DIAGNOSTIC PERFORMANCE OF AI-QCTISCHEMIA FOR ABNORMAL FFR ACROSS VARIOUS PLAQUE TYPES AND BURDENS
Topic: Hypertension, Atherosclerosis | |
| Type: Presentation - doctors , Number in the programme: 63 | |
| Kamila P.1, Nurmohamed N.2, Danad I.3, Knaapen P.2, Earls J.4, Bax J.1, Van Rosendael A.1 1 Cardiology, LUMC, Leiden, Netherlands, 2 Cardiology, Amsterdam UMC, Amsterdam, Netherlands, 3 Cardiology, Radboud UMC, Nijmegen, Netherlands, 4 , Cleerly Inc, Denver, United States | |
Aim: Quantitative evaluation of coronary plaque and vascular morphology using coronary CT angiography (CCTA), enhanced by artificial intelligence (AI-QCTISCHEMIA), offers reliable estimates of ischemia at the vessel level. This study assesses its diagnostic performance against abnormal invasive fractional flow reserve (FFR), focusing on calcified plaque burden and total plaque burden, indicated by percent atheroma volume (PAV). Methods: Symptomatic patients in the CREDENCE (n=305) and PACIFIC-1 (n=208) studies underwent CCTA, myocardial perfusion imaging (MPI; SPECT/PET), FFR-CT, and invasive coronary angiography with unbiased 3-vessel invasive FFR as reference standard. The performance of non-invasive tests was compared at the vessel level across tertiles of calcified plaque volume, non-calcified plaque volume, and PAV on an intention-to-diagnose basis (uninterpretable results were deemed abnormal). Results: In the CREDENCE study, AI-QCTISCHEMIA achieved AUC ROC values of 0.887, 0.850, and 0.816 to predict ischemia by FFR across calcified plaque tertiles 1 to 3. For non-calcified plaque burden, the AUC ROC values were 0.848, 0.790, and 0.890, while for PAV, the values were 0.863, 0.814, and 0.815, across all tertiles. In the PACIFIC-1 study, AI-QCTISCHEMIA achieved AUC ROC values of 0.811, 0.860, and 0.817 across the calcified plaque tertiles. Conclusion: AI-QCTISCHEMIA shows strong diagnostic performance for invasive FFR across various plaque characteristics and burdens. | |
