Coronary artery calcium scoring: expanding the new standard by photon-counting detector CT Part II: Impact of virtual monoenergetic image reconstructions with adjusted calcium scoring thresholds

OBJECTIVES: Part I of this study introduced a new photon-counting detector (PCD-)CT protocol for coronary artery calcium (CAC) scoring using 120 kVp, 75% dose, thin slices, quantum iterative reconstructions (IR) 2, leading to a significant reduction of score variability. The second part evaluated the potential of virtual monoenergetic image (VMI) reconstruction in further reducing score variability with PCD-CT.

MATERIALS AND METHODS: CAC scoring was performed on PCD-CT with a chest phantom containing nine calcifications using the optimized PCD-CT protocol from Part I. Images were reconstructed at different VMI levels (50-80 keV, 5 keV-steps), with adjusted CAC thresholds to maintain density equivalence to 70 keV. CAC scores, image noise, and calcification detectability were investigated. Results were compared to standard PCD-CT, EID-CT and previously proposed EID-CT protocols.

RESULTS: Using 65 keV reconstructions, score variability decreased by 9% compared to the optimized PCD-CT protocol from Part I, by 43% vs. the standard PCD-CT, by 78% vs. the standard EID-CT, and by 69% vs. the proposed EID-CT protocol. Image noise remained within targets, eliminating the risk of false-positives. Calcification detectability was comparable to the optimized PCD-CT protocol (7.1 ± 0.6 vs. 7.1 ± 0.8). Calcium volume and mass scores from the keV-optimized PCD protocol were closer to the physical reference compared to scores from the standard PCD protocol.

CONCLUSIONS: Score variability and calcification detectability in PCD-CT-based CAC scoring can be further improved when augmenting an optimized PCD-CT protocol at 65 keV. In addition to reducing the radiation dose, this protocol may enable more consistent CAC quantification and seems to perform even better than the proposed, multivendor EID-CT protocol.

KEY POINTS: Question Coronary calcium scoring lacks reproducibility. Adding virtual monoenergetic imaging with adapted thresholds to a pre-optimized photon-counting CT protocol may further improve score variability. Findings A 120 kVp, 75%-dose, thin-slice photon-counting CT protocol at 65 keV achieved the lowest coronary calcium score variability compared to previous protocols. Clinical relevance Minimizing variability in coronary calcium scoring improves the technical reliability of serial measurements. The additional use of virtual monoenergetic imaging further reduces variability in photon-counting CT, supporting a precise and consistent cardiovascular risk assessment.