Integrated head and neck imaging of symptomatic patients with stroke using simultaneous non-contrast cardiovascular magnetic resonance angiography and intraplaque hemorrhage imaging as compared with digital subtraction angiography

Intracranial and carotid atherosclerotic disease are the major causes of ischemic stroke worldwide, especially in Asia [1]. The risk assessment of ischemic stroke is based on the percentage of stenotic intracranial and carotid arteries, according to medical imaging findings [2, 3]. Digital subtraction angiography (DSA) is the gold standard for evaluating arterial stenoses. However, it has many disadvantages, including being invasive, costly, and complicated to perform. Cardiovascular magnetic resonance (CMR) high-resolution vessel wall imaging (HR-VWI), which provides a non-invasive and clear visualization of vessel lumens and walls, has been used increasingly in the study of atherosclerotic plaques. HR-VWI has been reported to provide significant data for the diagnosis, clinical treatment, and prognosis of patients with stroke [4‐6]. As a new CMR-based technique, simultaneous non-contrast angiography and intraplaque hemorrhage (SNAP) imaging can show the presence of both lumen and intraplaque hemorrhage (IPH) simultaneously with a wide field-of-view (FOV) that covers large vessels in both the head and the neck [7]. Recent studies have assessed the ability of HR-VWI and SNAP to characterize atherosclerotic plaque compositions, stenoses, and their association with ischemic stroke. The imaging functions of these modalities were also compared with those of conventional vascular imaging techniques [8‐12]. Previous research suggested that SNAP was efficient at identifying IPH and had satisfactory agreement with histopathology [8]. Several current guidelines recommend HR-VWI as the most appropriate diagnostic method to assess intracranial and extracranial atherosclerotic plaques, especially vulnerable plaques [13], and SNAP is suggested as a complement to identify vulnerable plaque components, mainly for IPH [14]. Compared with SNAP, HR-VWI examinations including multi-sequence imaging, are more time-consuming. If the diagnostic performance of SNAP and HR-VWI is similar to DSA in measuring the stenosis severity, SNAP could replace HR-VWI in some clinical circumstances and might improve diagnostic efficacy.

A growing body of evidence suggests that the composition and burden of plaques are important for stroke assessment and prognosis [13, 15]. IPH, as a plaque component, is particularly important for the prediction of plaque progression [16]. Previous literature demonstrate that the early detection of IPH enables clinicians to determine whether patients can benefit from carotid endarterectomy or stenting, which could eventually improve the prognosis of patients with stroke [17]. Exploring IPH-associated clinical factors via routine clinical biochemical examinations may remind clinicians to pay attention to the existence of IPH before an imaging examination, leading to diagnostic efficacy improvements.

There are few studies comparing the performance of SNAP and DSA in assessing arterial stenosis in the head and neck vessels, as well as the association between IPH and clinical factors. The present study aimed to compare the performance of SNAP with DSA in measuring luminal stenosis of the head and neck arteries. It also explored the value of SNAP in identifying IPH and the clinical factors associated with IPH to provide an effective imaging basis for improving diagnostic and therapeutic efficacy.

Although the primary indicator of ischemic stroke risk is the rate of stenosis, plaque components are becoming increasingly recognized as important in assessing stroke risk, as shown in the current study. IPH is an important risk factor for stroke compared with other plaque components [24]. IPH is common in asymptomatic and symptomatic patients with stroke, while symptomatic patients are at a greater risk for IPH than asymptomatic patients [25]. In the present study, we studied the symptomatic patients with stroke and investigated whether SNAP and HR-VWI were well-aligned with DSA for measuring the stenosis and diameter of both intracranial and carotid arteries. We also studied if SNAP was similar to HR-VWI in being able to detect IPH. The results showed that SNAP was more consistent with DSA at measuring luminal stenosis than HR-VWI. When measuring general segments, SNAP was also better than or similar to HR-VWI in the measurement of intracranial vascular segments. Additionally, SNAP had good inter-group consistency for the intracranial and carotid arteries and a good consistency with HR-VWI in identifying IPH. Based on these results, we analyzed IPH-related clinical factors. The main clinical factors associated with IPH were HDL, glucose, and cystatin C.

SNAP performance in detecting luminal stenosis has been studied previously [8‐12]. Wang et al. reported that SNAP had a good agreement with time-of-flight (TOF)-CMR angiography (TOF-CMRA) in measuring carotid artery lumens [8]. Shu et al. showed that, compared with CE-CMRA, SNAP was equally capable of measuring the carotid artery stenosis [9]. Several scholars have also suggested that SNAP had good consistency with TOF-MRA, magnetization prepared-rapid gradient echo (MPRAGE), and histology in characterizing plaque components, especially for IPH detection [10, 11]. SNAP was also shown to be consistent with HR-VWI in characterizing large intracranial vessels [12], while HR-VWI was found to have acceptable agreement with DSA in measuring the diameter of each carotid artery segment [7]. However, few studies have concentrated on the integrated imaging of intracranial and carotid vessel walls [26, 27], and no study has comprehensively assessed the imaging performance of SNAP and DSA on head and neck vascular walls. Our results showed that SNAP and HR-VWI were consistent with DSA in each intracranial and carotid vessel segment, and the results were reproducible and repeatable. SNAP is a non-contrast-enhanced integrated head and neck imaging technique that can be used for a wider range of individuals than conventional high-resolution imaging and DSA, the gold standard imaging technique, especially for contrast-sensitive patients and critically ill patients with poor liver and kidney functions.

