Diagnostic Impact of Intracranial Vessel Wall MRI on Ischemic Stroke Etiology in Young Adults Authors: Juhan Reimand, Siim Schneider, Riina Vibo, Liisa Kõrv, Dagmar Loorits, Karl Jõgi, Teele Tuularu, Nele Taba, Kateriine Orav, Janika Kõrv Published: 2026 ABSTRACT ABSTRACTBackgroundIntracranial vessel wall imaging (VWI) enables visualization of arterial wall pathology and may improve etiologic classification of ischemic stroke, but existing evidence is largely based on retrospective studies. We aimed to prospectively evaluate the diagnostic impact of VWI in young stroke patients.MethodsIn this two‐centre study, consecutive patients aged 18–59 years with acute ischemic stroke in the middle cerebral or anterior choroidal artery territory underwent contrast‐enhanced VWI within 2 weeks of stroke onset. Stroke etiology was assessed using the Trial of ORG 10172 in Acute Stroke Treatment (TOAST) and the Causative Classification of Stroke (CCS) classifications before and after incorporating VWI findings. Changes in etiologic classification and predictors of diagnostic change were analyzed.ResultsSixty‐nine patients were included (median age 53 years, 72% male). Incorporation of VWI findings resulted in a change in etiologic classification in 9/69 (13%) patients according to TOAST and 10/69 (14%) according to CCS. For TOAST classification, the proportion of large artery atherosclerosis increased from 17% to 28%, while cryptogenic stroke decreased from 22% to 14%. Baseline cryptogenic stroke was associated with etiologic reclassification after adjustment for age and sex (OR 7.8, 95% CI: 1.2–49.7; p = 0.029).ConclusionsVWI frequently identified intracranial atherosclerotic plaques not evident on luminal imaging, leading to etiologic reclassification predominantly toward large artery atherosclerosis. These findings suggest incremental diagnostic value in the etiologic assessment of young stroke, particularly in cryptogenic cases. Larger prospective studies are warranted to determine its role in routine diagnostic algorithms.Keywords: atherosclerosis, blood vessel, cryptogenic stroke, ischemic stroke, young adult Introduction Stroke represents a clinically heterogeneous condition with a broad spectrum of underlying etiologies [1]. Ischemic stroke remains a leading cause of mortality and long‐term disability worldwide [2]. Accurate determination of the ischemic stroke cause is essential, as it guides secondary prevention and influences prognosis [3]. Despite a comprehensive diagnostic evaluation, approximately one in four ischemic strokes remains classified as cryptogenic [4]. This is more common in young adults, in whom the etiologic spectrum is broader and differs from that of older populations [1, 5]. These observations reflect limitations of current diagnostic approaches and are particularly relevant given the substantial burden of stroke and the conflicting epidemiologic data regarding temporal incidence trends in this population [5, 6]. The diagnostic workup of ischemic stroke involves a multimodal approach aimed at identifying the underlying etiology [3, 5]. Neurovascular imaging modalities, including computed tomography angiography (CTA) and magnetic resonance angiography (MRA), play a central role in stroke evaluation. These techniques provide detailed information on the vessel lumen but are inherently limited in their ability to visualize the vessel wall and detect non‐stenotic or early pathological changes [7]. Intracranial vessel wall imaging (VWI) enables direct assessment of the arterial wall and characterization of disease‐specific features [8]. Prior studies have demonstrated that VWI can lead to substantial changes in the etiologic classification of stroke [9, 10]. However, most evidence is derived from retrospective analyses and selected patient cohorts, potentially introducing selection bias. Despite previous promising results, VWI has not been included in the routine diagnostic evaluation of young stroke [5]. Additional prospective data may provide a more accurate estimate of the diagnostic impact of VWI in routine clinical practice. Accordingly, in this two‐centre study, we aimed to prospectively evaluate the impact of VWI on stroke etiologic classification in young patients and to identify patient characteristics associated with additional diagnostic value. Method Study Population All consecutive patients aged 18 to 59 years with acute ischemic stroke, defined by the presence of a diffusion‐weighted imaging (DWI) lesion on brain magnetic resonance imaging (MRI) in the territory of the middle cerebral artery (MCA) or anterior choroidal artery (AChA), were prospectively included, irrespective of recanalization therapy. Patients were enrolled at two tertiary stroke centres, North Estonia Medical Centre (NEMC) and Tartu University Hospital (TUH), between September 2023 and February 2026. Recruitment at TUH was suspended from May to November 2025 due to MRI scanner replacement. All patients underwent a comprehensive etiologic workup, including brain MRI, imaging of cervical and intracranial arteries by CTA and/or MRA, transesophageal and transthoracic echocardiography, 24‐h Holter electrocardiography (ECG), routine blood tests (including total