Assessing internal carotid artery stenosis with a semiautomated computed tomography angiography tool and duplex ultrasound

Predictive stroke risk model with vision transformer-based Doppler features

BACKGROUND

Acute stroke is the leading cause of death and disability globally, with an estimated 16 million cases each year. The progression of carotid stenosis reduces blood flow to the intracranial vasculature, causing stroke. Early recognition of ischemic stroke is crucial for disease treatment and management.

PURPOSE

A computer-aided diagnosis (CAD) system was proposed in this study to rapidly evaluate ischemic stroke in carotid color Doppler (CCD).

METHODS

Based on the ground truth from the clinical examination report, the vision transformer (ViT) features extracted from all CCD images (513 stroke and 458 normal images) were combined in machine learning classifiers to generate the likelihood of ischemic stroke for each image. The pretrained weights from ImageNet reduced the time-consuming training process. The accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve were calculated to evaluate the stroke prediction model. The chi-square test, DeLong test, and Bonferroni correction for multiple comparisons were applied to deal with the type-I error. Only p values equal to or less than 0.00125 were considered to be statistically significant.

RESULTS

The proposed CAD system achieved an accuracy of 89%, a sensitivity of 94%, a specificity of 84%, and an area under the receiver operating characteristic curve of 0.95, outperforming the convolutional neural networks AlexNet (82%, p < 0.001), Inception-v3 (78%, p < 0.001), ResNet101 (84%, p < 0.001), and DenseNet201 (85%, p < 0.01). The computational time in model training was only 30 s, which would be efficient and practical in clinical use.

CONCLUSIONS

The experiment shows the promising use of CCD images in stroke estimation. Using the pretrained ViT architecture, the image features can be automatically and efficiently generated without human intervention. The proposed CAD system provides a rapid and reliable suggestion for diagnosing ischemic stroke.

Impaired dynamic cerebral autoregulation measured in the middle cerebral artery in patients with vertebrobasilar ischemia is associated with autonomic failure

OBJECTIVES

To assess whether vertebrobasilar artery ischemia (VBI) affects cortical cerebral blood flow (CBF) regulation.

MATERIAL AND METHODS

107 consecutive patients (mean age 65 ± 15 years; women 21) with VBI underwent structured stroke care with assessment of dynamic cerebral autoregulation (dCA) in both middle cerebral arteries (MCAs) by transfer function analysis using spontaneous oscillations of blood pressure (BP) and CBF velocity that yields by extraction of phase and gain information in the very low (0.02-0.07 Hz), low (0.07-0.15 Hz) and high frequency (0.15-0.5 Hz) ranges. Additionally, power spectrum analysis of BP and heart rate variability (HRV) was performed. The control group consists of 29 age- and sex-matched healthy persons.

RESULTS

Compared to controls, phase in the VBI patients was significantly reduced and gain increased in the very low frequencies (VLF), in the low (LF), phase was significantly reduced only ipsilaterally. In the high frequencies (HF), phase reduction was only marginally significant. BP power spectral density (PSD) was much higher in the patients than in the controls across all frequencies. In the PSD of heart rate variability the controls but not the patients exhibited a strong peak around 0.11Hz, while the patients, but not the controls, exhibit a strong peak around 0.36 Hz. In regression analysis, patient's phase and gain results were not related to age, sex, arterial hypertension, diabetes mellitus, renal dysfunction, heart failure as indicated by left ventricular ejection fraction, stroke subtype, presence or absence of cerebral small vessel disease.

CONCLUSION

Patients with VBI exhibit bilateral cortical autoregulation impairment in association with an autonomic nervous system disbalance.

CLINICALTRIALS

GOV IDENTIFIER

NCT04611672.

