Non-invasive imaging of individual histological carotid plaque characteristics: A diagnostic accuracy meta-analysis

Accurately detecting carotid plaque characteristics is crucial for identifying high-risk patients due to risk of cerebrovascular events and complications during revascularizations. Diagnostic accuracy of individual and overall carotid plaque characteristics using computed tomography (CT), magnetic resonance imaging (MRI), and ultrasound (US) compared to histology in patients with symptomatic/asymptomatic carotid plaques was aimed. After prospective registration on PROSPERO (CRD42022329690), Medline Ovid, Embase, Cochrane Library, and Web of Science were searched without any limitations. QUADAS-2 tool was used to study quality assessment, GRADE framework to assess evidence certainty, and univariate/bivariate random-effect meta-analyses for data analysis. Of 5960 studies screened, 107 were identified, resulting in 253 diagnostic accuracy comparisons of 16 plaque characteristics (28 CT, 120 MRI, and 105 US). CT detected intraplaque hemorrhage (IPH) and lipid-rich necrotic core (LRNC) with good accuracy (86 % [95 %CI 67-95] and 84 % [72-91], respectively) and exhibited very high accuracy for ulceration (92 % [87-95]; 76 % on MRI and 75 % on US) and calcification (90 % [58-98] vs. 89 % [87-91] on MRI). MRI identified LRNC and IPH with good accuracy (86 % [81-89] and 86 % [84-88], respectively), and differentiated between acute/subacute/old IPH (accuracy >87 %). US accurately detected ruptured fibrous cap (85 % [77-91]), comparable to MRI (85 % [79-90]), but demonstrated lower performance for other characteristics. Finally, CT detected overall carotid morphology with 89 % accuracy, followed by MRI (86 %; p = 0.374 to CT), and significantly lower by US (78 %; p < 0.001). CT identified key plaque features, especially ulceration and calcification. MRI provided thorough plaque assessment by detecting all features and differentiating IPH age. For overall morphology, CT and MRI surpassed US accuracy.