Fig. 2: CathAI performance to predict obstructive coronary artery stenosis.
From: CathAI: fully automated coronary angiography interpretation and stenosis estimation
a Receiver Operating Characteristic Curves (ROC) for CathAI prediction of obstructive (</≥70%) coronary stenosis in the test dataset. Black dot: AI-stenosis threshold chosen to optimize F1 score in the artery-level dataset. Red cross: Specificity when sensitivity is fixed at 80%. Blue cross: Sensitivity when specificity is fixed at 80%. b Confusion matrix for CathAI prediction of obstructive stenosis vs. REPORT-stenosis. c UOHI External Validation Dataset-ROC for CathAI prediction of obstructive (</≥70%) stenosis. Black dot: AI-stenosis threshold chosen to optimize F1 score in the artery-level dataset. Red cross: Specificity when sensitivity is fixed at 80%. Blue cross: Sensitivity when specificity is fixed at 80%. d UOHI External Validation Dataset-Confusion matrix for CathAI prediction of obstructive stenosis vs. expert adjudicators. AUC area under the curve, CI confidence interval, AI artificial intelligence.
