Copyright ©The Author(s) 2021.
World J Gastroenterol. Jun 7, 2021; 27(21): 2818-2833
Published online Jun 7, 2021. doi: 10.3748/wjg.v27.i21.2818
Table 2 Artificial intelligence applications in colorectal cancer pathology
No. of cases/data set
Machine learning method
Xu et al[38]Tumor classification: 6 classes (NL/ADC/MC/SC/PC/CCTA)717 patchesAlexNetAccuracy (97.5%)
Awan et al[87]Tumor classification: Normal/Low-grade cancer/High-grade cancer454 casesNeural networksAccuracy (97%, for 2-class; 91%, for 3-class)
Haj-Hassan et al[37]Tumor classification: 3 classes (NL/AD/ADC)30 multispectral image patchesCNNAccuracy (99.2%)
Kainz et al[88]Tumor classification: Benign/Malignant165 imagesCNN (LeNet-5)Accuracy (95%-98%)
Korbar et al[36]Tumor classification: 6 classes (NL/HP/SSP/TSA/TA/TVA-VA)697 casesResNetAccuracy (93.0%)
Yoshida et al[35]Tumor classification1328 colorectal biopsy WSIsMLAccuracy (90.1%, adenoma)
Alom et al[89]Tumor microenvironment analysis: Classification, Segmentation and Detection21135 patchesDCRN/R2U-NetAccuracy (91.1%, classification)
Bychkov et al[42]Prediction of colorectal cancer outcome (5-yr disease-specific survival).420 casesRecurrent neural networksHR of 2.3, AUC (0.69)
Weis et al[90]Evaluation of tumor budding401 casesCNNCorrelation R (0.86)
Ponzio et al[91]Tumor classification: 3 classes (NL/AD/ADC)27 WSIs (13500 patches)VGG16Accuracy (96 %)
Kather et al[34]Tumor classification: 2 classes (NL/Tumor)94 WSIsResNet18AUC (> 0.99)
Kather et al[34]Prediction of microsatellite instability360 TCGA- DX (93408 patches), 378 TCGA- KR (60894 patches)ResNet18AUC: TCGA-DX—(0.77, TCGA-DX; 0.84, TCGA-KR)
Kather et al[26]Tumor microenvironment analysis: classification of 9 cell types86 WSIs (100000)VGG19Accuracy (94%-99%)
Kather et al[26]Prognosis predictions1296 WSIsVGG19Accuracy (94%-99%)
Kather et al[26]Prognosis prediction934 casesDeep learning (comparison of 5 networks)HR for overall survival of 1.99 (training set) and 1.63 (test set)
Geessink et al[29]Prognosis prediction, quantification of intratumoral stroma129 casesNeural networksHRs of 2.04 for disease-free survival
Sena et al[40]Tumor classification: 4 classes (NL/HP/AD/ADC)393 WSIs (12,565 patches)CNNAccuracy (80%)
Shapcott et al[92]Tumor microenvironment analysis: detection and classification853 patches and 142 TCGA imagesCNN with a grid-based attention networkAccuracy (84%, training set; 65%, test set)
Sirinukunwattana et al[31]Prediction of consensus molecular subtypes of colorectal cancer1206 casesNeural networks with domain-adversarial learningAUC (0.84 and 0.95 in the two validation sets)
Swiderska-Chadaj et al[93]Tumor Microenvironment Analysis: Detection of immune cell, CD3+, CD8+28 WSIsFCN/LSM/U-NetSensitivity (74.0%)
Yoon et al[39]Tumor classification: 2 classes (NL/Tumor)57 WSIs (10280 patches)VGGAccuracy (93.5%)
Echle et al[46]Prediction of microsatellite instability8836 casesShuffleNet Deep learningAUC (0.92 in development cohort; 0.96 in validation cohort)
Iizuka et al[22]Tumor classification: 3 classes (NL/AD/ADC)4036 WSIsCNN/RNNAUCs (0.96, ADC; 0.99, AD)
Skrede et al[28]Prognosis predictions2022 casesNeural networks with multiple instance learningHR (3.04 after adjusting for established prognostic markers)