Diagnostic Performance of CTA and vessel wall MR imaging in Carotid webs—A Retrospective Study

Background: There currently is no consensus on the performance of Vessel wall MRI (VW MRI) with carotid webs (CaWs). No clear data exist on compare the performance of each diagnostic features in CaWs. Purpose: To set a standards about the diagnostic features and compare the performance of each diagnostic features in CaWs. Materials and methods : All patients (n = 165, median age, 56.9 ± 13.3) who has done both neck CTA and VW MRI in our radiology department, Zhongnan hospital, performed from February 2017 to April 2019 were included. All CTA and VW MRI images were read by two experienced radiologists. Interobserver agreement for the two reader’s decision of CTA was checked by using kappa statistics. Sensitivity, specificity, and area under the curve (AUC) of each diagnostic features at CTA and VW MRI were compared. Results: 10 patients with carotid web was found. The feature “double lumen sign” of CTA and “valve sign” and “contrast stasis” of VW MRI showed good specificity (specificity, 98.6% vs 92.9% vs 93.6%, respectively; P<0.001). The feature “double lumen sign” of VW MRI compared to it of CTA showed higher sensitivity (sensitivity, 80.0% [8 of 10] vs 50.0% [5 of 10], respectively; P<0.001) and better diagnostic performances (area under the curve [AUC], 0.93 [95% CI: 0.68, 0.97] and 0.74 [95% CI: 0.54, 0.94], respectively; P<0.001). Conclusion: The feature “double lumen sign” and “a line or septum” of CTA and “valve sign” and “contrast stasis” of VW MRI may be a crucial performance of carotid web.


Introduction
The carotid web is an intraluminal shelf-like projection within the lumen of the carotid bifurcation (1). It has been confirmed as a severe risk of recurrent ipsilateral cryptogenic ischemic stroke (2)(3)(4). Nevertheless, the prevalence of CaWs is indeterminate, range from 2.5% (5) -37% (6)(7)(8). CTA was usually selected as defining imaging modality in series of CaWs as a noninvasive method (1,7,8). The diagnostic feature "filling defect" and "a (thin) line or septum" was considered as important evidences to indicate carotid web (2,5,9). No clear data exist on compare the performance of each diagnostic features in CaWs.
Currently, no consensus on the performance of VW MRI with CaWs had established.
A MRI study using 2D FSE sequences and cine FSE sequences, black blood, 5 patients with CaWs diagnosed by 3D contrast-enhanced MR angiography was recorded with its performance (14). It had depicted the feature "protrusion" of the carotid web into the lumen and "a thin septum" dividing the lumen and "contrast stasis" after contrast enhancement. In our studies, a 3.0-T system (Prisma; Siemens, German) and 64 coil tunnel was used to acquire images using 3D FSE SPACE sequences (black blood). T1 weighted imaging (T1WI), T2 weighted imaging (T2WI) and T1 weighted imaging with fat suppression and contrast enhancement (T1WI-CE) was used to acquire carotid vessel wall imaging. And we summarized and concluded the performance of VW MRI with CaWs into six features based on those patient diagnosed by CTA.
Patients with carotid webs diagnosed by CTA was collected and aimed to summarize 5 their diagnostic features and to compare the diagnostic performance of each diagnostic features. In this studies, we performed this retrospective study to analyze each patient's image, and we hypothesized that each modality has its advantage in different diagnostic features.

Materials and Methods
All patients (n = 165, median age, 56.9 ± 13.3) who has done both neck CTA and VW MRI in our radiology department, Zhongnan hospital, performed from February 2017 to April 2019 were included. This research was approved by the ethics committee, Zhongnan Hospital of Wuhan University, Wuhan University. Those patient with stent-deployed or poor image quality that cannot be judged (n=12) were excluded. All CTA (n = 153, median age, 56.9 ± 13.3) images were read by two experienced radiologists. All carotids were evaluated in standard and oblique projections (thin cuts) in order to get better view of each carotid arteries.
Discrepancies were settled by consensus. Each patients diagnosed with CaWs will be recorded with their characteristics based on following diagnostic criterion.

CT angiography
A Siemens Somatom Definition 64 slice CT scanner (Siemens, German) or Philips Ingenuity CT 64 slice CT scanner (Phillips, North America) was use to acquire CTA images. Parameters: tube voltages: 140kv, tube current: 31mAs, thickness 1.5mm, with injection of 1.5ml/kg iopromide contrast material (Bayer Schering Pharma AG).

