Clinical Evaluation of High-Resolution MRI Combined With DWI in Identifying Vulnerable Carotid Plaque

Background: High-resolution magnetic resonance imaging combined with diffusion weighted imaging is used to identify vulnerable plaques (VP) and their characteristic components, and apparent diffusion coefficient (ADC) correlation analysis with serum inflammatory markers to assess plaque vulnerability. Methods: In this study, 60 eligible patients were included, including 29 patients in VP group and 31 patients in non-VP group (N group). The average ADC value, serum inflammatory marker levels (high-sensitivity C-reactive protein, myeloperoxidase, and erythrocyte sedimentation rate) of the 2 groups were measured, and the characteristics of different plaque components and ADC levels of vascular wall in VP group were compared, to evaluate the correlation between serum inflammatory markers and the mean value of plaque ADC. Results: The results showed that the ADC mean value of the plaques in the VP group was significantly lower than that in the N group, and the levels of hypersensitive C-reactive protein and myeloperoxidase were correlated with the ADC mean value of the plaques. Conclusion: The ADC value of plaque measured by high-resolution magnetic resonance imaging combined with diffusion weighted imaging sequence can quantify the identification of VP and its characteristic components, reflect the inflammation of plaque to a certain extent, and thus prevent and treat stroke and other adverse outcomes more effectively.

C arotid atherosclerotic plaque is one of the main causes of ischemic stroke. 1 Plaque stenosis is not the only factor leading to ischemic stroke. Previous studies have shown that plaque composition and vulnerability are closely related to it. 2 Inflammation is considered to be an important factor involved in the development, progression and rupture of atherosclerotic plaques, leading to thrombosis. 3 Effective identification of active plaque inflammation is a key link for accurate identification of vulnerable plaques (VPs). In recent years, high-resolution magnetic resonance imaging (HR-MRI) of intracranial arteries has been proved to have higher reproducibility and clinical significance, and is an important technique for in vivo evaluation of intracranial arterial walls and plaques. 4,5 For extracranial plaques, carotid ultrasound, and computed tomography angiography are more dependent. Ultrasound, computed tomography angiography, and digital subtraction angiography are not sensitive to inflammation. Although positron emission tomography-computed tomography can reflect inflammation through uptake rate, it is expensive and has potential radiation damage to patients. In arthritis studies, MRI combined with diffusion weighted imaging (DWI) can show early changes in inflammation. 6,7 HR-MRI has been shown to be a sensitive tool for morphologic identification of atherosclerotic plaques, compatible with histology. 8 Skinner ad colleagues confirmed that HR-MRI combined with DWI or SWI functional imaging had higher spatial resolution and tissue component recognition for carotid artery VPs. [9][10][11] DWI is the only noninvasive imaging examination method that can detect molecular diffusion movement in vivo. It can accurately distinguish plaque components from normal vascular wall tissues by showing microscopic pathologic changes reflecting the degree of molecular diffusion movement restriction and combining with quantifiable apparent diffusion coefficient (ADC). However, there are few studies on this topic at present. Therefore, correlation analysis was conducted between serum inflammatory markers and imaging parameters in this study to explore the value of DWI in evaluating vulnerable carotid plaques.

Patients
Between September 2018 and March 2020, 60 patients in the Department of Neurology China-Japan Union Hospital of Jilin University were selected, including 43 males and 17 females, aged between 39 and 80 years. The present study was approved by the Ethics Committee of China-Japan Union hospital of Jilin University (approval no. 2020010812; Changchun, China). Patients who participated in this research had complete clinical data. Signed informed consents were obtained from the patients and/or the guardians.

