BACKGROUND AND PURPOSE:Embolic stroke of undetermined source (ESUS) may be associated with carotid artery plaques with <50% stenosis. Plaque vulnerability is multifactorial, possibly related to intraplaque hemorrhage (IPH), lipid-rich necrotic core, perivascular adipose tissue (PVAT), and calcifications. Machine learning (ML)-based plaque classification is increasingly popular but often limited in clinical interpretability by black-box nature. We applied an explainable ML approach, using noncalcified plaque components and calcification features with the SHapley Additive exPlanations (SHAP) framework to classify plaques as culprit or nonculprit. METHODS:This was a retrospective, cross-sectional study. Patients with unilateral anterior circulation ESUS with calcified carotid plaques in neck computed tomography (CT) angiography were analyzed. Calcification-level features were derived from manual segmentations. Plaque-level features were assessed by a neuroradiologist and by semi-automated software. Plaques were classified as culprit if ipsilateral to stroke side. Eight classifiers were benchmarked, and a gradient-boosted decision tree (CatBoost) was further tuned. SHAP explained model decisions. RESULTS:Seventy patients yielded 116 calcified plaques (270 calcifications). Model based on five plaque- and calcification-level features achieved ROC-AUC (receiver operating characteristic area under the curve) 0.79 and precision-recall-AUC 0.86, outperforming classification based on plaque thickness ≥3 mm (ROC-AUC 0.59, p = 0.04) and IPH presence (ROC-AUC 0.51, p = 0.003). SHAP identified plaque thickness and PVAT volume as the most influential features with potential thresholds of >2.6 mm and ≥112 mm3, respectively.f CONCLUSIONS: ML model trained with noncalcified plaque and calcification features can classify culprit calcified carotid plaque better than conventional criteria. Using clinically interpretable features with SHAP, the model explained its decisions and suggested hypothesis-generating thresholds.

BACKGROUND: A modified computed tomography angiography (CTA)–based Carotid Plaque Reporting and Data System (Plaque-RADS) classification was applied to a cohort of patients with embolic stroke of undetermined source to test whether high-risk Plaque-RADS subtypes are more prevalent on the ipsilateral side of stroke. With the widespread use of CTA for stroke evaluation, a CTA-based Plaque-RADS would be valuable for generalizability. METHODS: A retrospective observational cross-sectional study was conducted at a single integrated health system comprised of 3 hospitals with a comprehensive stroke center between October 1, 2015, and April 1, 2017. Patients with unilateral anterior circulation stroke and <50% carotid stenosis on CTA were retrospectively identified. Maximum plaque thickness and ulceration were assessed by a neuroradiologist blinded to the stroke side. A semiautomated segmentation software measured intraplaque hemorrhage volumes. Modified CTA-based Plaque-RADS classification was defined as (1) no plaque, (2) plaque thickness <3 mm, (3) plaque thickness ≥3 mm or ulcerated, and (4) plaque with intraplaque hemorrhage >50 mm 3 irrespective of plaque thickness. High-risk plaque subtypes (Plaque-RADS 3 and 4) were compared with low-risk subtypes (Plaque-RADS 1 and 2). RESULTS: Ninety-four patients (55% women; median age, 66 years) were included. CTA-based Plaque-RADS categories for plaques ipsilateral to the stroke side were as follows: (1) 14.9%, (2) 42.6%, (3) 41.5%, and (4) 1.1%. Carotid plaques contralateral to stroke side were Plaque-RADS: (1) 21.3%, (2) 46.8%, (3) 31.9%, and (4) 0%. When compared with the contralateral side, plaques ipsilateral to the stroke side were significantly associated with high-risk Plaque-RADS subtypes in a mixed-effects logistic model adjusting for age and sex (adjusted odds ratio, 2.10 [95% CI, 1.20–3.71]; P =0.01). CONCLUSIONS: Carotid plaque ipsilateral to the stroke side was significantly associated with CTA-based high-risk Plaque-RADS subtypes in an embolic stroke of undetermined source cohort. A CTA-based Plaque-RADS classification may be useful for identifying potentially causative carotid plaque phenotypes in patients with embolic stroke of undetermined source.

Background Complex aortic plaque (CAP) is a potential embolic source in patients with cryptogenic stroke (CS). We review CAP imaging criteria for transesophageal echocardiogram (TEE), computed tomography angiography (CTA), and magnetic resonance imaging and calculate CAP prevalence in patients with acute CS. Methods and Results PubMed and EMBASE databases were searched up to December 2022 in accordance with the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses guideline. Two independent reviewers extracted data on study design, imaging techniques, CAP criteria, and prevalence. The Cochrane Collaboration tool and Guideline for Reporting Reliability and Agreement Studies were used to assess risk of bias and reporting completeness, respectively. From 2293 studies, 45 were reviewed for CAP imaging biomarker criteria in patients with acute CS (N=37 TEE; N=9 CTA; N=6 magnetic resonance imaging). Most studies (74%) used ≥4 mm plaque thickness as the imaging criterion for CAP although ≥1 mm (N=1, CTA), ≥5 mm (N=5, TEE), and ≥6 mm (N=2, CTA) were also reported. Additional features included mobility, ulceration, thrombus, protrusions, and assessment of plaque composition. From 23 prospective studies, CAP was detected in 960 of 2778 patients with CS (0.32 [95% CI, 0.24–0.41], I2=94%). By modality, prevalence estimates were 0.29 (95% CI, 0.20–0.40; I2=95%) for TEE; 0.23 (95% CI, 0.15–0.34; I2=87%) for CTA and 0.22 (95% CI, 0.06–0.54; I2=92%) for magnetic resonance imaging. Conclusions TEE was commonly used to assess CAP in patients with CS. The most common CAP imaging biomarker was ≥4 mm plaque thickness. CAP was observed in one‐third of patients with acute CS. However, high study heterogeneity suggests a need for reproducible imaging methods.

