Minireviews
Copyright ©The Author(s) 2021.
Artif Intell Gastroenterol. Apr 28, 2021; 2(2): 42-55
Published online Apr 28, 2021. doi: 10.35712/aig.v2.i2.42
Table 1 Recent developments in artificial intelligence assisted diagnosis
AI category
Data adopted
Advantages
Control
Ref.
ANNPreoperative serum AFP, tumor number, size and volumeThe ANN showed higher AUCs in identifying tumor grade (0.94) and MVI (0.92)LR model (0.85 and 0.85)[20]
CNNEnhanced MRIThe CNN showed comparable accuracy (90%)Traditional multiphase MRI (89%)[24,25]
Open-source framework “caffe” based CNN modelDWICNN trained with three sets of b-values found better grading accuracy (80%)CNN trained with different b-values (65%, 68%, 70%)[26]
CNNNonenhanced MRIThe deeply supervised and pretrained CNN model performed better in characterizing HCC (accuracy 77.00 ± 1.00%)CNN-based method pretrained by ImageNet (65.00 ± 1.58%)[27]
DL-based segmentation modelContrast-enhanced CTThe model with a combination of 2D multiphase strategy showed higher ability of segmenting active part from the tumorsTraditional CT estimation[28-30]
RF based ML modelHE-stained histopathological imagesThe classifying model showed an AUC of 0.988 in the test set and 0.886 in the external validation set-[31]
1D CNNHyperspectral and HE-stained imagesThe models had a higher average AUC of 0.950RF (0.939) and SVM (0.930) models[33]
Shiny and Caret packages-based prediction modelClinical and laboratorial informationThe optimal model had an AUC of 0.943Single factor-based predictors (0.766, 0.644 and 0.683)[34]