Inclusion and Exclusion Criteria
Inclusion criteria: (i) the age of the patient fell within the range 18 to 80 years; (ii) carotid plaque had been identified by carotid ultrasound; (iii) patients have completed the baseline data examination; and (iv) the patients and their families were willing to sign informed consent forms. Exclusion criteria: (i) the patient has serious life-threatening diseases, such as acute cerebral infarction; acute coronary syndrome; hemorrhagic disease or bleeding tendency; epilepsy; dysfunction of heart, lung, liver and kidney; high fever, infectious diseases, electrolyte disorders; autoimmune diseases. (ii) Patients could not perform the examinations applied in this study, such as: metal prostheses or pacemaker implantation in vivo; claustrophobia; psychiatric disorders; drug and alcohol abuse; convulsions. (iii) HR-MRI quality score was ≤ 2, and subsequent image data analysis cannot be performed. (iv) Patients who have received intravenous thrombolysis and interventional therapy. (v) Patients with chronic inflammation were excluded.

Study Groups
According to the American College of Cardiology's modified the American Heart Association (AHA) classification standard for MRI, the plaque signal performance on high-resolution MRI was reviewed and divided into VP group (n = 29) and non-VP group (N group, n = 31). 12 Through the signal characteristics of each series IV, V, and VI plaque into the VP group, the rest of the type into N groups.

Cervical Vascular Examination
Mindray DC-8 color Doppler ultrasound imaging instrument (Mindray Medical International Limited) was used with the probe frequency of 9 Hz to scan the bilateral carotid arteries and vertebral arteries respectively, and the presence of carotid artery plaque was detected according to color Doppler flow imaging and echo characteristics.

Carotid HR-MRI Examination
Scanning was performed using a German Siemens (model Skyra) 3.0 T superconducting magnetic resonance scanner, and axial imaging was performed using a head-neck combined 20-channel coil and a carotid artery dedicated 4-channel coil. The carotid artery bifurcation as shown in the magnetic resonance angiography reconstructed automatically by 3-dimensional time-offlight (3D-TOF) was taken as the center to scan the upper and lower 3 cm of the carotid artery bifurcation. In this study, the

Image Postprocessing
Image quality of all sequences of plaque HR-MRI and DWI was evaluated by two experienced physicians according to the HR-MRI quality rating criteria, which divided image quality from low to high into 1 to 4, excluding images with score ≤ 2. 13 The AHA plaque typing method was followed, which was based on the modified plaque classification standard for nuclear magnetics in which plaque types I-II, III, VII, and VIII are classified as stable plaque (N group), and plaque types IV to V and VI are categorized as VP (VP group). 6,14 The main observation targets were lipid core (LC), fibrous cap (FC), calcification and intraplaque hemorrhage (IPH) of plaque.
With Siemens postprocessing workstation, the DWI sequence ADC diagram is reconstructed. With the measurement tool of DICOM browser RadiAnt DICOM Viewer 5.5.0, the ADC average of all the plaques and their characteristic components is independently measured. Determination of ADC mean value of plaque components: (i) In conventional HR-MRI sequence images, the area of signal change of target components (normal vessel wall, LC, FC, and IPH) was manually designated as the region of interest (ROI); (ii) manually copy the ROI of the target component into the ADC diagram at the corresponding level; (iii) measure ADC values of 3 times of target ROI at each level and take the average value. When 2 or more plaques are present in the same patient, the average value   ADC indicates apparent diffusion coefficient; ESR, erythrocyte sedimentation rate; hs-CRP, high-sensitivity C-reactive protein; MPO, myeloperoxidase. of each plaque is measured according to the above method and then taken as the average ADC value of the plaque of the patient. The ADC values of image quality and ROI above were analyzed by 2 radiologists with 15 and 13 years of experience in a blind way, and consensus was reached through discussion in case of disagreement.

Test Method
Serum levels of inflammatory markers hypersensitive C-reactive protein (hs-CRP), myeloperoxidase (MPO), and erythrocyte sedimentation rate (ESR) were measured in all subjects. The hs-CRP was determined by antu A2000 automatic immune analyzer, and the reagent was used by the hs-CRP detection kit (magnetic particle chemiluminescence). MPO was determined using the human MPO assay kit (enzyme-linked immunoassay). The ESR was determined by Westergren method.

