Table 4.

Summary of AI models for CRC prognosis session. (C-index, concordance index; LR, logistic regression; DT, decision tree; GB, gradient boosting; LightGBM, light gradient boosting machine; CNN, convolutional neural network; AUC, area under curve; PET-CT, positron emission tomography or computed tomography; HR, hazard ratio; GSEA, gene set enrichment analysis; PPI, protein-protein interaction; HE, hematoxylin and eosin; WSI, whole slide image; MLP, multilayer perceptron; AdaBoost, adaptive boosting; LSTM, long short-term memory; EHR, electronic health record; SVM, support vector machine; NB, naïve Bayesian; KNN, K-nearest neighbors; NN, neural network; RF, random forest).

Topic	Task	Dataset	Model	Performance	Year	Ref.
Recurrence	Recurrence perdition of stage II CRC	Clinicopathological data of 350 patients after curative resection for stage II CRC	Nomogram	C-index: 0.585 in the validation set	2020	[88]
	Recurrence prediction of Stage IV CRC after tumor resection	EHR data from 999 patients of stage IV CRC	LR, DT, GB and LightGBM	LightGBM: AUC: 0.761	2020	[89]
	Recurrence prediction of local tumor	PET-CT images from 84 patients	CNN, Proportional hazards model	C-index: 0.64	2019	[90]
	Risk prediction of recurrence of gastrointestinal stromal tumor	Clinical data of 2560 patients	Proportional hazards, Non-linear model	AUC: 0.88	2012	[91]
	Recurrence perdition after surgery	Clinicopathological data of 1320 nonmetastatic CRC patients	NomogramCOX regression	C-index: 0.77	2008	[92]
Survival	Genetic risk factors Identification	National Center for Biotechnology Information Gene Expression Omnibus	GSEA, PPI network, Cox Proportional Hazard regression	4 sub-networks and 8 hub genes as potential therapeutic targets	2021	[93]
	Prognostic prediction for stage III CRC	Clinicopathological data of 215 patients	CNN, GB	HR: 8.976 and 10.273	2020	[94]
	Outcome prediction	12,000,000 HE images	CNN	HR: 3.84 and 3.04 with established prognostic markers	2020	[95]
	Survival prediction	7180 HE images of 25 patients	CNN	Nine-class accuracy: >94%	2019	[96]
	Survival prediction	PET-CT images of 84 patients	CNN, proportional hazards model	C-index: 0.64	2019	[90]
	Outcome prediction, and remaining lifespan prediction	SEER	tree-based ensemble model	Accuracy: 0.7069, Sensitivity: 0.8452, Specificity: 0.66	2019	[97]
	Outcome prediction	75 WSIs from stage I and II CRC patients with surgical resection	CNN	F1: 0.67	2019	[98]
	Outcome prediction	EHR data of 58,152 patients	CNN	AUC: 0.922, Sensitivity: 0.837, specificity: 0.867, PPV: 0.532	2019	[99]
	Prediction of Stages and Survival Period	Clinicopathological data of 4021 patients	RF, SVM, LR, MLP, KNN, and AdaBoost	RF: F-measure: 0.89, Accuracy: 84%, AUC: 0.82 ± 0.10	2019	[100]
	1/2/5 years Survival prediction	SEER data	DNN	AUC: 0.87	2019	[101]
	Outcome prediction	Digitized HE tumor tissue microarray samples of 420 patients	CNN, LSTM	LSTM: AUC: 0.69, histological grade AUC: 0.57, the visual risk score AUC: 0.58	2018	[102]
	5-year survival prediction	EHR data of 1127 CRC patients	Ensemble (bagging and voting) classifier	Ensemble voting model AUC: 0.96	2017	[103]
	5-year survival prediction	EHR data of 334,583 cases from Robert Koch Institute	SVM, LR, NB, DT, KNN, LR, NN, RF	Average accuracy of the clinicians: 59%, ML: 67.7%	2015	[104]