cholesterol, high‐density lipoprotein cholesterol, low‐density lipoprotein cholesterol, triglycerides, and C‐reactive protein [CRP]), and screening for prothrombotic conditions [11]. Patients were excluded if they met any of the following criteria: (1) inability to undergo MRI due to factors such as agitation, claustrophobia, morbid obesity, medical instability, or non‐MR‐compatible implants or devices; (2) contraindication to gadolinium‐based contrast agents; (3) inability or refusal to provide informed consent; (4) failure to complete the VWI protocol within the predefined 2‐week time window after stroke onset. This study is a substudy of the Estonian Young Stroke Registry, which is a prospective ongoing hospital‐based registry of all consecutive young stroke patients (aged 18–54 years) hospitalized at Tartu University Hospital and North Estonia Medical Centre, as described previously [11]. For the present analysis, the age range was extended to 18–59 years to increase the number of eligible patients and ensure feasible recruitment. This cutoff remained within age ranges previously used in studies of young stroke [12]. The study was approved by the Research Ethics Committee of the University of Tartu (approval no. 384/M‐21). Written informed consent was obtained from all participants prior to enrolment. MRI Protocol As part of routine etiologic workup, all patients underwent conventional post‐stroke MRI sequences, including DWI, fluid attenuated inversion recovery (FLAIR), and susceptibility‐weighted imaging (SWI). In addition, all patients underwent a high‐resolution contrast‐enhanced VWI examination within 2 weeks of stroke onset (median 6 days; interquartile range [IQR]: 5–8 days). In both centres, the VWI protocol consisted of a localizer sequence, three‐dimensional (3D) time‐of‐flight angiography (TOF), pre‐contrast 3D T1‐weighted black‐blood imaging, contrast‐enhanced MR angiography (CEMRA), T2‐weighted and 3D post‐contrast T1‐weighted black‐blood (BB) imaging, as previously recommended [7, 13]. T1‐weighted BB sequences employed vendor‐specific flow‐suppression techniques to reduce signal from flowing blood and cerebrospinal fluid, thereby enhancing vessel wall visualization [7]. Axial high‐resolution T1‐weighted BB and T2‐weighted imaging slabs were positioned to cover the circle of Willis and distal intracranial internal carotid arteries. Post‐contrast T1‐weighted BB imaging was performed after intravenous administration of 0.1 mmol/kg gadobutrol (Gadovist, Bayer AG). T2‐weighted imaging was acquired after contrast administration and CEMRA to allow adequate contrast agent distribution within the vessel wall prior to post‐contrast T1‐weighted BB imaging. Depending on the MRI system, total scan time was approximately 35–45 min, of which the T1‐weighted BB pre‐ and post‐contrast sequences each accounted for approximately 5–8 min. In NEMC, VWI was performed on two Signa Premier 3 T scanners (GE Healthcare, Milwaukee, USA) using 48‐channel head phased‐array coils and identical protocols (n = 37 + 4). Pre‐ and post‐contrast axial T1‐weighted motion‐sensitized driven equilibrium (MSDE) sequences had an in‐plane reconstructed resolution (RR) of 0.47 × 0.47 mm with a slice thickness (ST) of 0.5 mm. In addition, axial T2‐weighted sequences had an RR of 0.18 × 0.18 mm, ST 1.0 mm; axial 3D TOF angiography had an RR of 0.20 × 0.20 mm, ST 1.0 mm; and coronal CEMRA had an RR of 0.26 × 0.26 mm, ST 0.89 mm. In TUH, VWI was performed on two different scanners. The majority of patients (n = 23) underwent VWI on a Philips Achieva 3 T (Philips Healthcare, Best, The Netherlands) scanner using a 16‐channel dedicated neurovascular phased‐array coil. 3D pre‐ and post‐contrast T1‐weighted BB sequences had an RR of 0.34 × 0.34 mm, ST 1.3 mm. Axial T2‐weighted sequence had an RR of 0.47 × 0.47 mm, ST 1.0 mm; axial 3D TOF angiography had an RR of 0.35 × 0.35 mm, ST 1.0 mm; and coronal CEMRA had an RR of 0.50 × 0.50 mm, ST 0.98 mm. A subset of patients in TUH (n = 5) underwent VWI on a Siemens MAGNETOM Vida (Siemens Healthineers, Erlangen, Germany) using a 20‐channel head/neck receive coil. Sagittal 3D pre‐ and post‐contrast T1‐weighted BB sequences had an RR of 0.72 × 0.72 mm, ST 0.70 mm. Axial T2‐weighted sequence had an RR of 0.30 × 0.30 mm, ST 1.0 mm; 3D TOF angiography had an RR of 0.42 × 0.42 mm, ST 0.45 mm; and coronal CEMRA had an RR of 0.63 × 0.63 mm, ST 0.9 mm. Acquired resolutions are provided in Table S1. Etiologic Classification VWI scans were independently assessed by two neuroradiologists (J.R. and D.L., each with > 10 years of experience in neuroradiology) to evaluate stroke etiology based on infarct pattern, luminal findings, vessel wall morphology (including eccentric or concentric wall thickening and vascular remodeling), and signal characteristics such as contrast enhancement. Both readers were blinded to clinical information, except for prior imaging findings. Diagnostic confidence was classified by study neuroradiologists as high or low. High confidence indicated that a specific radiologic diagnosis was favored without substantial diagnostic uncertainty, whereas low confidence indicated an uncertain interpretation or relevant alternative diagnoses. CTA scans (available in 67/69, 97%) were re‐evaluated by study neuroradiologists prior to VWI assessment to ensure comprehensive imaging assessment for subsequent etiologic classification. The etiology of ischemic stroke was independently evaluated by two stroke neurologists (S.S., K.O.) using the Trial of ORG 10172 in Acute Stroke Treatment [14] (TOAST) and the Causative Classification of Stroke [15] (CCS) based on all available clinical, laboratory, and imaging data. The initial classification was performed without access to VWI results and was subsequently repeated after incorporating VWI findings. Discrepancies in both radiologic and etiologic classification were resolved by consensus. Statistical Analysis Statistical analysis was performed using R software (version 4.5.1). A two‐sided p‐value < 0.05 was considered statistically significant. Group comparisons for continuous variables were performed using the Wilcoxon rank‐sum test and categorical variables were compared using the Fisher's exact test. Effect sizes were reported as Hodges–Lehmann estimates with corresponding 95% confidence intervals (CI) for continuous variables and as odds ratios (OR) with corresponding 95% CI for categorical variables. Interobserver agreement between the two neuroradiologists for the radiologic diagnosis and between the two neurologists for stroke etiologic classification was evaluated using Cohen's kappa (κ). The impact of VWI on TOAST and CCS etiologic classification was assessed by comparing the proportion of diagnoses before and after VWI. Then factors associated with TOAST reclassification after VWI were investigated. First, logistic regression analyses were performed to evaluate whether baseline TOAST category was associated with TOAST reclassification after VWI both without adjustment and after adjusting for age and sex, with cardioembolism as the reference category. Second, as an exploratory analysis, clinical characteristics, vascular risk factors, blood biomarkers, and the presence of ≥ 50% stenosis on CTA [16] were compared between patients with and without TOAST classification change. The selected variables were chosen based on prior evidence suggesting relevance to stroke etiology and atherosclerotic activity. Exploratory analyses were not adjusted for multiple comparisons and should be interpreted as hypothesis‐generating. As sensitivity analyses, the proportion of TOAST reclassifications was also assessed in: (i) patients aged ≤ 50 years; and (ii) after excluding patients who underwent mechanical thrombectomy. Results Study Population A total of 93 patients with DWI lesions in the MCA or AChA territory were screened for inclusion. Of these, 10 declined to provide informed consent, 7 were excluded because VWI could not be performed within the predefined time window, 4 due to critical condition and/or severe neurological deficit, and 3 due to incomplete VWI or excessive motion artifacts. The final study cohort comprised 69 patients (41 from NEMC, 28 from TUH). The median age was 53 years (IQR 47–55), 50 (72%) were male and the median baseline National Institutes of Health Stroke Scale (NIHSS) score was 3 (IQR 1–6). Vascular risk factors were prevalent, including current smoking in 35/69 (51%), hypertension in 35/69 (51%), and dyslipidemia in 54/69 (78%) patients, with a median body mass index (BMI) of 28 kg/m2 (IQR 24–32). Intravenous thrombolysis was administered in 23/69 patients (33%) and 12/69 (17%) underwent mechanical thrombectomy. Baseline characteristics of the study population are summarized in Table 1. TABLE 1Participant clinical characteristics by study centre.Total n = 69 (100%)NEMC n = 41 (59%)TUH n = 28 (41%)Sex, male (%)50 (72%)29 (71%)21 (75%)Age (median [IQR])53 (47–55)52 (48–53)54 (44–57)NIHSS, baseline (median [IQR])3 (1–6)2 (1–6)4 (1–7)mRS, premorbid (median [IQR])0 (0–0)0 (0–0)0 (0–1)**mRS, 7‐day (median [IQR])1 (0–2)1 (0–2)1 (0–2)IVT (%)23 (33%)9 (22%)14 (50%)*MT (%)12 (17%)7 (17%)5 (18%)Time from stroke onset to VWI, days (median [IQR])6 (5–8)7 (5–10)6 (5–7)Stroke risk factorsBMI (median [IQR])28.1 (24.4–32.0)28.3 (25.8–32.0)28.0 (22.3–31.6)Atrial fibrillation (%)5 (7%)4 (10%)1 (4%)Current smoking (%)35 (51%)18 (44%)17 (61%)Hypertension (%)35 (51%)24 (59%)11 (39%)Dyslipidemia (%)54 (78%)33 (80%)21 (75%)Diabetes (%)9 (13%)5 (12%)4 (14%)PFO (%)20 (29%)13 (32%)7 (26%)Previous TIA and/or stroke (%)6 (9%)6 (15%)0 (0%) Note: Data are presented as n (%) or median (IQR). Categorical variables were analyzed using Fisher's exact test and continuous variables using the Wilcoxon rank‐sum test. IQR is reported as the 1st and 3rd quartiles. *p < 0.05, **p < 0.01 for comparisons between study centres.Abbreviations: BMI, body mass index; IQR, interquartile range; IVT, intravenous thrombolysis; mRS, Modified Rankin Scale; MT, mechanical thrombectomy; NEMC, North Estonia Medical Centre; NIHSS, National Institutes of Health Stroke Scale; PFO, patent foramen ovale; TIA, transient ischemic attack; TUH, Tartu University Hospital; VWI, vessel wall imaging. Participant clinical characteristics by study centre. Note: Data are presented as n (%) or median (IQR). Categorical variables were analyzed using Fisher's exact test and continuous variables using the Wilcoxon rank‐sum