Spectral Black-Blood Photon-Counting CT for High-Resolution Carotid Vessel Wall Imaging

OBJECTIVES: To evaluate the feasibility of a novel single-phase spectral black-blood CT (BBCT) technique for high-resolution carotid vessel wall imaging and explore its potential impact on plaque feature visualization. METHODS: In this IRB-approved retrospective study, clinically indicated head and neck photon-counting detector (PCD)-CT angiography (CTA) examinations were used to generate high-resolution BBCT reconstructions from a single angiographic-phase data set using spectral iodine suppression. Nine iodine ratio (IR) settings were evaluated to determine the balance between intraluminal suppression and vessel wall preservation. Objective image quality was assessed using attenuation and contrast-to-noise ratio (CNR). Three readers independently assessed overall image quality, lumen visualization, wall-lumen delineation, plaque conspicuity, and high-risk plaque features, including intraplaque hemorrhage (IPH), lipid-rich necrotic core (LRNC), and thin fibrous cap, and compared BBCT with conventional CTA reconstructions (0.4 mm, Qr48 kernel). Inter-reader agreement was assessed using Gwet's agreement coefficient (AC). RESULTS: A total of 40 patients (mean age: 64.1±11.3 y; 57.5% women) were included. Lumen attenuation progressively decreased with lower IR values (range: 51.0 to -176.6 HU), whereas vessel wall attenuation remained relatively preserved (range: 11.4 to 55.8 HU, P=0.004). Lumen-to-wall CNR decreased from 13.6±6.1 to 2.1±1.7, whereas wall-to-fat CNR increased from 7.2±3.8 to 15.3±6.8, with IR 2.0 representing the best balance between lumen suppression and wall conspicuity. Compared with conventional CTA, BBCT improved wall-lumen delineation across readers (median score: 2 to 3 vs. 4; all P<0.001) and plaque conspicuity in evaluable plaques (median score: 3 vs. 4; all P≤0.03), whereas overall image quality was preserved (all P≥0.45). Agreement improved for IPH (AC: 0.63 vs. 0.45), LRNC (0.62 vs. 0.50), and thin fibrous cap assessment (0.70 vs. 0.33). However, the number of cases classified as positive for IPH varied across readers and was lower on BBCT (1 to 9 cases) than on CTA (11 to 14 cases), suggesting possible subtraction-related sensitivity loss. CONCLUSIONS: Spectral BBCT was feasible for high-resolution carotid vessel wall imaging, with preserved overall image quality, improved wall-lumen delineation, and higher plaque conspicuity compared with conventional CTA. Improved agreement for selected plaque features supports its potential as a complementary vessel wall reconstruction; however, possible subtraction-related sensitivity loss suggests BBCT should be interpreted in conjunction with conventional CTA rather than as a standalone reconstruction.