Radiology. 2026 Mar;318(3):e251586. doi: 10.1148/radiol.251586.
ABSTRACT
Background Supracardiac atherosclerosis is a major cause of stroke. Although vessel wall MRI (vwMRI) can assess plaques, current techniques lack full head-neck-aortic arch coverage and efficiency. Purpose To develop and evaluate a head-neck-aortic vwMRI protocol for supracardiac atherosclerosis, compare its performance with that of supra-aortic CT angiography (CTA), and determine its feasibility in helping identify causes of embolic stroke of undetermined source (ESUS). Materials and Methods We developed a noncontrast head-neck-aortic vwMRI protocol integrated with neural network-based acceleration and multichannel coils, enabling a scan time of approximately 15 minutes. Then, participants with acute ischemic stroke or transient ischemic attack were recruited from October 2024 to March 2025. They received routine supra-aortic CTA and this new vwMRI protocol. Plaque vulnerabilities, vascular stenosis, and stroke etiologies were compared between the modalities. The accuracy of vwMRI for detecting calcification was evaluated using CTA as the reference standard, and the McNemar test was used to compare the findings between CTA and vwMRI. Results Among 108 participants who underwent imaging with both modalities (mean age, 56 years ± 14 [SD]; 89 male participants), the plaque detection rate was higher with vwMRI than CTA (88 of 108 participants [81.5%] vs 75 of 108 participants [69.4%]; P = .002). Moreover, vwMRI helped achieve 91% (98 of 108 participants) accuracy for calcification detection. vwMRI helped identify intraplaque hemorrhage in 27.8% (30 of 108) of participants and demonstrated an ulceration rate comparable with that of CTA (28 of 108 participants [25.9%] vs 25 of 108 participants [23.1%]; P = .51). CTA helped identify 49 participants with vascular stenosis, whereas vwMRI helped identify 47 participants with causes including atherosclerosis (n = 41), thrombus (n = 3), and dissection (n = 3). For 38 participants initially classified with ESUS via CTA, vwMRI enabled reclassification of 16, reducing the percentage of participants with ESUS from 35.2% (38 of 108 participants) to 20.4% (22 of 108 participants) (P < .001). Conclusion This head-neck-aortic vwMRI protocol was feasible and demonstrated a higher plaque detection rate than supra-aortic CTA. © RSNA, 2026 Supplemental material is available for this article. See also the editorial by von Kummer in this issue.
PMID:41842667 | DOI:10.1148/radiol.251586