Evaluation of low-dose chest scans for coronary artery calcium scoring using photon-counting computed tomography with...

Coronary artery calcium scoring (CACS) is a noninvasive method for quantifying coronary artery calcification plaque burden. Although low-dose chest computed tomography (LD-CT) scans with energy-integrating detector CT can be used for CACS, its clinical utility is constrained by higher mean differences and wider limits of agreement. In contrast, photon-counting CT offers significantly enhanced image quality compared to energy-integrating detector CT, making it a more promising tool for accurate CACS. This study investigated the difference in CACS obtained from LD-CT using photon-counting detector CT as compared to standard CACS CT (CAC-CT).

This prospective study included 105 patients (mean age 64.6±11.5 years) who underwent both CAC-CT and LD-CT scans on the same day using photon-counting detector CT. CAC-CT served as a reference standard. Ten groups generated by LD-CT were reconstructed using different strength levels of quantum iterative reconstruction (QIR; 0-4) and slice thickness (3 and 1.5 mm), with the LD_3mm-QIR0 to LD_{1.5 mm-QIR4} groups being designated. The accuracy of CACS detection via LD-CT was evaluated using the intraclass correlation coefficient (ICC; >0.9: excellent) with two-way mixed effects and absolute agreement, Bland-Altman analysis, and weighted kappa (excellent: >0.8) being used for risk categorization agreement. The Wilcoxon signed-rank test was used to compare the radiation doses between CAC-CT and LD-CT. P<0.05 was considered to indicate statistical significance.

Coronary artery calcium was detected in 77 (73.3%) patients via CAC-CT. Each LD-CT group demonstrated high sensitivity (96.1-100%) and specificity (100%) in detecting coronary artery calcium, with all cases detected in the LD_{1.5 mm-QIR0} and LD_{1.5 mm-QIR1} groups. Excellent agreement in CACS was observed between the LD-CT groups and the CAC-CT group (ICC: 0.983-0.993). LD_{1.5 mm-QIR2} showed the narrowest limits of agreement, while LD_{1.5 mm-QIR1} showed the lowest mean bias. The volume CT dose index was reduced by approximately 56.5% with LD-CT as compared to with CAC-CT (1.0 vs. 2.3 mGy; P<0.001).

The LD-CT scan derived from photon-counting detector CT demonstrated excellent agreement with the standard CACS scan in terms of CACS and significantly reduced radiation dose. The reconstruction protocol using a 1.5-mm slice thickness with QIR level 1 improved coronary artery calcium detection, quantification, and risk categorization.