Peak systolic velocity ratio for evaluation of internal carotid artery stenosis correlated with plaque morphology: substudy results of the ANTIQUE study

Background

Accurate assessment of carotid stenosis severity is important for proper patient management. The present study aimed to compare the evaluation of carotid stenosis severity using four duplex sonography (DUS) measurements, including peak systolic velocity (PSV), PSV ratio in stenosis and distal to stenosis (PSVICA/ICA ratio), end-diastolic velocity (EDV), and B-mode, with computed tomography angiography (CTA), and to evaluate the impact of plaque morphology on correlation between DUS and CTA.

Methods

Consecutive patients with carotid stenosis of ≥40% examined using DUS and CTA were included. Plaque morphology was also determined using magnetic resonance imaging. Spearman's correlation and Kendall's rank correlation were used to evaluate the results.

Results

A total of 143 cases of internal carotid artery stenosis of ≥40% based on DUS were analyzed. The PSVICA/ICA ratio showed the highest correlation [Spearman's correlation r = 0.576) with CTA, followed by PSV (r = 0.526), B-mode measurement (r = 0.482), and EDV (r = 0.441; p < 0.001 in all cases]. The worst correlation was found for PSV when the plaque was calcified (r = 0.238), whereas EDV showed a higher correlation (r = 0.523). Correlations of B-mode measurement were superior for plaques with smooth surface (r = 0.677), while the PSVICA/ICA ratio showed the highest correlation in stenoses with irregular (r = 0.373) or ulcerated (r = 0.382) surfaces, as well as lipid (r = 0.406), fibrous (r = 0.461), and mixed (r = 0.403; p < 0.01 in all cases) plaques. Nevertheless, differences between the mentioned correlations were not statistically significant (p > 0.05 in all cases).

Conclusion

PSV, PSVICA/ICA ratio, EDV, and B-mode measurements showed comparable correlations with CTA in evaluation of carotid artery stenosis based on their correlation with CTA results. Heavy calcifications and plaque surface irregularity or ulceration negatively influenced the measurement accuracy.

Ultrasonography Grading of Internal Carotid Artery Disease: Multiparametric German Society of Ultrasound in Medicine (DEGUM) versus Society of Radiologists in Ultrasound (SRU) Consensus Criteria

PURPOSE

 We sought to determine the diagnostic agreement between the revised ultrasonography approach by the German Society of Ultrasound in Medicine (DEGUM) and the established Society of Radiologists in Ultrasound (SRU) consensus criteria for the grading of carotid artery disease.

MATERIALS AND METHODS

 Post-hoc analysis of a prospective multicenter study, in which patients underwent ultrasonography and digital subtraction angiography (DSA) of carotid arteries for validation of the DEGUM approach. According to DEGUM and SRU ultrasonography criteria, carotid arteries were independently categorized into clinically relevant NASCET strata (normal, mild [1-49 %], moderate [50-69 %], severe [70-99 %], occlusion). On DSA, carotid artery findings according to NASCET were considered the reference standard.

RESULTS

 We analyzed 158 ultrasonography and DSA carotid artery pairs. There was substantial agreement between both ultrasonography approaches for severe (κw 0.76, CI95 %: 0.66-0.86), but only fair agreement for moderate (κw 0.38, CI95 %: 0.19-0.58) disease categories. Compared with DSA, both ultrasonography approaches were of equal sensitivity (79.7 % versus 79.7 %; p = 1.0) regarding the identification of severe stenosis, yet the DEGUM approach was more specific than the SRU approach (70.2 % versus 56.4 %, p = 0.0002). There was equality of accuracy parameters (p > 0.05) among both ultrasonography approaches for the other ranges of carotid artery disease.

CONCLUSION

 While the sensitivity was equivalent, false-positive identification of severe carotid artery stenosis appears to be more frequent when using the SRU ultrasonography approach than the revised multiparametric DEGUM approach.

Duplex ultrasound for diagnosing symptomatic carotid stenosis in the extracranial segments

BACKGROUND

Carotid artery stenosis is an important cause of stroke and transient ischemic attack. Correctly and rapidly identifying patients with symptomatic carotid artery stenosis is essential for adequate treatment with early cerebral revascularization. Doubts about the diagnostic value regarding the accuracy of duplex ultrasound (DUS) and the possibility of using DUS as the single diagnostic test before carotid revascularization are still debated.