test. IQR is reported as the 1st and 3rd quartiles. *p < 0.05, **p < 0.01 for comparisons between study centres. Abbreviations: BMI, body mass index; IQR, interquartile range; IVT, intravenous thrombolysis; mRS, Modified Rankin Scale; MT, mechanical thrombectomy; NEMC, North Estonia Medical Centre; NIHSS, National Institutes of Health Stroke Scale; PFO, patent foramen ovale; TIA, transient ischemic attack; TUH, Tartu University Hospital; VWI, vessel wall imaging. Baseline Etiologic Classification Following independent assessment and consensus review by both neurologists, the baseline TOAST etiologies were cardioembolism (23/69, 33%), stroke of undetermined etiology (cryptogenic stroke; 15/69, 22%), large artery atherosclerosis (LAA; 12/69, 17%), small vessel occlusion (SVO; 12/69, 17%), and other determined etiology (7/69, 10%). Interobserver agreement for baseline TOAST classification was substantial (κ = 0.79; 95% CI: 0.68–0.90, p < 0.001). Vessel wall abnormalities consistent with a stroke etiology were identified in 21/69 (30%) cases, most frequently intracranial atherosclerotic disease (ICAD; 14/69, 20%). Interobserver agreement for VWI interpretation between radiologists was substantial (κ = 0.74; 95% CI: 0.59–0.88; p < 0.001) and radiologic diagnostic confidence was considered high in 58/69 (84%) cases. Among patients who underwent thrombectomy, vessel wall changes attributable to the procedure were observed in 7/12 (58%). Impact of VWI on Etiologic Classification Following consensus review, VWI led to a change in TOAST classification in 9/69 patients (13%), of whom five were reclassified from cryptogenic stroke to LAA, one from cryptogenic stroke to other determined etiology, one from cardioembolism to LAA, one from cardioembolism to cryptogenic stroke, and one from SVO to LAA. Sensitivity analyses showed similar reclassification rates in patients aged ≤ 50 years (3/29 reclassified, 10%) and after excluding patients who underwent mechanical thrombectomy (9/57 reclassified, 16%). Changes in consensus TOAST etiologic classification after VWI are illustrated in Figure 1A, while independent and consensus classifications are shown in Table S2. The proportion of LAA increased from 17% to 28% (12 to 19), while the proportion of cryptogenic stroke decreased from 22% to 14% (15 to 10). Among patients with baseline cryptogenic stroke, reclassification occurred in 6/15 (40%) cases, whereas it was uncommon in patients with baseline cardioembolism (2/23, 9%), SVO (1/12, 8%), and was not observed in patients with LAA (0/12) and other determined etiology (0/7). Following incorporation of VWI findings, interobserver agreement for TOAST classification increased to κ = 0.83 (95% CI: 0.73–0.93; p < 0.001). Similar changes were observed in CCS, with reclassification occurring in 10/69 (14%) patients (Figure 1B, Table S3). Representative MRI findings are shown in Figure 2. Impact of VWI on TOAST and CCS etiologic classification. Sankey diagram illustrating the impact of VWI on etiologic classification according to TOAST (A) and CCS (B). The left side represents baseline etiologic categories and the right side shows post‐VWI classifications with connecting flows indicating transitions between categories. Numbers within nodes indicate the number of patients in each category. Two nodes containing a single patient do not display a patient count. The width of each flow is proportional to the number of patients. TOAST transitions were cryptogenic→LAA (n = 5), cryptogenic→other (n = 1), CE → LAA (n = 1), CE → cryptogenic (n = 1), and SVO → LAA (n = 1). CCS transitions were possible CE → evident LAA (n = 4), other cryptogenic→evident LAA (n = 2), possible LAA → evident LAA (n = 1), evident CE → probable LAA (n = 1), evident SVO → evident LAA (n = 1), and other cryptogenic→evident other (n = 1). Categories not shown did not occur in the study cohort. CCS, Causative Classification of Stroke; CE, cardioembolism; LAA, large artery atherosclerosis; SVO, small vessel occlusion; TOAST, Trial of ORG 10172 in Acute Stroke Treatment; VWI, vessel wall imaging. Three representative cases illustrating changes in TOAST etiologic classification following VWI. Case 1 (A–E). A 51‐year‐old female presented with a left MCA stroke. DWI (A) shows an acute ischemic lesion in the left basal ganglia region. No obvious stenosis is seen on CTA (B). VWI demonstrates eccentric wall thickening in the anterior and superior wall of the M1 segment of the MCA with intermediate signal on non‐contrast axial MSDE (C), heterogeneous signal on axial T2 (D), and intense enhancement on post‐contrast axial MSDE (E). TOAST classification changed from SVO to large artery atherosclerosis. Case 2 (F–J). A 59‐year‐old male with a left MCA stroke. DWI demonstrates an acute ischemic lesion in the left temporoparietal region (F). Atherosclerotic changes with both calcified and non‐calcified components are seen on CTA (G). VWI shows concentric wall thickening in the distal internal carotid artery with intermediate signal on non‐contrast axial MSDE (H), and intense enhancement on post‐contrast coronal (I) and axial (J) MSDE. TOAST classification changed from cryptogenic stroke to large artery atherosclerosis. Case 3 (K–O). A 