The size of juxtaluminal hypoechoic area in ultrasound images of asymptomatic carotid plaques predicts the occurrence of stroke

Abstract

Objective: To test the hypothesis that the size of a juxtaluminal black (hypoechoic) area (JBA) in ultrasound images of asymptomatic carotid artery plaques predicts future ipsilateral ischemic stroke.

Methods: A JBA was defined as an area of pixels with a grayscale value <25 adjacent to the lumen without a visible echogenic cap after image normalization. The size of a JBA was measured in the carotid plaque images of 1121 patients with asymptomatic carotid stenosis 50% to 99% in relation to the bulb (Asymptomatic Carotid Stenosis and Risk of Stroke study); the patients were followed for up to 8 years.

Results: The JBA had a linear association with future stroke rate. The area under the receiver-operating characteristic curve was 0.816. Using Kaplan-Meier curves, the mean annual stroke rate was 0.4% in 706 patients with a JBA <4 mm(2), 1.4% in 171 patients with a JBA 4 to 8 mm(2), 3.2% in 46 patients with a JBA 8 to 10 mm(2), and 5% in 198 patients with a JBA >10 mm(2) (P < .001). In a Cox model with ipsilateral ischemic events (amaurosis fugax, transient ischemic attack [TIA], or stroke) as the dependent variable, the JBA (<4 mm(2), 4-8 mm(2), >8 mm(2)) was still significant after adjusting for other plaque features known to be associated with increased risk, including stenosis, grayscale median, presence of discrete white areas without acoustic shadowing indicating neovascularization, plaque area, and history of contralateral TIA or stroke. Plaque area and grayscale median were not significant. Using the significant variables (stenosis, discrete white areas without acoustic shadowing, JBA, and history of contralateral TIA or stroke), this model predicted the annual risk of stroke for each patient (range, 0.1%-10.0%). The average annual stroke risk was <1% in 734 patients, 1% to 1.9% in 94 patients, 2% to 3.9% in 134 patients, 4% to 5.9% in 125 patients, and 6% to 10% in 34 patients.

Conclusions: The size of a JBA is linearly related to the risk of stroke and can be used in risk stratification models. These findings need to be confirmed in future prospective studies or in the medical arm of randomized controlled studies in the presence of optimal medical therapy. In the meantime, the JBA may be used to select asymptomatic patients at high stroke risk for carotid endarterectomy and spare patients at low risk from an unnecessary operation.