Quantitative parameters of carotid plaque based on dual-layer spectral-detector CT for stratifying cerebral small vessel disease

Abstract

Cerebral small vessel disease (CSVD) is a major cause of vascular cognitive impairment and stroke, and has been linked to carotid atherosclerotic plaque characteristics. This study aimed to investigate the association between quantitative parameters of non-calcified carotid plaques derived from dual-layer spectral-detector computed tomography (DLSCT) and CSVD burden, and to construct classification models incorporating clinical and imaging variables. This cross-sectional study enrolled 216 patients who underwent carotid DLSCT and brain magnetic resonance imaging. CSVD burden was assessed using a total score (0–4) and categorized as high (score ≥2) or low (score <2) burden. Four logistic regression models were constructed to classify individuals with high CSVD burden, including a clinical model, a model integrating clinical variables with conventional CT features, a model integrating clinical variables with DLSCT-derived parameters, and a model integrating clinical variables, conventional CT features and DLSCT-derived parameters. Model performance was evaluated using optimism-corrected AUC, calibration slope, Brier score, and decision curve analysis. Age and Z-Zeff were identified as independent correlates, with Z-Zeff demonstrating a particularly notable inverse association (OR = 0.386–0.387, p < 0.001). Models incorporating Z-Zeff showed higher discrimination (AUC = 0.748 and 0.759, respectively) compared with the clinical model and the model integrating clinical variables with conventional CT features (AUC = 0.624 and 0.639). Decision curve analysis suggested that models including Z-Zeff had higher net benefit across a range of threshold probabilities. DLSCT-derived Zeff represents an independent quantitative imaging biomarker associated with CSVD burden. Integrating Z-Zeff with clinical variables was associated with improved discrimination for high CSVD burden, supporting further investigation of DLSCT-derived parameters as imaging correlates of CSVD burden.