57‐year‐old male with a left MCA stroke. DWI reveals multiple acute ischemic lesions in the left cerebral hemisphere (K). CTA (not shown) and TOF (L) do not demonstrate significant atherosclerotic changes or stenosis. VWI reveals bilateral subtle M1 concentric vessel wall thickening with intermediate signal on axial T1 BB (M). Axial post‐contrast T1 BB with motion artifacts demonstrates intense concentric enhancement in a left M2 branch (N, O) and bilateral M1 segments (O). TOAST classification changed from cryptogenic stroke to other determined etiology (vasculitis). BB, black‐blood; CTA, computed tomography angiography; DWI, diffusion‐weighted imaging; MCA, middle cerebral artery; MSDE, motion‐sensitized driven equilibrium; SVO, small vessel occlusion; TOAST, Trial of ORG 10172 in Acute Stroke Treatment; TOF, time‐of‐flight angiography; VWI, vessel wall imaging. Predictors of TOAST Reclassification After VWI In logistic regression analysis, baseline cryptogenic stroke was associated with TOAST reclassification after VWI without adjustment (OR 7.0; 95% CI: 1.2–41.5; p = 0.032) and after adjustment for age and sex (OR 7.8, 95% CI: 1.2–49.7; p = 0.029; Table S4), with wide CI observed in both models. In an exploratory analysis, we assessed the association of selected clinical markers with TOAST reclassification (Table 2). CRP showed a nominal association with TOAST classification change (uncorrected p = 0.029, Hodges–Lehmann estimate: 2.00; 95% CI: 0.38–3.65). TABLE 2Exploratory comparison of characteristics according to TOAST reclassification after VWI.Total n = 69TOAST not changed n = 60TOAST changed n = 9Effect (95% CI) p Sex, male (%)50 (72%)43 (72%)7 (78%)OR: 1.38 (0.23; 14.90)1.000Age (median [IQR])53 (47–55)53 (47–55)56 (49–57)HL: 3.00 (−2.00; 7.00)0.208Baseline NIHSS (median [IQR])3.0 (1.0–6.0)3.0 (1.0–6.5)1.0 (0.0–4.0)HL: −1.00 (−3.00; 1.00)0.350Thrombolysis (%)23 (33%)20 (33%)3 (33%)OR: 1.00 (0.15; 5.29)1.000Thrombectomy (%)12 (17%)12 (20%)0 (0%)OR: 0.00 (0.00; 2.38)0.342CTA ≥ 50% stenosis (%)15 (22%)12 (21%)3 (33%)OR: 1.90 (0.27; 10.58)0.407Centre, NEMC (%)41 (59%)37 (62%)4 (44%)OR: 0.50 (0.09; 2.60)0.469Stroke risk factorsAtherosclerosis composite score (median [IQR])2.0 (1.0–4.0)2.0 (1.0–4.0)3.0 (3.0–4.0)HL: 1.00 (0.00; 2.00)0.106BMI ≥ 30 (%)22 (33%)17 (29%)5 (56%)OR: 2.96 (0.56; 16.87)0.142Current smoking (%)35 (51%)29 (48%)6 (67%)OR: 2.12 (0.41; 14.28)0.477Hypertension (%)35 (51%)28 (47%)7 (78%)OR: 3.93 (0.67; 41.80)0.151Diabetes (%)9 (13%)7 (12%)2 (22%)OR: 2.13 (0.18; 14.83)0.333Dyslipidemia (%)54 (78%)46 (77%)8 (89%)OR: 2.41 (0.28; 115.59)0.672Coronary artery disease (%)5 (7%)5 (8%)0 (0%)OR: 0.00 (0.00; 7.82)1.000Cardiac arrhythmia (%)5 (7%)5 (8%)0 (0%)OR: 0.00 (0.00; 7.82)1.000Heart failure (%)7 (10%)7 (12%)0 (0%)OR: 0.00 (0.00; 4.92)0.582Previous TIA (%)5 (7%)5 (8%)0 (0%)OR: 0.00 (0.00; 7.82)1.000PFO (%)20 (29%)17 (29%)3 (33%)OR: 1.23 (0.18; 6.60)1.000Blood biomarkersTotal cholesterol, mmol/L (median [IQR])4.90 (4.10–5.90)4.90 (4.05–5.90)4.90 (4.20–5.90)HL: 0.00 (−1.00; 0.90)0.922LDL, mmol/L (median [IQR])3.30 (2.52–4.09)3.30 (2.53–4.07)3.58 (2.44–4.23)HL: 0.08 (−0.86; 0.98)0.936HDL, mmol/L (median [IQR])1.10 (0.90–1.40)1.13 (0.90–1.41)1.02 (0.98–1.12)HL: −0.04 (−0.25; 0.13)0.682Triglycerides, mmol/L (median [IQR])1.24 (1.00–1.66)1.24 (1.00–1.77)1.23 (1.02–1.49)HL: −0.01 (−0.35; 0.33)0.949CRP, mg/L (median [IQR])1.60 (0.98–3.00)1.50 (0.70–2.00)3.20 (2.46–7.00)HL: 2.00 (0.38; 3.65)0.029 Note: CTA ≥ 50% stenosis was assessed in the intracranial internal carotid artery and middle cerebral artery M1–M2 segments; occlusions were not included. The atherosclerosis composite score was calculated as the sum of selected risk factors (BMI ≥ 30, current smoking, hypertension, diabetes, dyslipidemia, coronary artery disease, previous TIA) with one point assigned per factor (max 7 points). For CRP analysis, two patients with concomitant bacterial infection were excluded. Categorical variables were compared using Fisher's exact test and continuous variables using the Wilcoxon rank‐sum test. IQR is reported as the 1st and 3rd quartiles. Effect sizes with corresponding 95% confidence intervals refer to the Hodges–Lehmann estimator when the Wilcoxon rank‐sum test is used (continuous variables; > 0 indicates higher values in the “TOAST changed” group) and to the odds ratio when Fisher's exact test is used (categorical variables; > 1 indicates a higher proportion in the “TOAST changed” group). No correction for multiple testing was performed for this exploratory analysis.Abbreviations: BMI, body mass index; CI, confidence interval; CRP, C‐reactive protein; CTA, computed tomography angiography; HDL, high‐density lipoprotein cholesterol; HL, Hodges–Lehmann; IQR, interquartile range; LDL, low‐density lipoprotein cholesterol; NEMC, North Estonia Medical Centre; NIHSS, National Institutes of Health Stroke Scale; OR, odds ratio; TIA, transient ischemic attack; TOAST, Trial of ORG 10172 in Acute Stroke Treatment; VWI, vessel wall imaging. Exploratory comparison of characteristics according to TOAST reclassification after VWI. Note: CTA ≥ 50% stenosis was assessed in the intracranial internal carotid artery and middle cerebral artery M1–M2 segments; occlusions