Abstract WP143: Differentiating Carotid Artery Webs From Large Artery Atherosclerosis Using Hounsfield Units and Clot Perviousness

Le résumé fourni par la source

Background: Carotid artery webs (CWs) are an increasingly recognized cause of ischemic stroke. Differentiating CW from soft plaque due to large artery atherosclerosis (LAA) in the proximal internal carotid artery (ICA) can be challenging but is clinically important, as the two entities carry different recurrence risks. This study evaluated whether CWs can be distinguished from LAA lesions using Hounsfield unit (HU) measurements and clot perviousness. Methods: We retrospectively analyzed patients with ipsilateral CW and middle cerebral artery (M1) occlusion from an ongoing registry at a Comprehensive Stroke Center. A comparison group of patients with extracranial carotid severe ipsilateral LAA was selected by age matching (±2 years) to the CW group. HU measurements were obtained at the thrombus site on non-contrast head CT (HCT); for cases without a hyperdense MCA sign, CT angiography (CTA) was co-registered for localization. Relative HU (rHU) was used instead of absolute HU to account for inter-scan variability in image acquisition and patient-specific factors such as hematocrit, which can influence attenuation values, thereby improving comparability across patients. The optimal rHU cut point for differentiating CW from LAA was determined using Youden’s method. Clot perviousness was calculated as the HU difference between CTA and HCT, with >10 HU considered pervious. Logistic regression assessed associations with thrombus attenuation index (TAI). Results: Forty-nine patients were included (27 LAA, 22 CW), mean age 56 ± 11 years; 49% were male. CW patients were more often Black (23% vs. 0%, p = 0.038), while LAA patients had higher rates of congestive heart failure (26% vs. 5%, p = 0.044). The median TAI was higher in CW compared with LAA (36.3 HU [IQR 23–46] vs. 21.5 HU [IQR 2.5–35.4], p=0.028)The optimal rHU threshold for distinguishing CW from LAA was 1.13 (sensitivity 86%, AUC 0.60). In the unadjusted analysis, CW clots were more often pervious (TAI ≥10 HU) compared with LAA (OR 4.35; 95% CI, 1.03–18.37; p = 0.045). This association remained significant after adjustment for confounders (adjusted OR 5.50; 95% CI, 1.20–25.22; p = 0.028). Conclusions: CWs can be differentiated from LAA lesions using HU characteristics and clot perviousness, with CW clots showing higher attenuation on CTA and greater perviousness. These imaging markers may improve diagnostic accuracy and guide tailored stroke prevention strategies.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.