PurposeAssess whether photon-counting computed tomography (PCCT) improves discrimination of vulnerable coronary soft-plaque components by extending one-dimensional Hounsfield Unit (HU) thresholding to a simple, interpretable two-dimensional linear rule.ApproachWe generated a synthetic cohort of N=225 coronary plaque phantoms with randomized anatomy, tissue composition (lipid-rich, fibrotic, calcified), and iodine concentrations. Ultra–high-resolution PCCT data were reconstructed into polychromatic T3D, high energy threshold, material-specific, and virtual monoenergetic images (VMIs). Voxel-wise logistic regression implemented single-image (1D) and dual-image (2D) decision rules; performance was assessed by the area under the receiver operating characteristic curve (ROC-AUC). Partial-volume behavior was quantified as correctness versus Euclidean distance to the nearest out-of-class voxel using isotonic regression with a phantom-level bootstrap.ResultsCombining T3D with low-keV VMI yielded the best separation of lipid-rich and fibrous soft-plaque subtypes. A 2D linear rule on T3D + VMI50 achieved AUC=0.925 (95% CI: 0.912, 0.937), exceeding 1D thresholding on T3D (AUC=0.850; 95% CI: 0.821, 0.875) and on VMI50 (AUC=0.814; 95% CI: 0.780, 0.843). Correctness increased with distance to the nearest out-of-class voxel and was ≥95% for voxels at distances D≥0.28 mm (lipid-rich) and D≥0.43 mm (fibrous) (lower 95% CI bounds: 0.20 and 0.41 mm). Accuracy degraded below these thresholds.ConclusionsA transparent, affine 2D threshold that combines routinely reconstructed PCCT images improves voxel-wise discrimination of lipid-rich versus fibrous plaque over conventional HU binning, yielding higher AUCs with tighter 95% confidence intervals. The derived boundary-distance guidance indicates where voxel-level decisions remain reliable, supporting interpretable, clinically pragmatic plaque assessment.
Goldmann et al. (Fri,) studied this question.