were not included. The atherosclerosis composite score was calculated as the sum of selected risk factors (BMI ≥ 30, current smoking, hypertension, diabetes, dyslipidemia, coronary artery disease, previous TIA) with one point assigned per factor (max 7 points). For CRP analysis, two patients with concomitant bacterial infection were excluded. Categorical variables were compared using Fisher's exact test and continuous variables using the Wilcoxon rank‐sum test. IQR is reported as the 1st and 3rd quartiles. Effect sizes with corresponding 95% confidence intervals refer to the Hodges–Lehmann estimator when the Wilcoxon rank‐sum test is used (continuous variables; > 0 indicates higher values in the “TOAST changed” group) and to the odds ratio when Fisher's exact test is used (categorical variables; > 1 indicates a higher proportion in the “TOAST changed” group). No correction for multiple testing was performed for this exploratory analysis. Abbreviations: BMI, body mass index; CI, confidence interval; CRP, C‐reactive protein; CTA, computed tomography angiography; HDL, high‐density lipoprotein cholesterol; HL, Hodges–Lehmann; IQR, interquartile range; LDL, low‐density lipoprotein cholesterol; NEMC, North Estonia Medical Centre; NIHSS, National Institutes of Health Stroke Scale; OR, odds ratio; TIA, transient ischemic attack; TOAST, Trial of ORG 10172 in Acute Stroke Treatment; VWI, vessel wall imaging. Discussion In this study, we prospectively investigated the impact of intracranial VWI in a young cohort with acute ischemic stroke in the MCA or AChA territory. We found that VWI resulted in a substantial rate of etiologic reclassification according to both TOAST and CCS criteria, with a predominant shift from cryptogenic stroke and cardioembolism to LAA. This was primarily driven by the detection of intracranial plaques that were not apparent or were present but not etiologically classifiable on conventional luminal imaging. These findings suggest that VWI provides additional diagnostic value in the etiologic assessment of young stroke, especially in cases where the cause remains unclear despite a comprehensive conventional assessment. The main finding of our study was that VWI led to a change in TOAST etiologic classification in 13% of cases. Previous retrospective studies have demonstrated an even greater impact of intracranial VWI, with reclassification rates up to 55% of cases [9, 10]. The lower reclassification rate in our cohort was most likely due to the prospective nature of our study, in which all patients regardless of suspected stroke etiology or luminal imaging findings underwent VWI. This suggests that the diagnostic impact of VWI in routine clinical practice may be lower than previously reported. In our cohort, only a small minority of patients with baseline TOAST classifications of cardioembolism, SVO, or other determined etiology underwent reclassification after VWI, suggesting that if the baseline diagnosis is evident, VWI is less likely to provide additional diagnostic value. A common factor of previous retrospective studies is the inclusion of patients with a high degree of stenosis on luminal imaging [17, 18, 19]. Interestingly, CTA ≥ 50% stenosis was not predictive of a change in etiologic classification following VWI in our cohort, implying that VWI might still provide diagnostic information even in the absence of significant luminal stenosis. This is consistent with pathological studies demonstrating that advanced atherosclerosis may be present even without high‐grade luminal narrowing [20]. Similar findings were recently reported in a prospective young stroke cohort, in which approximately half of vessel wall lesions had no corresponding MRA abnormality [21]. This observation may also partly reflect the fact that significant luminal stenosis frequently provides sufficient evidence for etiologic classification on conventional imaging alone. Etiologic reclassification occurred most frequently in cryptogenic strokes, with 40% of these cases being assigned to a different etiology after VWI. This finding is clinically relevant, as uncertainty regarding stroke mechanism remains a major challenge in young stroke populations. Cryptogenic stroke has been reported to account for up to 40% of ischemic strokes [1], with higher proportions in young stroke populations. The observed rate of cryptogenic stroke in our cohort was at the lower end of the reported range, which may partly be explained by the inclusion of patients up to 59 years of age and the restriction to strokes in the MCA and AChA territories. Accurate etiologic classification is particularly relevant in this population, as it directly influences secondary prevention strategies, including aggressive vascular risk factor control. In our study, the etiologic classification after VWI predominantly shifted toward LAA. Similar observations have been reported in previous studies of cryptogenic stroke, where VWI frequently identified unrecognized intracranial atherosclerotic plaques [9, 10, 22]. Notably, most prior VWI studies have been derived from East Asian cohorts, where intracranial atherosclerosis is more prevalent. Our findings suggest that these observations may also be applicable to