Everyone has now heard of Chat GPT, but artificial intelligence (AI) also leads to very concrete applications in the medical field. Since October, the colorectal cancer screening service at the Nord Franche-Comté hospital (HNFC) has had two Olympus boxes capable of analyzing images in digestive endoscopy. In summary, the AI is able to detect polyps, these growths which form inside the colon and which must be removed to prevent the appearance of cancer. Pascal Mathis, director of the HNFC, describes this new equipment as avant-garde.
“The consultations will be shorter”
“We remain attentive to the development of this new technology”, assures the director. These devices, at a total cost of 60,000 euros, were financed by the Montbéliard committee of the League once morest cancer. Its president Alain Monnier says he is “very proud of this investment. The consultations will be shorter, the hospital will be able to consult more patients and the care will be more efficient”.
The second deadliest cancer
Colorectal cancer is the cause of 17,000 deaths per year in France for 43,000 new cases diagnosed. It is the second deadliest following that of the lung. Screening tests are strongly recommended from the age of 50 because, detected early enough, it is cured in 90% of cases. This cancer is preluded by the appearance of polyps.
“AI is a psychological aid”
“By chaining colonoscopies, it is possible to miss very fine ones at the end of the day”, explains Doctor Serge Fratté, head of the gastroenterology department. “AI is a psychological aid. We can hope to eliminate several hundred more polyps per year thanks to it,” says the doctor. The HNFC performs more than 3,000 colonoscopies annually, 1,800 of which reveal the presence of polyps. “The goal now is to stop having the six out of 1,000 patients who get colorectal cancer following having had a colonoscopy within four years. »
Thanks to its huge store of data, the Olympus box detects polyps live during the examination, including the most subtle, invisible or almost invisible to the naked eye. Moreover, the AI is not just an image bank, it corrects itself and perfects itself following each examination. Concretely, during colonoscopy, the AI integrates with the video system and a green square appears around the polyp when it is detected. The location is kept in memory (so the doctor can easily return to it to remove it).
Doctor Fratté hopes that this progress will be followed by an increase in the number of screenings: currently, it is only carried out by 37% of the population concerned in our region.