Statistical Analysis
The obtained research data were statistically analyzed using SPSS software version 25.0 (IBM Corp.), and comparisons between the data groups were performed by using the χ 2 test. Bland-Altman method was used to evaluate the consistency of ADC mean values of plaques. Multiple comparisons were performed using single-factor analysis of variance (ANOVA) and least significant difference test. The correlation between serum inflammatory markers (hs-CPR, MPO, and ESR) and the mean value of plaque ADC was analyzed by the Pearson correlation analysis.

Comparison of Baseline Data
The number of cases with carotid plaque in VP group and N group were statistically analyzed. The general baseline data and risk factors of arterial plaque formation, including sex, age, smoking and alcohol history, hypertension history, diabetes history, coronary heart disease history, TG, CHOL, LDL, HDL, Hcy levels, were not significantly different.

ADC Comparisons Between Different Groups
The ADC of the overall plaque and the characteristic components of the vascular wall and VP measured by the two doctors were 1.3% and 3.2%, respectively, that were outside the consistency limit. Therefore, the ADC average values measured by the 2 physicians are consistent. Mean ADC results of each characteristic component of the vascular wall and plaque were performed using ANOVA. Homogeneity of variance test was as follows: P = 0.146 (P > 0.05), namely, homogeneity of variance of ADC mean values of vessel wall, lipid core, fibrous cap, and IPH. ANOVA univariate analysis showed that P = 0.00 (P < 0.05), there were statistically significant differences in ADC values between the vascular wall, lipid core, fibrous cap, and IPH as a whole. Multiple comparisons of least significant difference test showed that P = 0.00 (P < 0.05; Table 1). The mean value of ADC values in VP group and N group was (1.71 0.21 vs. 1.87 0.18×10 −3 mm 2 /s; P = 0.005). The serum levels of inflammatory markers (hs-CRP, MPO, and ESR) in the VP group were higher than that in the N group (   (Table 3).

Comparison Plaque Characteristics and Images Between Cervical Vascular Ultrasound and the DWI Sequences of HR-MRI
The assessment of VPs by cervical vascular ultrasound and ADC maps of HR-MRI combined with DWI is consistent. The size of the posterior wall plaque of the common carotid artery biforked was 0.97×0.24 cm, ADC was high signal (Fig. 1), the ADC mean value of each component was 1.79×10 −3 mm 2 /s, hs-CRP was 0.55 mg/dL, MPO was 18.14 mg/dL, and ESR was 6 mg/dL. A mixed echo plaque was observed in the left carotid sinus to the anterior wall of the initial segment of the internal carotid artery, with a size of about 2.2×0.30 cm. ADC was isosignal (Fig. 2). The ADC mean value of each component was 1.62×10 −3 mm 2 /s, hs-CRP 2.79 mg/dL, MPO 65.49 mg/dL, and ESR 8 mg/dL. Calcium spots were observed on the anterior wall of the carotid sinus, with a size of 1.2×0.25 cm. ADC was low signal (Fig. 3). The ADC mean value of each component was 1.51×10 −3 mm 2 /s, hs-CRP 1.29 mg/dL, MPO171.7 mg/dL, and ESR 6 mg/dL. HR-MRI combined with DWI can clearly show the characteristic components of vessel wall and plaque (lipid core, fibrous cap, and IPH), and can be quantitatively identified by measuring ADC value. The higher the ADC value, the lower the level of serum inflammatory markers. HR-MRI is more effective in identifying VPs and has obvious advantages compared with ultrasonography ( Figs. 1-3).