European young stroke populations. A recent study found intracranial vessel wall lesions in approximately one‐quarter of young stroke patients, further supporting the potential contribution of nonstenotic intracranial atherosclerosis in this population [21]. Various atherosclerotic plaque characteristics, including plaque enhancement, plaque thickness, T1‐weighted hyperintensity, and stenosis severity, have been shown to be associated with symptomatic plaques and stroke recurrence [23]. However, the most prominent feature, plaque enhancement, is also known to occur in asymptomatic intracranial plaques [24], limiting its specificity for identifying the culprit lesion. Although no composite scoring system has yet been established to consistently distinguish culprit from non‐culprit plaques, it is likely that a combination of plaque characteristics would provide greater diagnostic accuracy than any single feature alone [25]. ICAD is one of the most common causes of stroke worldwide, accounting for up to half of the ischemic strokes in Asian countries [26]. In European countries, the reported prevalence of symptomatic ICAD is considerably lower, ranging from 3% to 43% [26, 27]. These estimates are based on heterogeneous methodologies, including different stenosis thresholds and the use of transcranial Doppler ultrasound, and may lack sensitivity [26]. It is therefore plausible that more sensitive imaging techniques, such as VWI, may increase recognition of intracranial atherosclerosis and its contribution to stroke etiology. The shift toward LAA observed after VWI in our cohort is consistent with this hypothesis. Based on current literature, it remains unclear which patients benefit most from VWI and in which clinical situations VWI should routinely be used. Recently, consensus recommendations from the Society for Magnetic Resonance Angiography suggested that VWI is likely useful for identifying symptomatic nonstenotic disease and differentiating causes of intracranial arterial narrowing [28]. Cryptogenic stroke was not included among the proposed indications, although in an earlier survey from the American Society of Neuroradiology, 41% of respondents reported using VWI in this setting [29]. Similarly, a recently proposed diagnostic algorithm for young stroke did not include VWI as part of the routine evaluation [5]. Our findings suggest that cryptogenic stroke may represent an additional clinical scenario in which VWI could provide meaningful diagnostic information, whereas its additional value appears more limited when stroke etiology is already well established, such as in established LAA or dissection. Given the additional imaging time and costs associated with VWI, it would be useful to identify clinical markers associated with increased diagnostic yield. In our exploratory analysis, higher CRP values showed a nominal association with TOAST reclassification. However, given the small number of reclassified patients and the multiple unadjusted comparisons, this finding may have arisen by chance and should be interpreted cautiously. Although elevated CRP has previously been associated with atherosclerotic activity [30, 31] and plaque enhancement on VWI is considered a marker of local inflammatory activity within the vessel wall [32], the potential relationship between inflammatory markers and the diagnostic yield of VWI requires confirmation in larger prospective studies. An additional factor potentially influencing the diagnostic yield of VWI is prior thrombectomy. Procedure‐related vessel wall abnormalities were observed in more than half of patients who underwent thrombectomy in our cohort. Although thrombectomy was not associated with TOAST reclassification, no patient with TOAST classification change had undergone the procedure. Excluding patients who underwent thrombectomy resulted in a similar TOAST reclassification rate (16%), indicating that procedure‐related vessel wall abnormalities did not drive the observed etiologic reclassifications. It is nevertheless plausible that previous thrombectomy may complicate VWI interpretation by mimicking or obscuring underlying etiologic vessel wall lesions. The main strength of this study is the prospective application of intracranial VWI in a consecutive cohort of young stroke patients, irrespective of baseline etiologic suspicion, with imaging performed shortly after stroke onset. Unlike previous studies that predominantly included selected patient groups or did not specifically assess changes in formal stroke etiologic classification after VWI, our design provides an estimate of the diagnostic impact of VWI in a setting that more closely reflects routine clinical practice. The two‐centre design, together with the high interobserver agreement between neurologists and neuroradiologists, supports the reproducibility and generalizability of the findings across different clinical settings. However, some limitations should be considered when interpreting the findings. The absence of a definitive reference standard for intracranial vessel wall pathology limits diagnostic certainty. In this context, the identification of intracranial atherosclerotic plaques does not necessarily establish causality, and incidental or