Fig 1.
The 3 year and 8 year stroke rate of patients in the 5 risk categories in the testing datasets.
Table 1.
Characteristics of development and testing datasets.
Fig 2.
Performance of the deep learning model for predicting 3 year stroke occurrence in (A) testing dataset 1 and (B) testing dataset 2.
Fig 3.
Sensitivity and specificity of the DNN model for predicting 3 year stroke occurrence in different testing time periods under (A) the high specificity operating point and (B) the high sensitivity operating point.
Table 2.
Performance of currently available stroke risk assessment scores and the deep learning model.