Emerging molecules, tools, technology, and future of surgical knife in gastroenterology

Table 1 Some of the applications of machine learning in gastroenterology[49]

Modality	Clinical presentation/ diagnosis	Application
Upper GI endoscopy	Barrett’s oesophagus	Identification of early cancerous lesion; target site for biopsy; endoscopic assistance
	Oesophageal cancer	In the diagnosis of SCC
	H. pylori infection	Atrophy vs metaplasia
	Gastric cancer	Tumor vs non tumorous tissue; depth of invasion
Capsule endoscopy	GI bleed	Source of bleed; detecting pathologic lesions such as erosions and ulcers
	Celiac	Finding villous atrophy
Colonoscopy	Colorectal cancer	Bowel preparation assessment; adenoma detection; assistance
	Ulcerative colitis	Severity and relapses
Ultrasound-based test-fibro scan/elastography	Various liver diseases; benign & malignant	Fibrosis stage
	Pancreatic diseases	Tumour assessment, degree of intrapancreatic fat
	GI pathology	Survival prediction in colorectal cancer; identification of MSI; HCC vs cholangiocarcinoma; predict prognosis and survival in HCC

GI: Gastrointestinal; SCC: Squamous cell carcinoma; H. Pylori: Helicobacter pylori; HCC: Hepatocellular carcinoma.

Table 2 Current use and application of newer tools and technology in gastrointestinal surgery

Disease/pathology	Aim	Artificial intelligence used	Ref.
GERD, atrophic gastritis	Diagnosis	Artificial neural network	Pace et al[45], 2005; Lahner et al[50], 2005
Acute pancreatitis	Prediction of prognosis- comparing with RANSON, APACHE score	Artificial neural network	Yang et al[51], 2019
Ulcerative colitis	Prediction of prognosis after apheresis therapy	Artificial neural network	Takayama et al[52], 2015
Colorectal cancer	To predict lymph node metastasis	SVM	Ichimasa et al[53], 2018
	Recognition of ureter in endoscopic images	Convolutional neural network	Harangi et al[54], 2017
	To distinguish normal colonic mucosa from malignant lesions with confocal laser endomicroscopy	Fractal analysis and neural network modelling	Ştefănescu et al[55], 2016
	Differentiation of colonic carcinoma from adenoma and healthy mucosa using hyperspectral imaging	Artificial neural network	Jansen-Winkeln et al[56], 2021
Gastric cancer	To define safe dissection planes	Deep learning model based on u net	Kumazu et al[57], 2021
GI bleed	Predicting mortality, recurrent bleeding, need for therapeutic intervention	Artificial neural network	Das et al[58], 2003
GI polyp	On colonoscopy	SVM	Billah et al[59], 2017
HBV related fibrosis	Shear wave elastography	Convolutional neural network	Wang et al[60], 2019
Pancreatic adenocarcinoma	Image analysis of endoscopic ultrasound	Multilayer perception network	Săftoiu et al[61], 2012

GI: Gastrointestinal; GERD: Gastroesophageal reflux disease; APACHE: Accuracy of Acute Physiology and Chronic Health Evaluation; HBV: Hepatitis B virus; SVM: Support vector machine.