OBJECTIVES

To estimate the accuracy of DUS in individuals with symptomatic carotid stenosis verified by either digital subtraction angiography (DSA), computed tomography angiography (CTA), or magnetic resonance angiography (MRA).

SEARCH METHODS

We searched CRDTAS, CENTRAL, MEDLINE (Ovid), Embase (Ovid), ISI Web of Science, HTA, DARE, and LILACS up to 15 February 2021. We handsearched the reference lists of all included studies and other relevant publications and contacted experts in the field to identify additional studies or unpublished data.

SELECTION CRITERIA

We included studies assessing DUS accuracy against an acceptable reference standard (DSA, MRA, or CTA) in symptomatic patients. We considered the classification of carotid stenosis with DUS defined with validated duplex velocity criteria, and the NASCET criteria for carotid stenosis measures on DSA, MRA, and CTA. We excluded studies that included < 70% of symptomatic patients; the time between the index test and the reference standard was longer than four weeks or not described, or that presented no objective criteria to estimate carotid stenosis.

DATA COLLECTION AND ANALYSIS

The review authors independently screened articles, extracted data, and assessed the risk of bias and applicability concerns using the QUADAS-2 domain list. We extracted data with an effort to complete a 2 × 2 table (true positives, true negatives, false positives, and false negatives) for each of the different categories of carotid stenosis and reference standards. We produced forest plots and summary receiver operating characteristic (ROC) plots to summarize the data. Where meta-analysis was possible, we used a bivariate meta-analysis model.

MAIN RESULTS

We identified 25,087 unique studies, of which 22 were deemed eligible for inclusion (4957 carotid arteries). The risk of bias varied considerably across the studies, and studies were generally of moderate to low quality. We narratively described the results without meta-analysis in seven studies in which the criteria used to determine stenosis were too different from the duplex velocity criteria proposed in our protocol or studies that provided insufficient data to complete a 2 × 2 table for at least in one category of stenosis. Nine studies (2770 carotid arteries) presented DUS versus DSA results for 70% to 99% carotid artery stenosis, and two (685 carotid arteries) presented results from DUS versus CTA in this category. Seven studies presented results for occlusion with DSA as the reference standard and three with CTA as the reference standard. Five studies compared DUS versus DSA for 50% to 99% carotid artery stenosis. Only one study presented results from 50% to 69% carotid artery stenosis. For DUS versus DSA, for < 50% carotid artery stenosis, the summary sensitivity was 0.63 (95% confidence interval [CI] 0.48 to 0.76) and the summary specificity was 0.99 (95% CI 0.96 to 0.99); for the 50% to 69% range, only one study was included and meta-analysis not performed; for the 50% to 99% range, the summary sensitivity was 0.97 (95% CI 0.95 to 0.98) and the summary specificity was 0.70 (95% CI 0.67 to 0.73); for the 70% to 99% range, the summary sensitivity was 0.85 (95% CI 0.77 to 0.91) and the summary specificity was 0.98 (95% CI 0.74 to 0.90); for occlusion, the summary sensitivity was 0.91 (95% CI 0.81 to 0.97) and the summary specificity was 0.95 (95% CI 0.76 to 0.99). For sensitivity analyses, excluding studies in which participants were selected based on the presence of occlusion on DUS had an impact on specificity: 0.98 (95% CI 0.97 to 0.99). For DUS versus CTA, we found two studies in the range of 70% to 99%; the sensitivity varied from 0.57 to 0.94 and the specificity varied from 0.87 to 0.98. For occlusion, the summary sensitivity was 0.95 (95% CI 0.80 to 0.99) and the summary specificity was 0.91 (95% CI 0.09 to 0.99). For DUS versus MRA, there was one study with results for 50% to 99% carotid artery stenosis, with a sensitivity of 0.88 (95% CI 0.70 to 0.98) and specificity of 0.60 (95% CI 0.15 to 0.95); in the 70% to 99% range, two studies were included, with sensitivity that varied from 0.54 to 0.99 and specificity that varied from 0.78 to 0.89. We could perform only a few of the proposed sensitivity analyses because of the small number of studies included.

