Research Progress of Artificial Intelligence Image Analysis in Systemic Disease-Related Ophthalmopathy

Table 3

Summary of TAO diagnosis model based on deep learning method.

Study	Task	Sample size	AI model	Output
Lin et al. [29]	Detection	160 MRI images of the eyes	A deep learning system based on a deep convolution neural network	The accuracy, sensitivity, specificity, and scores of network A were , , , and , respectively; the accuracy, sensitivity, specificity, and scores of network B were , , , and , and the AUC values of the two networks were 0.922.
Song et al. [30]	Detection and identification	1135 orbital CT images	A ResNet-18 derived network	The accuracy, sensitivity, and specificity were 0.87, 0.88, and 0.85, respectively, and the AUC value was 0.919.
Salvi et al. [31]	Classification and progression prediction	242 patients’ ophthalmic examination results	A back propagation neural network model with 17 input parameters	The classification accuracy of the model was 78.3%, and the accuracy of predicting progress was 69.2%.