Evaluated whether coronary artery calcium scoring (CACS) can be derived directly from contrast-enhanced CCTA using the PCCT-specific PureCalcium virtual noniodine reconstruction, potentially eliminating the separate true noncontrast (TNC) calcium-score acquisition.
Tested on a first-generation dual-source photon-counting CT in a phantom and 67 patients, comparing PureCalcium with conventional virtual noncontrast (VNC) and TNC reference images.
PureCalcium substantially outperformed conventional VNC. In vivo, correlation with TNC calcium scores was very high (r = 0.97; ICC = 0.98), although PureCalcium still systematically underestimated absolute calcium burden.
Conventional VNC showed much greater underestimation: median CACS 24.3 with VNC vs 232.2 with TNC, compared with 169.3 with PureCalcium vs 232.2 with TNC.
Importantly for clinical risk stratification, calcium-score category agreement was excellent with PureCalcium (κ = 0.88) versus only moderate with VNC (κ = 0.60).
PCCT takeaway: spectral information from a single contrast-enhanced CCTA acquisition can be used to separate iodine from calcium much more effectively than conventional VNC.
This creates the possibility of obtaining both coronary anatomy and clinically useful calcium scoring from one scan, potentially reducing additional acquisitions, scan time, and radiation exposure. Absolute CACS underestimation remains an important limitation.
