Comparison of the Agatston score acquired with photon-counting detector CT and energy-integrating detector CT: ex vivo study of cadaveric hearts

Abstract

The purpose of this study was to compare the correlation and agreement between AS derived from either an energy-integrating detector CT (EID-CT) or a photon-counting detector CT (PCD-CT). Reproducibility was also compared. In total, 26 calcified coronary lesions (from five cadaveric hearts) were identified for inclusion. The hearts were positioned in a chest phantom and scanned in both an EID-CT and a prototype PCD-CT. The EID-CT and PCD-CT acquisition and reconstruction parameters were matched. To evaluate the reproducibility, the phantom was manually repositioned, and an additional scan was performed using both methods. The EID-CT reconstructions were performed using the dedicated calcium score kernel Sa36. The PCD-CT reconstructions were performed with a vendor-recommended kernel (Qr36). Several monoenergetic energy levels (50–150 keV) were evaluated to find the closest match with the EID-CT scans. A semi-automatic evaluation of calcium score was performed on a post-processing multimodality workplace. The best match with Sa36 was PCD-CT Qr36 images, at a monoenergetic level of 72 keV. Statistical analyses showed excellent correlation and agreement. The correlation and agreement with regards to the Agatston score (AS) between the two methods, for each position as well as between the two positions for each method, were assessed with the Spearman´s rank correlation. The correlation coefficient, rho, was 0.98 and 0.97 respectively 0.99 and 0.98. The corresponding agreements were investigated by means of Bland–Altman plots. High correlation and agreement was observed between the AS derived from the EID-CT and a PCD-CT. Both methods also demonstrated excellent reproducibility.

A prospective study comparing the quality of coronary computed tomography angiography images from photon counting and energy integrating detector systems

“…Consequently, it is crucial to authenticate the applicability of this method in relation to the PCD-technology. Although there were only 14 patients with a positive AS, the comparison between EID-CT and PCD-CT showed good correlation and agreement, confirming previous results (39). To explore any over/underestimation of the AS, the PCD-CT data were reconstructed as 70 keV and 80 keV VMIs.…”

“…To explore any over/underestimation of the AS, the PCD-CT data were reconstructed as 70 keV and 80 keV VMIs. VMI 70 keV showed a tendency for overestimation of the AS, whereas VMI 80 keV showed a tendency to underestimate the AS for the PCD-CT (39). Therefore, in terms of the AS, the best match between EID-CT and PCD-CT is likely somewhere between these two keV-levels, although larger studies are needed to validate this result.…”

Coronary Computed Tomography Angiography-Based Calcium Scoring

“…of 2 was about 3%. Skoog et al 37 showed high agreement between PCD- and energy-integrating detector CT–derived CACS using virtual monoenergetic reconstructions at 72 keV in a study of cadaveric hearts.…”

Photon Counting Detector CT-Based Virtual Noniodine Reconstruction Algorithm for In Vitro and In Vivo Coronary Artery Calcium Scoring: Impact of Virtual Monoenergetic and Quantum Iterative Reconstructions

“…19,31 Several ex vivo and in vivo studies already demonstrated the feasibility of accurate CACS with PCD CT using TNC scans. 21,[32][33][34][35][36][37] Emrich et al 20 evaluated this VNI algorithm on PCD CCTA data and showed high accuracy using postprocessing settings at 70 keVand QIR 2. This is in line with the results of the present study that additionally confirms excellent correlation and agreement of CACS VNI with CACS TNC at every investigated VMI level from 55 to 80 keV and every QIR level from strength 1 to 4.…”

“…However, whereas dual-energy-based VNC images are obtained using 2-material separation into soft tissue and iodine, 7,8 PCD enables multimaterial differentiation, including the separation between iodine and calcium 19,31 . Several ex vivo and in vivo studies already demonstrated the feasibility of accurate CACS with PCD CT using TNC scans 21,32–37 . Emrich et al 20 evaluated this VNI algorithm on PCD CCTA data and showed high accuracy using postprocessing settings at 70 keV and QIR 2.…”