Identification of symptomatic carotid atherosclerotic plaques using dual-layer spectral detector computed tomography-derived perivascular adipose tissue parameters and plaque morphology.

Objectives

To establish and validate a multidimensional model for identifying symptomatic carotid plaques based on the combination of plaque morphological characteristics and perivascular adipose tissue (PVAT) parameters obtained from dual-layer spectral detector computed tomography (DLCT).

Methods

This single-center study included 256 patients with carotid atherosclerosis. Patients were divided into symptomatic and asymptomatic groups based on the presence or absence of transient ischemic attack or ipsilateral ischemic stroke within 2 weeks before DLCT. The study comprised a retrospective training cohort (n = 133), a retrospective internal validation cohort (n = 58), and a prospectively enrolled independent temporal test cohort (n = 65). Conventional plaque characteristics and multiple PVAT parameters were compared between the two groups. Imaging variables independently associated with symptomatic carotid plaques were identified using multivariable logistic regression and incorporated into classification models.

Results

Degree of stenosis, plaque burden, and the effective atomic number (Zeff) of PVAT were independently associated with symptomatic carotid plaques. Model 2 incorporated all three parameters demonstrating strong discriminatory performance, with areas under the receiver operating characteristic curves of 0.907, 0.825, and 0.848 in the training, validation, and test sets, respectively. Incorporation of PVAT Zeff improved the discriminative performance and reclassification beyond conventional stenosis and plaque-burden assessment.

Conclusions

Degree of stenosis, plaque burden, and PVAT Zeff were independently associated with symptomatic carotid plaques. Adding PVAT Zeff to conventional plaque characteristics improved model discrimination and provided complementary information for the characterization of symptomatic lesions.