether the lesion is suspicious, and how to proceed according to standard clinical practice and guidelines.
For this reason, educating and training specialists should remain a high priority in order to optimise identification and management of pre-cancerous lesions, thereby successfully preventing or delaying the development of colorectal cancer.
At the Corpus, we are looking forward to learning about more exciting advances in the field of endoscopy applied to gastroenterology at the upcoming UEG week virtual meeting, where this subject features in several prominent sessions. If you want to know more about how we have already approached the topic of diagnostics in gastroenterology, particularly in relation to ulcerative colitis and Crohn’s disease, visit our website: https://www.the-corpus.com/.
References
1. Attardo S et al. World J Gastroenterol 2020;26:5606-5616
2. Mascarenhas M et al. Ann Gastroenterol 2021;34:1-10
3. FDA News Release. https://www.fda.gov/news-events/press-announcements/fda-authorizes-marketing-first-device-uses-artificial-intelligence-help-detect-potential-signs-colon (Accessed June 2021)
4. Repici A et al. Gastroenterology 2020;159(2):512-520
5. Stauffer CM & Pfeifer C. StatPearls 2021 https://www.ncbi.nlm.nih.gov/books/NBK559274/ (Accessed June 2021)
6. Cosmo Pharmaceuticals. Press Release. https://www.cosmopharma.com/news-and-media/press-releases-and-company-news/2020/200220 (Accessed June 2021)
The use of AI to aid diagnosis has increasingly been heralded as an imminent evolution in the world of gastroenterology. As the FDA approves the first AI endoscopic device, what will it mean for physicians?
