Ultra-high resolution photon-counting detector coronary CT minimizes overestimation bias compared to invasive reference

Abstract

Background Photon-counting detector (PCD) coronary CT angiography (CCTA) at ultra-high-resolution (UHR) is a promising tool for the detailed evaluation of the coronary arteries. However, correlation with invasive quantitative coronary angiography (QCA) has not been thoroughly investigated. We here evaluated the efficacy of UHR-CCTA against invasive QCA in patients suspected of coronary artery disease (CAD). Methods Retrospectively, patients suggestive of CAD were included if they had undergone UHR-CCTA on a PCD-CT system showing coronary stenosis which clinically indicated subsequent invasive coronary angiography and no prior coronary interventions. CCTA datasets were reconstructed in 0.6 mm, 0.4 mm, and UHR 0.2 mm slice thicknesses. The extent of stenosis was compared between QCA and CCTA using univariate analysis of variance with post-hoc testing and Bland-Altman plots. Diagnostic performance was assessed based on the detection of relevant coronary stenosis (≥50 %) as confirmed by QCA. Results Forty-nine patients (71 ± 9 years; 37 % male) were included. Stenosis evaluation for 103 segments revealed decreasing mean stenosis diameter with improving spatial resolution (61.4 % for 0.6 mm, 55.3 % for 0.4 mm, 50.9 % for UHR 0.2 mm; p ≤ 0.001). Bias between CCTA and QCA decreased with increasing resolution (13.2 %, limits of agreement [LoA] 30 vs. 9.4 %, 28.1 vs. 5.2 %, 23). UHR-CCTA reconstructions showed superior diagnostic accuracy and positive predictive value (PPV) for detecting relevant CAD compared to lower resolutions (61.2 vs. 61.2 % vs. 71.4 and 53.7 % vs. 53.9 vs. 61.8 %, respectively). Conclusions UHR-CCTA with photon-counting detector CT demonstrated a decrease in overestimation bias and an increase in PPV.