AUTHORS' CONCLUSIONS

This review provides evidence that the diagnostic accuracy of DUS is high, especially at discriminating between the presence or absence of significant carotid artery stenosis (< 50% or 50% to 99%). This evidence, plus its less invasive nature, supports the early use of DUS for the detection of carotid artery stenosis. The accuracy for 70% to 99% carotid artery stenosis and occlusion is high. Clinicians should exercise caution when using DUS as the single preoperative diagnostic method, and the limitations should be considered. There was little evidence of the accuracy of DUS when compared with CTA or MRA. The results of this review should be interpreted with caution because they are based on studies of low methodological quality, mainly due to the patient selection method. Methodological problems in participant inclusion criteria from the studies discussed above apparently influenced an overestimated estimate of prevalence values. Most of the studies included failed to precisely describe inclusion criteria and previous testing. Future diagnostic accuracy studies should include direct comparisons of the various modalities of diagnostic tests (mainly DUS, CTA, and MRA) for carotid artery stenosis since DSA is no longer considered to be the best method for diagnosing carotid stenosis and less invasive tests are now used as reference standards in clinical practice. Also, for future studies, the participant inclusion criteria require careful attention.

CT Angiography Manual Multiplanar Vessel Diameter Measurement vs. Semiautomated Perpendicular Area Minimal Caliber Computation of Internal Carotid Artery Stenosis

Objective: To determine the diagnostic agreement of CT angiography (CTA) manual multiplanar reformatting (MPR) stenosis diameter measurement and semiautomated perpendicular stenosis area minimal caliber computation of extracranial internal carotid artery (ICA) stenosis.

Methods: We analyzed acute cerebral ischemia CTA at our tertiary stroke center in a 12-month period. Prospective NASCET-type stenosis grading for each ICA was independently performed using (1) MPR to manually determine diameters and (2) perpendicular stenosis area with minimal caliber semiautomated computation to grade luminal constriction. Corresponding to clinically relevant NASCET strata, results were grouped into severity ranges: normal, 1–49%, 50–69%, and 70–99%, and occlusion.

Results: We included 647 ICA pairs from 330 patients (median age of 74 [66–80, IQR]; 38–92 years; 58% men; median NIHSS 4 [1–9, IQR]). MPR diameter and semiautomated caliber measurements resulted in stenosis grades of 0–49% in 143 vs. 93, 50–69% in 29 vs. 27, 70–99% in 6 vs. 14, and occlusion in 34 vs. 34 ICAs (p = 0.003), respectively. We found excellent reliability between repeated manual CTA assessments of one expert reader (ICC = 0.997; 95% CI, 0.993–0.999) and assessments of two expert readers (ICC = 0.972; 95% CI, 0.936–0.988). For the semiautomated vessel analysis software, both intrarater reliability and interrater reliability were similarly strong (ICC = 0.981; 95% CI, 0.952–0.992 and ICC = 0.745; 95% CI, 0.486–0.883, respectively). However, Bland–Altman analysis revealed a mean difference of 1.6% between the methods within disease range with wide 95% limits of agreement (−16.7–19.8%). This interval even increased with exclusively considered vessel pairs of stenosis ≥1% (mean 5.3%; −24.1–34.7%) or symptomatic stenosis ≥50% (mean 0.1%; −25.7–26.0%).

Conclusion: Our findings suggest that MPR-based diameter measurement and the semiautomated perpendicular area minimal caliber computation methods cannot be used interchangeably for the quantification of ICA steno-occlusive disease.

