Description:
Introduction: Coronary atherosclerosis and vascular morphology imaging by quantitative CT angiography (AI-QCT) provides detailed plaque and vessel measures of the coronary tree. These features combined in a tool (AI-QCT ISCHEMIA ) accurately estimate vessel level ischemia. We investigated the diagnostic performance of AI-QCT ISCHEMIA for abnormal invasive fractional flow reserve (FFR) across subgroups of calcified plaque. Methods: Symptomatic patients within the CREDENCE (N=305) and PACIFIC (N=208) studies underwent CCTA, myocardial perfusion imaging (MPI; single photon emission computed tomography [SPECT] and/or positron emission tomography [PET]), fractional flow reserve by CT (FFR-CT ) and invasive coronary angiography in conjunction with unbiased 3-vessel invasive FFR measurements as reference standard. The diagnostic performance of the non-invasive tests was compared on vessel-level across the tertiles of calcified plaque volume, on an intention-to-diagnose basis. Results: Within the CREDENCE study, AI-QCT ISCHEMIA had an area under the receiver operating characteristics curve (AUC ROC) of 0.887, 0.850, 0.816 across tertile 1-3 of calcified plaque, respectively, which was significantly higher than FFR-CT and MPI. Within the PACIFIC study, the AUC ROC was 0.811, 0.860, 0.817 across the calcified plaque tertiles 1-3, which was significantly better than SPECT, and comparable to FFR-CT ad PET. The specificity of AI-QCT ISCHEMIA was lower with increased calcified plaque, being approximately 70% for the most calcified vessels amongst both cohorts. Conclusion AI-QCT ISCHEMIA , a tool to assess coronary ischemia based on quantitative coronary plaque and vascular morphology retained good diagnostic performance for invasive FFR across calcified plaque subgroups.