Artificial Intelligence in Gastroenterology
1Swedish Medical Center , Seattle, USA
2University of Cincinnati College of Medicine, Cincinnati, USA
3Olive View-UCLA Medical Center, Sylmar, USA
4Stanford University School of Medicine, Stanford, USA
5All India Institute of Medical Sciences, New Delhi, India
Artificial Intelligence in Gastroenterology
Description
The concept of Artificial Intelligence (AI) encompasses computational automated programs that mimic human intelligence, including learning and problem solving. This concept deeply intertwines with machine learning (ML) and deep learning (DL), wherein computer-based methods analyze data to formulate predictive models. Over the past decade, AI as well as ML and DL domains have been widely and increasingly utilized in the interpretation and management of clinical disorders, including gastroenterological and hepatologic disorders. Within the realm of gastroenterology, AI has been investigated in a variety of disease processes, with particular value and promise in endoscopy to enhance detection of malignant and pre-malignant lesions (in esophagus, stomach, and colon), gastrointestinal bleeding, inflammatory bowel disease (IBD), and pancreatobiliary diseases. For example, AI has been successfully utilized to detect malignant and pre-malignant colonic lesions, resulting in improved diagnostic accuracy as well as cancer prognostication. AI algorithms are also being studied to identify patients with IBD and celiac disease and to predict treatment response in such patients. Using images collected during wireless capsule endoscopy, AI has been utilized to detect small bowel bleeding. Amongst pancreatobiliary diseases, AI has been studied for early detection of pancreatic cancer during endoscopic ultrasound, and amongst liver diseases for detection of fibrosis in viral hepatitis and fatty liver disease.
AI technology is evolving rapidly, though certain obstacles have been identified. A major limitation currently is the paucity of high quality datasets for development of ML and DL algorithms. Additional gaps in this field include varying performance of AI algorithms in different GI diseases, the heterogeneity of performance metrics, and lack of rigorous randomized controlled trials comparing AI-assisted approaches to current non-AI-based approaches.
This Special Issue will focus on the evolution and applications of AI in GI. Original research and review articles are welcome.
Potential topics include but are not limited to the following:
- AI in esophageal pre-malignant (e.g. Barrett’s esophagus) and malignant diseases (esophageal cancer)
- AI in gastric pre-malignant (H. pylori, intestinal metaplasia) and malignant diseases (gastric cancer)
- AI in colonic pre-malignant (polyps) and malignant diseases (colorectal cancer)
- AI in pancreatic pre-malignant (mucinous cysts – IPMN and MCN) and malignant diseases (pancreatic cancer), as well as benign diseases (chronic pancreatitis)
- AI in biliary pre-malignant (indeterminate strictures) and malignant diseases (cholangiocarcinoma)
- AI in gastrointestinal bleeding (small bowel or variceal or gastric or colonic)
- AI in IBD (identification and treatment response), celiac diseases, parasitic diseases, and other benign luminal diseases
- AI in liver diseases – fatty liver disease, hepatitis, cirrhosis
- AI in GI endoscopy – advances and metrics
- ML and DL datasets for development of AI algorithms
- Performance metrics of AI algorithms
- Development of new therapeutic targets via synthesis of molecular, genetic, and clinical data from large datasets
- AI technologies to monitor patients’ health remotely