3D Convolutional Neural Network for Predicting Clinical Outcome from Coronary Computed Tomography Angiography in Patients with Suspected Coronary Artery Disease
This observational analysis shows improved predictive performance in coronary artery disease patients, indicating the value of combining deep learning with clinical scores.
Key Points
The CNN achieved an AUC of 0.872 for predicting the composite cardiac endpoint in the test cohort.
Integration of Morise score and eoCAD with CNN increased predictive accuracy from 0.652 to 0.920 AUC.
Data for this analysis came from a study involving 5562 patients with suspected coronary artery disease.
The findings suggest that deep learning enhances risk stratification for cardiac events in CAD patients.