Introduction: In patients with cryptogenic stroke (CS), complex aortic plaque may be a potential underlying etiology. We performed a systematic review to determine the prevalence of complex aortic plaque in CS patients. Methods: A systematic review and meta-analysis were performed according to PRISMA guidelines (PROSPERO: CRD42022300865). PubMed and EMBASE databases were searched from Jan 1980 to Nov 2021 for studies assessing aortic (ascending, arch, descending) plaque by transesophageal echocardiogram (TEE), CT/CTA, or MRI in at least 10 CS patients. Prevalence rates were pooled using a random-effects model. I 2 statistics assessed heterogeneity. An Egger’s test assessed publication bias. Results: From 2712 articles, 31 met inclusion criteria. Ascending, arch, and descending aorta were assessed in 65%, 100%, 55% of studies, respectively. Studies investigated aortic plaque by TEE (84%), CT/CTA (19%) and MRI (16%). The prevalence of complex aortic plaque in 4666 CS patients was heterogeneous across studies and yielded a summary prevalence of 30% (95% CI 23-38%, I 2 = 96%; Figure 1) contrasting with 11% (95% CI 5-20%, I 2 =83%) in 677 patients without stroke. Prevalence rates in women and men were 26% (95% CI 14%-43%, I 2 = 94%) and 35% (95% CI 21-52%, I 2 = 97%), respectively. To investigate geographic differences, 14 studies from Europe were pooled (32%, 95% CI 23-43%, I 2 =93%), 3 from the Middle East (34%, 95% CI 11-67%, I 2 =94%) and 3 from the US (28%, 95% CI 13-51%, I 2 =90%). No publication bias was detected (p=0.66). Sources of heterogeneity included patient selection, imaging technology (e.g, transducer frequency) and plaque measurement criteria. Conclusions: Studies suggest a prevalence rate of complex aortic plaque in approximately 30% of CS patients. However, significant heterogeneity in the results indicate a need for less variability in CS patient selection and more reproducible imaging methods/criteria for detecting complex aortic plaque.

Background Accurate differentiation of pseudoprogression (PsP) from tumor progression (TP) in glioblastomas (GBMs) is essential for appropriate clinical management and prognostication of these patients. In the present study, we sought to validate the findings of our previously developed multiparametric MRI model in a new cohort of GBM patients treated with standard therapy in identifying PsP cases. Methods Fifty-six GBM patients demonstrating enhancing lesions within 6 months after completion of concurrent chemo-radiotherapy (CCRT) underwent anatomical imaging, diffusion and perfusion MRI on a 3 T magnet. Subsequently, patients were classified as TP + mixed tumor (n = 37) and PsP (n = 19). When tumor specimens were available from repeat surgery, histopathologic findings were used to identify TP + mixed tumor (> 25% malignant features; n = 34) or PsP (< 25% malignant features; n = 16). In case of non-availability of tumor specimens, ≥ 2 consecutive conventional MRIs using mRANO criteria were used to determine TP + mixed tumor (n = 3) or PsP (n = 3). The multiparametric MRI-based prediction model consisted of predictive probabilities (PP) of tumor progression computed from diffusion and perfusion MRI derived parameters from contrast enhancing regions. In the next step, PP values were used to characterize each lesion as PsP or TP+ mixed tumor. The lesions were considered as PsP if the PP value was < 50% and TP+ mixed tumor if the PP value was ≥ 50%. Pearson test was used to determine the concordance correlation coefficient between PP values and histopathology/mRANO criteria. The area under ROC curve (AUC) was used as a quantitative measure for assessing the discriminatory accuracy of the prediction model in identifying PsP and TP+ mixed tumor. Results Multiparametric MRI model correctly predicted PsP in 95% (18/19) and TP+ mixed tumor in 57% of cases (21/37) with an overall concordance rate of 70% (39/56) with final diagnosis as determined by histopathology/mRANO criteria. There was a significant concordant correlation coefficient between PP values and histopathology/mRANO criteria (r = 0.56; p < 0.001). The ROC analyses revealed an accuracy of 75.7% in distinguishing PsP from TP+ mixed tumor. Leave-one-out cross-validation test revealed that 73.2% of cases were correctly classified as PsP and TP + mixed tumor. Conclusions Our multiparametric MRI based prediction model may be helpful in identifying PsP in GBM patients.

Despite advancements in multi-modal imaging techniques, a substantial portion of ischemic stroke patients today remain without a diagnosed etiology after conventional workup. Based on existing diagnostic criteria, these ischemic stroke patients are subcategorized into having cryptogenic stroke (CS) or embolic stroke of undetermined source (ESUS). There is growing evidence that in these patients, non-cardiogenic embolic sources, in particular non-stenosing atherosclerotic plaque, may have significant contributory roles in their ischemic strokes. Recent advancements in vessel wall MRI (VW-MRI) have enabled imaging of vessel walls beyond the degree of luminal stenosis, and allows further characterization of atherosclerotic plaque components. Using this imaging technique, we are able to identify potential imaging biomarkers of vulnerable atherosclerotic plaques such as intraplaque hemorrhage, lipid rich necrotic core, and thin or ruptured fibrous caps. This review focuses on the existing evidence on the advantages of utilizing VW-MRI in ischemic stroke patients to identify culprit plaques in key anatomical areas, namely the cervical carotid arteries, intracranial arteries, and the aortic arch. For each anatomical area, the literature on potential imaging biomarkers of vulnerable plaques on VW-MRI as well as the VW-MRI literature in ESUS and CS patients are reviewed. Future directions on further elucidating ESUS and CS by the use of VW-MRI as well as exciting emerging techniques are reviewed.