Assessing Ischemic Stroke with Convolutional Image Features in Carotid Color Doppler

Stroke is a leading cause of disability and death worldwide. Early and accurate recognition of acute stroke is critical for achieving a good prognosis. The novel automated system proposed in this study was based on convolutional neural networks (CNNs), which were used to identify lesion findings on carotid color Doppler (CCD) images in patients with acute ischemic stroke. An image database composed of 1032 CCD images from 106 patients with acute ischemic stroke (549 images) and from 79 normal controls (483 images) was retrospectively analyzed. Taking the consensus of two neuroradiologists as the gold standard, different CNN models with and without transfer learning were evaluated with 10-fold cross-validation. The diagnostic information provided from individual color channels was also explored. AlexNet, which was trained from scratch, achieved an accuracy of 91.67%, a sensitivity of 93.33%, a specificity of 90.20% and an area under the receiver operating characteristic curves (AUC) of 0.9432. Other transferred models achieved accuracies between 77.69% and 83.94%. In channel comparisons, the green channel had the best performance, with an accuracy of 87.50%, a sensitivity of 97.78%, a specificity of 78.43% and an AUC of 0.9507. The proposed CNN architecture, as a computer-aided diagnosis system, suggests using automatic feature extraction from CCD images to predict ischemic stroke. The developed scheme has the potential to provide diagnostic suggestions in clinical use.

INVERSE PROBLEM ALGORITHM APPLICATION TO SEMI-QUANTITATIVE ANALYSIS OF 272 PATIENTS WITH ISCHEMIC STROKE SYMPTOMS: CAROTID STENOSIS RISK ASSESSMENT FOR FIVE RISK FACTORS

This study proposes the inverse problem algorithm (IPA) with five risk factors applied to the semi-quantitative analysis of carotid stenosis 272 patients with suspected ischemic stroke. The IPA is known to provide a substantiated machine learning-based prediction of the expected outcomes by solving an inverse matrix of variable coefficients. In case of carotid stenosis prediction, such risk factors as patient’s age, mean arterial pressure (MAP), glucose AC, low-density lipoprotein-cholesterol (LDL-C), and C-Reactive protein (CRP) were assessed for the main group of 217 patients. Their results were processed by the STATISTICA program with a customized loss function ([Formula: see text]), yielding the first-order nonlinear semi-empirical formula with 16 terms. The loss function was calculated via the total mismatch between the theoretical predictions and true carotid stenosis cases (%) for all 217 patients. Thus, the carotid stenosis (%) compromised solution array [[Formula: see text]] was optimized using [Formula: see text] individual data points via the proposed algorithm. The results showed a complete regression with loss function [Formula: see text]=2.3543, variance [Formula: see text]=87.46%, and correlation coefficient [Formula: see text]. The reference group of 55 more patients with the same preliminary diagnosis and symptoms was selected to validate the method predictive feasibility, which was found quite satisfactory. The decreasing order of three dominant risk factors was as follows: CRP, glucose AC, and MAP, whereas age and LDL-C weakly influenced the program computation results. The IPA showed a strong convergence by its default characteristic. The reduction of the number of variables in computation deteriorated the prediction accuracy, exhibiting the algorithm’s high sensitivity to the number of variables.

Imaging modalities to diagnose carotid artery stenosis: progress and prospect

In the past few decades, imaging has been developed to a high level of sophistication. Improvements from one-dimension (1D) to 2D images, and from 2D images to 3D models, have revolutionized the field of imaging. This not only helps in diagnosing various critical and fatal diseases in the early stages but also contributes to making informed clinical decisions on the follow-up treatment profile. Carotid artery stenosis (CAS) may potentially cause debilitating stroke, and its accurate early detection is therefore important. In this paper, the technical development of various CAS diagnosis imaging modalities and its impact on the clinical efficacy is thoroughly reviewed. These imaging modalities include duplex ultrasound (DUS), computed tomography angiography (CTA) and magnetic resonance angiography (MRA). For each of the imaging modalities considered, imaging methodology (principle), critical imaging parameters, and the extent of imaging the vulnerable plaque are discussed. DUS is usually the initial recommended CAS diagnostic examination. However, for the therapeutic intervention, either MRA or CTA is recommended for confirmation, and for added information on intracranial cerebral circulation and aortic arch condition for procedural planning. Over the past few decades, the focus of CAS diagnosis has also shifted from pure stenosis quantification to plaque characterization. This has led to further advancement in the existing imaging tools and development of other potential imaging tools like Optical coherence tomography (OCT), photoacoustic tomography (PAT), and infrared (IR) thermography.

