Table 2.

Application of AI in HPV testing.

Reference	Year	Aim of study	Number of subjects	Methods	Results
Wong et al. (29)	2019	Identifying high-grade lesions and in triaging equivocal smears	605 cervical cytology samples	Decision tree, random forest SVM-linear SVM-nonlinear	Specificity: 94.32% Specificity: 90.91% Specificity: 90.91% Specificity: 90.91%
Pathania D et al. (30)	2019	Point-of-care HPV screening	Training sets: 13000 images Validation: 35 cervical specimens	CNN	Sensitivity: down to a single cell specificity: 100%
Tian R et al. (31)	2019	Predicting cervical lesion grades	10 HPV+ cases 10 CIN1 cases 14 CIN2+cases	Random forest unsupervised clustering	Accuracy 0.814

HPV, human papillomavirus; CIN, cervical intraepithelial neoplasia; SVM, support vector machine; CNN, convolutional neural network.