non‐culprit lesions may have contributed to the observed shift toward LAA. The relatively small sample size and, particularly, the small number of patients with etiologic reclassification limited statistical power and precision in the subgroup analyses. Consequently, associations identified in the exploratory analyses are accompanied by substantial uncertainty and should be considered hypothesis‐generating. Moreover, the upper age limit of 59 years was higher than that used in some young stroke studies, which may limit direct comparability; however, a sensitivity analysis restricted to patients aged ≤ 50 years showed a similar rate of TOAST reclassification. It is also possible that some eligible patients were inadvertently not screened for inclusion, but we do not believe this substantially affected the composition of the cohort. In addition, interpretation of VWI findings may be subject to observer bias for both radiologists and neurologists. However, the use of independent blinded raters and the observed high interobserver agreement likely mitigate this limitation. Furthermore, VWI was performed using different MRI systems and acquisition protocols, with differences in spatial resolution that may have affected the detection and characterization of vessel wall abnormalities and therefore represent a potential source of confounding. All examinations were performed at 3 T using a common core VWI protocol, and this technical heterogeneity reflects implementation of VWI across different clinical settings. Finally, we did not evaluate whether etiologic reclassification after VWI resulted in changes in secondary prevention management or clinical outcomes. Although such diagnostic changes may have important clinical implications, this remains to be established. Conclusion In our prospective study, VWI resulted in a substantial rate of etiologic reclassification in young stroke patients, indicating that it may provide incremental diagnostic value in etiologic assessment. Our findings suggest that VWI may be particularly informative in cryptogenic stroke, where uncertainty regarding stroke mechanism remains a major clinical challenge. Larger prospective studies are warranted to confirm these findings and to determine whether VWI should be routinely performed in all young patients with cryptogenic stroke or selectively in patients with additional high‐risk features. Author Contributions Juhan Reimand: conceptualization, formal analysis, funding acquisition, investigation, methodology, writing – original draft, writing – review and editing. Siim Schneider: conceptualization, funding acquisition, writing – original draft, writing – review and editing, methodology, investigation. Riina Vibo: conceptualization, data curation, investigation, funding acquisition, writing – review and editing, methodology, supervision. Liisa Kõrv: data curation, investigation, writing – review and editing. Dagmar Loorits: investigation, writing – review and editing. Karl Jõgi: investigation, writing – review and editing. Teele Tuularu: investigation, writing – review and editing. Nele Taba: data curation, formal analysis, methodology, writing – review and editing. Kateriine Orav: investigation, writing – review and editing. Janika Kõrv: conceptualization, supervision, investigation, funding acquisition, writing – review and editing, methodology. Funding This study received funding from a research grant from Estonian Research Council (PRG1915). Ethics Statement The study was conducted in accordance with the ethical principles of the World Medical Association Declaration of Helsinki. Written informed consent was obtained from all subjects before the study. The study has been approved by the Research Ethics Committee of the University of Tartu (approval no. 384/M‐21). Conflicts of Interest Janika Kõrv is a member of the Editorial Board of the European Journal of Neurology. She was not involved in the review or editorial decision‐making process for this manuscript. The authors report no other conflicts of interest related to this manuscript. Supporting information Table S1: Acquired in‐plane resolution of VWI sequences. Table S2: TOAST classification before and after incorporation of VWI findings. Table S3: CCS classification before and after incorporation of VWI findings. Table S4: Logistic regression models for predictors of TOAST reclassification after VWI. Acknowledgments The authors would like to thank the study participants and personnel in the departments of neurology and radiology in both NEMC and TUH for their assistance in various parts of the study. Data Availability Statement The data that support the findings of this study are available from the corresponding author upon reasonable request. References Associated Data Supplementary Materials Table S1: Acquired in‐plane resolution of VWI sequences. Table S2: TOAST classification before and after incorporation of VWI findings. Table S3: CCS classification before and after incorporation of VWI findings. Table S4: Logistic regression models for predictors of TOAST reclassification after VWI. Data Availability Statement The data that support the findings of this study are available from the corresponding author upon reasonable request.