Monday, May 2, 2022 (HealthDay News) – Artificial intelligence reduces the rate of colorectal neoplasia by almost half, according to a survey published online March 15. Gastroenterology.
Michael B., MD, Florida Mayo Clinic Jacksonville. Wallace and colleagues enrolled 230 patients in one of eight centers in Italy, the United Kingdom and the United States through colorectal cancer screening in two days on the same day, with or without artificial intelligence (AI; deep learning CADe) after colonoscopy.
The researchers found that the adenoma miss rate (AMR) in the arm with AI was 15.5 percent (38 of 246) and non-AI colonoscopy in the first arm (adjusted odds ratio) was 32.4 percent (80 of 247). [OR], 0.38). AMR was particularly low in AI-first cases for lesions ≤5 mm (15.9 vs. 35.8 percent; or, 0.34) and nonpolypoid lesions (16.8 v. 45.8 percent; or: 0.24). Furthermore, AMR was lower in both proximal (18.3 vs. 32.5 percent; or, 0.46) and distal colon (10.8 vs. 32.1 percent; or, 0.25). In the second colonoscopy, the mean number of adenomas was lower in the AI-I group than in the non-AI colonoscopy-I group. In the AI and non-AI first arms, the lie-negative rates were 6.8 percent (three of 44 patients) and 29.6 percent (13 of 44 patients), respectively (or, 0.17).
“AI has resulted in an almost double reduction in the rate of colorectal neoplasia, with standard colonoscopy supporting AI-advantage to reduce perceptual errors for small and delicate lesions,” the authors wrote.
One author revealed the financial relationship with Cosmo Artificial Intelligence-AI Limited, which funded the study.
Abstract / Full Text (may require subscription or payment)