TY - JOUR T1 - The importance of high-quality ‘big data’ in the application of artificial intelligence in inflammatory bowel disease JF - Frontline Gastroenterology JO - Frontline Gastroenterol SP - 258 LP - 262 DO - 10.1136/flgastro-2022-102342 VL - 14 IS - 3 AU - James J Ashton AU - Johanne Brooks-Warburton AU - Patrick B Allen AU - Tony C Tham AU - Sami Hoque AU - Nicholas A Kennedy AU - Anjan Dhar AU - Shaji Sebastian A2 - , Y1 - 2023/05/01 UR - http://fg.bmj.com/content/14/3/258.abstract N2 - The rise of artificial intelligence (AI) in healthcare provides an opportunity to improve clinical care and patient outcomes. Gastroenterology may be seen as a leader in the application of AI, through automatic image recognition and interpretation in endoscopy, and analysis of images gathered through video capsule endoscopy.1 Despite this, the clinical translation of AI into routine practice has lagged behind its application in research settings. Inflammatory bowel disease (IBD) presents specific clinical challenges for which AI may have solutions, including prediction of therapeutic response, novel subgroup classification, precise molecular diagnosis, complication risk stratification and endoscopic image analysis for scoring of severity of mucosal inflammation in both ulcerative colitis and Crohn’s disease.2 We are also moving to an era of big data in IBD research, with consortia collecting and collating large cohorts of patients with available genomic, and other multiomic, data.3 4 This resource presents an unrivalled opportunity to alter the landscape of disease prediction and classification in IBD, and usher in routine personalisation of diagnosis and treatment (figure 1). A key part of reliable, reproducible and applicable use of AI for patients, is the quality of the clinical phenotyping. Lack of in-depth clinical data, systemic bias in data entry, lack of longitudinal outcomes and missing data all pose huge challenges to application of AI, with algorithms reliant on high-quality data input to give high-quality output.5 While this constitutes a challenge, it also creates an opportunity to develop robust systems to gather prospective and retrospective data, in structured ways and to use routinely collected data for alternative purposes.Figure 1 Potential applications of artificial intelligence and ‘big data’ to personalise management in inflammatory bowel disease. Created with BioRender.com.Within this opinion article, we focus on the use of AI to predict outcomes, response to therapy, complications and novel disease … ER -