Carotid Consensus Panel duplex criteria can replace modified University of Washington criteria without affecting accuracy

The decision to intervene for internal carotid stenosis often depends on the degree of stenosis seen on duplex ultrasound (US). The aim of this study is to compare the diagnostic accuracy of two criteria: modified University of Washington (UW) and 2003 Carotid Consensus Panel (CCP). All patients undergoing US in an accredited (IAC) vascular laboratory from January 2010 to June 2015 were reviewed ( n=18,772 US exams). Patients receiving a neck computed tomography angiography (CTA) within 6 months of the US were included in the study ( n=254). The degree of stenosis was determined by UW/CCP criteria and confirmed on CTA images using North American Symptomatic Carotid Endarterectomy Trial (NASCET)/European Carotid Surgery Trial (ECST) schema. Kappa analysis with 95% confidence intervals (CIs) were utilized to determine duplex–CTA agreement. A total of 417 carotid arteries from 221 patients were assessed in this study. The modified UW criteria accurately classified 266 (63.9%, kappa = 0.321, 95% CI 0.255 to 0.386) cases according to NASCET-derived measurements. The sensitivity, specificity, and accuracy at ≥ 60% stenosis were 65.7%, 81.3%, and 81.9%. The CCP criteria resulted in 296 (70.9%) accurate diagnoses (kappa = 0.359, 95% CI 0.280 to 0.437). At ≥ 70% stenosis, the sensitivity, specificity and accuracy were 38.8%, 91.6%, and 87.1% for NASCET. Comparison of the duplex results to ECST-derived CTA measurements revealed a similar trend (UW 53.1%, κ = 0.301 vs CCP 62.1%, κ = 0.315). The CCP criteria demonstrate a higher concordance rate with measurements taken from CTAs. The CCP criteria may be more sensitive in classifying clinically significant degrees of stenosis without a loss in diagnostic accuracy.

Multi-parametric ultrasound criteria for internal carotid artery disease-comparison with CT angiography

INTRODUCTION

The German Society of Ultrasound in Medicine (known by its acronym DEGUM) recently proposed a novel multi-parametric ultrasound approach for comprehensive and accurate assessment of extracranial internal carotid artery (ICA) steno-occlusive disease. We determined the agreement between duplex ultrasonography (DUS) interpreted by the DEGUM criteria and CT angiography (CTA) for grading of extracranial ICA steno-occlusive disease.

METHODS

Consecutive patients with acute cerebral ischemia underwent DUS and CTA. Internal carotid artery stenosis was graded according to the DEGUM-recommended criteria for DUS. Independent readers manually performed North American Symptomatic Carotid Endarterectomy Trial-type measurements on axial CTA source images. Both modalities were compared using Spearman's correlation and Bland-Altman analyses.

RESULTS

A total of 303 acute cerebral ischemia patients (mean age, 72 ± 12 years; 58 % men; median baseline National Institutes of Health Stroke Scale score, 4 [interquartile range 7]) provided 593 DUS and CTA vessel pairs for comparison. There was a positive correlation between DUS and CTA (r s  = 0.783, p < 0.001) with mean difference in degree of stenosis measurement of 3.57 %. Bland-Altman analysis further revealed widely varying differences (95 % limits of agreement -29.26 to 22.84) between the two modalities.

CONCLUSION

Although the novel DEGUM criteria showed overall good agreement between DUS and CTA across all stenosis ranges, potential for wide incongruence with CTA underscores the need for local laboratory validation to avoid false screening results.