Clinical application of artificial intelligence in colonoscopy and endoscopy for gastrointestinal bleeding: New techniques and outcomes.Review

Background and aims: Artificial intelligence (AI) is rapidly expanding in various fields of medicine.In recent years, researchers have used artificial intelligence and especially machine learningalgorithms to analyze high-volume medical data and perform clinical tasks such as identifyingand classifying upper and lower gastrointestinal-bleeding (GI-bleeding) in endoscopy and colonoscopyimages and videos, and following It has been used as an influential factor to enhance riskassessment and advance diagnosis, treatment and standard care around endoscopy and colonoscopyand other medical decision-making. Our goal is to examine studies related to clinical applicationsof artificial intelligence in gastrointestinal endoscopy and colonoscopy and its potential rolein medical decision-making regarding gastrointestinal-bleeding.Method: we searched for this topic in reliable scientific databases PubMed according to thekeywords, which included artificial intelligence, machine learning, medicine, gastrointestinalbleeding, endoscopy, colonoscopy and classification. all keywords queries were considered as asupplementary file.Results: The results show the improvement of the process of examination, diagnosis, classificationand differentiation and overall analysis of images and videos in endoscopy and colonoscopywith the help of artificial intelligence techniques. Studies have told us that more accurate andfaster and automatic diagnosis of bleeding sites, blood remains, wounds, tumors and various intestinaldiseases, inflammatory areas, digestive infections such as Helicobacter pylori infection,cancer, identification of the depth of cancer invasion, Dysplasia in Barrett’s esophagus, Predictionof disease recurrence, prediction of pathological diagnosis and many abnormalities are thingsthat artificial intelligence has brought to us in the field of endoscopy and colonoscopy. Further, ithas been seen that if the endoscopist or colonoscopist performs artificial intelligence and endoscopyor colonoscopy together, the accuracy of endoscopy or colonoscopy diagnosis increases.The use of artificial intelligence systems, especially machine-learning, with several prospectivepatient-based studies during gastrointestinal endoscopy and colonoscopy significantly better diagnosedgastrointestinal-bleeding in patients with suspected bleeding and the rate of gastrointestinal-bleeding, and in the class Automatic classification of bleeding types performed better thanprevious methods. Also, in a study, it was shown that experts with an automated artificial intelligencesystem they were able to accurately identify small bowel angioectasia, which is the mostcommon cause of bleeding in patients with obscure gastrointestinal-bleeding. All this goes handin hand with artificial intelligence having a potential and influential role in medical treatment andrecommendations and in medical decision-making generally.Conclusion: In this review, we highlighted the future insights of artificial intelligence in endoscopyand colonoscopy in medical decision-making, especially in the field of gastrointestinal-bleeding.However, artificial intelligence will continue to develop and be used in daily clinical practiceand will increase its role in medicine. For further progress and development in this field, we mustfocus on integrating artificial intelligence systems with current endoscopy and colonoscopy platformsand electronic medical records, develop training modules to teach clinicians how to use artificial intelligence tools, and determine the best tools for regulation and confirmation of newartificial intelligence technologies.

کلیدواژه ها:

artificial intelligence ، medicine ، colonoscopy ، endoscopy ، gastrointestinal bleeding ، classification