DISCUSSION
In the past, many scholars have done research on the degree of arterial stenosis and the shape of plaque, however, the occurrence of adverse outcome events such as cerebral infarction or TIA does not depend entirely on the volume of carotid plaque or the degree of carotid stenosis. Carotid VPs are present in 20% of cerebrovascular patients with carotid stenosis rate <50%, 15 and in 6% to 8% of patients with cerebrovascular disease without carotid stenosis. 16 Therefore, accurate identification of VPs and their compositional characteristics can better prevent the occurrence of adverse outcome events such as cerebral infarction or TIA.
On the basis of previous studies that applied HR-MRI-based DWI sequences to identify inflammatory lesions in other diseases, this study examined the association between ADC values of HR-MRI-based DWI sequences and plaque-related inflammatory markers to identify changes in plaque active inflammation. 17,18 The consistency of ADC values of plaques and their characteristic component ROI measured by the 2 physicians was analyzed using Bland-Altman method, which is the basis for quantitative analysis of imaging information to identify lesions, and is consistent with the conclusions of previous studies on quantitative identification of lesions using HR-MRI combined with DWI sequence. 19,20 HR-MRI combined with quantifiable functional FIGURE 2. Imaging findings of plaques with thrombosis. Carotid arterial ultrasound: The arterial wall is not smooth, and part of the wall is thickened. Left: A mixed echogenic plaque is seen on the anterior and posterior walls of the carotid sinus to the origin of the internal carotid artery. Right: The initial segment of the common carotid artery was occluded about 1 cm after the segmental lumen was occluded, and the lumen was filled with parenchymal heterogeneous hypoechoic thrombus without blood flow, as shown by the circle (A). A hypoechoic plaque was seen from the carotid sinus to the posterior wall of the initial segment of the internal carotid artery; a hypoechoic plaque was seen on the anterior wall of the carotid sinus. High-resolution magnetic resonance imaging of the neck plaque showed that the right common carotid artery, external carotid artery, and proximal internal carotid artery were not clearly displayed, as shown (B). In the right common carotid artery, a strip of equal or slightly shorter T1 and equal or slightly longer T2 signal shadows were seen, involving a length of about 9.1 cm, and DWI showed slightly high signal; bilateral internal carotid arteries had eccentric thickening of the lumen at the beginning, local It can be seen that the arc is slightly shorter T1 and slightly longer T2 signal shadow, and DWI is isointense, as shown by the arrows (C-F). DWI indicates diffusion weighted imaging.
imaging has improved the deficiencies of traditional MRI in the identification of microlesion structures. HR-MRI showed that ADC values of the characteristic components of VPs (vascular wall, lipid, FC, and IPH) were statistically different (Table 1). In the DWI sequence of MRI, the weighted sum of quantized MRI signals from inside and outside the cells was performed, and the results were expressed as ADC. ADC can deduce the microenvironment of cells by measuring the limitation of different microstructures on the diffusion of water molecules. 21 This study found increased serum inflammatory markers (hs-CRP, MPO) in response to active plaque inflammation in a subset of plaques not defined as VP by conventional HR-MRI sequences and a correlation between serum hs-CRP, MPO, and ADC values (Figs. 1-3), which may provide new insights into the noninvasive diagnosis of VP.
The VP group had higher levels of serum inflammatory markers (hs-CRP, MPO) than the N group, the VP group had lower mean ADC values than the N group, and the levels of serum inflammatory markers (hs-CRP, MPO) were inversely correlated with the ADC values of the plaques, whereas hs-CRP, MPO were serum markers reflecting active inflammation of the plaques. Therefore, the ADC values could be applied to reflect the inflammation of the plaques, and complement the application of conventional HR-MRI for assessment of plaque vulnerability.
It is also demonstrated that hs-CRP and MPO participate in inflammatory response and are closely related to the formation of vulnerable cervical plaques. Although some evidence 22 has shown that ESR has a certain diagnostic effect on unstable plaques, in this study, there was no significant difference between the VP group and the N group, indicating that the diagnosis of ESR on VPs remains to be explored.
The present study has certain limitations. First, although motion detection technology was added in this study to improve the accuracy of ADC images, data bias caused by image resolution and motion pseudoerror cannot be excluded. Secondly, the sample size of the present case was small, and there may be a certain bias associated with some of the conclusions. Finally, the lack of a gold standard of pathology as a verification tool in this study is a limitation of current research.

CONCLUSION
The DWI sequence and its ADC of HR-MRI can be used to evaluate the carotid artery VPs and their characteristic components in a convenient and noninvasive manner, reflecting that plaque active inflammation is superior to other imaging examinations and has unique advantages in the prevention and treatment of ischemic stroke.