VW MRI
A 3.0-T system (Prisma; Siemens, German) was used to acquire images using 3D FSE SPACE sequences (black blood) and acquisition parameters are detailed in Table 1. 64-channel head neck coil was used to ensure the good image quality. Each person 6 included three sequences for about thirty minutes scanning. T1 weighted imaging (T1WI), T2 weighted imaging (T2WI) and T1 weighted imaging with fat suppression and contrast enhancement (T1WI-CE) after injection of 0.1 mmol/kg gadoliniumbased contrast material (Bayer Schering Pharma AG).

CT angiography
CaW is an intraluminal shelf-like projection within the lumen of the carotid bifurcation ( Figure 1) (6,8,9,15). Recent studies have used CTA as the defining imaging modality in series of CaWs. Three features of diagnostic criteria can be summarize into three points: (1) "A filling-defect ", other description: a shelf-like, a protrusion.
(2) " A thin line or a septum", dividing the lumen in axial.
(3) "Double-lumen sign", similar with dissection. b) Differential diagnosis It need to be distinguish with flow artifacts near the carotid bulb.

Reference standard
CTA was usually selected as defining imaging modality in this study. Each patients' diagnosis character of each modality were be recorded according to the abovementioned diagnostic criterion. Discrepant opinions were discussed and settled by consensus. In additions, the degree of stenosis of each patient with CaWs was recorded based on NASCET criteria (mild: 1-29%, low moderate: 30-50%, highmoderate: 50-69%, severe: 70-99%) (13,14). And any hyperdense focus with relative Hounsfield unit >1,000 was identified as vessel wall calcification.

Statistical Analysis
Interobserver agreement in CTA of detecting carotid webs was checked by using kappa statistics. Two-sided Cohen's Kappa (κ) coefficient was used to evaluate the inter-rater agreement for the two reader's diagnostic results.  Sensitivity, specificity, and area under the curve (AUC) of each diagnostic features at CTA and VW MRI were showed in Fig. 3 and 2D FSE sequences and cine FSE sequences had recorded carotid web with its 12 performance (14). It had supported the feature "protrusion" and "a thin septum" and "contrast stasis". But it did not provide quantitative measurements of carotid wall thickness and distensibility and did not make a systematic explanation of its performances on MR due to the small sample size. Our study using 3D FSE SPACE sequences (black blood), 3.0T MR, to acquire the carotid vessel wall imaging, and it can demonstrate the vessel wall in any direction and can summarize the signal on multi-sequences which can provide more evidences to distinguish atherosclerosis and dissection. Feature "Contrast stasis" means residual contrast-enhanced blood accumulation behind the carotid web on T1 weighted imaging with fat suppression and contrast enhancement (T1WI-CE), which can give evidence of for a proposed link between a morphological finding (carotid web) the likelihood of thrombus formation(1). It had proved good specificity (specificity, 93.6%, P < 0.001) but with lower sensitivity (sensitivity, 30.0%, P < 0.001), which means not all carotid web has contrast stasis or it could not been easily catched. Or, VW MRI may not be the best method to show it. Computational fluid dynamics may could provide good depictions with increased recirculation zones and regional increased wall shear stress metrics that are associated with disturbed flow (20).
CTA has good depiction on anastomotic stenosis of lumen, as our research had proved the feature "filling defect" has a good sensitivity. But it was weak in distinguish the component of a filling defect. VW MRI has great advantage to provide complementary and important information regarding vessel wall composition, wall mechanics, and blood flow (21). Especially, CaWs sometimes has the similar diagnostic performance with atherosclerotic plaques and carotid dissection (5,22). On the one hand, atherosclerotic plaques can also mimic a CaW because they typically occur at the carotid bifurcation and has the same feature " a 13 filling defect" (9). Both may appear as relatively focal endoluminal protrusions, therefore, vessel wall imaging is very important in making an accurate diagnosis.

Ethics approval and consent to participate
Informed consent is acquired.

Consent for publication
Agree.

Availability of data and material
Not applicable.

Competing interests
Not applicable.  Figure 1 Diagnostic criterion of CTA. Three key points: (1) "A filling-defect", a shelf-like, a protrusion Tables.docx