AUTHOR=Chen Guihua , Shen Jun 

TITLE=Artificial Intelligence Enhances Studies on Inflammatory Bowel Disease

JOURNAL=Frontiers in Bioengineering and Biotechnology

VOLUME=Volume 9 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/bioengineering-and-biotechnology/articles/10.3389/fbioe.2021.635764

DOI=10.3389/fbioe.2021.635764

ISSN=2296-4185

ABSTRACT=Inflammatory bowel disease (IBD) studies have accumulated extensive digital images and medical record resources, requiring the employment of artificial intelligence (AI) for big data analysis. AI-assisted diagnosis and image analysis have proven effective in many fields including gastroenterology. Studies on IBD, a complex multi-factor disease, can also be boosted utilizing AI technology. The complete etiology of IBD is still uncertain. However, some IBD risk genes have been identified using AI with existing gene analysis techniques such as genome-wide association studies (GWAS). Similar to other gastrointestinal diseases, AI-assisted diagnosis and prognosis prediction have shown promising results in IBD studies. Before deep-learning development, support vector machines (SVM) achieved the best performance for classification and regression. The convolutional neural network was then considered an excellent tool for image analysis among several deep-learning models. Finally, artificial neural networks can represent the multifaceted interactions between clinical characteristics, environment, and demographics. AI application faces hurdles when the accuracy of diagnosis and prognosis relies on the best adjustability between AI technology and the dataset. This review aims to summarize AI application in the field of IBD study and objectively evaluate the performance of these methods to understand the algorithm-dataset combination in studies.