To the best of our knowledge, CMR can evaluate IPH easier and faster than other plaque components. CMR is also more amenable to in-clinic evaluation. SNAP and HR-VWI were all found to be in good agreement with pathologic findings in the identifying IPH in the previous literatures [11, 28]. In the current study, we found that SNAP and HR-VWI had good consistency in the identifying of IPH, and no significant difference in the number of IPH and non-IPH patients among those with different degrees of stenosis, which could be related to the relatively small sample size or a potentially complicated relationship between stenosis and various plaque components [29, 30]. Additionally, mild stenosis might be partly caused by plaque remodeling [31]. Therefore, the number of IPH lesions was observed to increase as the degree of stenosis increased; however, no significant correlation was found between stenosis and the presence of IPH lesions, indicating the study of plaque remodeling is needed. But, overall, we found that as the degree of stenosis increased, the presence of IPH also increased. Our study included the assessment of both carotid and intracranial arteries, and we came to a similar conclusion as those of a previous study [32]. That study demonstrated an increase in the degree of carotid artery stenosis independently associated with the presence of IPH. IPH was detected in both carotid and intracranial vessels. Moreover, IPH was detected in the major intracranial vessels more often than in the carotid artery. Therefore, the integration of head and neck imaging can help detect and localize more IPH lesions, which is helpful for diagnosing etiologies, giving treatment guidance, and providing prognostic evaluations for symptomatic patients with stroke. SNAP is a one-stop solution for measuring narrowness and identifying IPH and has a shorter scan time than HR-VWI. However, it is worth noting that the bright signal in the T1w intracranial artery images in our study was defined as IPH with no histological evidence provided, which is a limitation of the study. On the other hand, this image interpretation is similar to previous radiological studies on intracranial IPH [33, 34] and has been supported by some specimen studies [28].

A previous study found that IPH in carotid plaques using MPRAGE imaging was associated with age, sex, degree of stenosis, and hyperlipidemia [35], which was, not observed in our study. This may be related to the cohort characteristics where patients with primarily moderate to severe arterial stenosis were included; some also had partial occlusions. In addition, rather than considering carotid plaques, we included all intracranial and carotid arterial plaques that could have influenced the analytic results, owing to the high incidence of intracranial plaques in Chinese patients and the high incidence of extracranial plaques in European and American patients [36]. A comprehensive assessment of the presence of plaques enables us to identify the etiology of stroke better.

The specific relationship between cystatin C and IPH has not been previously reported. Cystatin C is an inhibitor of elastase that can affect the extracellular matrix remodeling of atherogenesis. Although cystatin C levels could also indicate renal dysfunction, a number of scholars have reported the association of higher cystatin C levels,with an increased risk of cardiovascular-related death [37], a higher degree of intracranial and extracranial artery stenosis, and the presence of vulnerable plaques in patients with acute ischemic stroke [38, 39]. Increased serum cystatin C levels have also been shown to be associated with subclinical atherosclerosis and recurrent stroke [40, 41]. Therefore, increased serum cystatin C levels could be related to the progression of arteriosclerosis plaque. In the present study, we showed that patients with IPH also tended to have significantly higher serum cystatin C levels than those without IPH. But our study had a small sample size, we still need further research in the future.

Diabetes is known to play an important role in stroke. The Rotterdam Study showed that IPH is associated with increased serum insulin levels but not with increased glucose levels [42]. Previous studies have also found that diabetes was highly associated with lipid-rich necrotic cores (LRNC) of plaques and that the volume ratio of LRNC > 22.0% in carotid plaques could be an independent risk factor for acute cerebral infarction [43]. In our study, although the glucose levels were significantly different between the IPH and non-IPH groups, diabetes was not found to be significantly associated with IPH. We speculate that the varying glucose levels were a result of the varying serum insulin levels.

Hypertension is an important risk factor for clinical cardiovascular events. A study [44] on carotid arteries found that increases in both systolic and diasotlic blood pressure were associated with a range of plaque components, including IPH. Hypertension also affected vessel wall volume and fibrous cap. Hypertension may contribute to fibrous cap rupture and IPH, but we did not observe a similar result, perhaps due to the small sample size. Low diastolic blood pressure has been found to be associated with IPH in patients with asymptomatic atherosclerosis [45]; however, the underlying mechanism needs further study in the future.

This study has several limitations. First, our study had a small sample size and was a single-center study. Second, there might have been some bias in our population selection, which primarily selected for symptomatic patients with moderate to severe stroke. Third, we did not have histopathologic results of plaque as the gold standard to identify IPH, especially in the intracranial arteries. Finally, a small percentage of the clinical data were missing due to the retrospective nature of this study; we attempted to decrease the influence of missing data on the conclusions as much as possible. Therefore, the conclusions should be considered carefully when generalizing the findings.

Among patients with recent symptomatic stroke, CMR SNAP is consistent with DSA in measuring the diameter of each intracranial and carotid arterial segment and stenotic atherosclerotic lesions. SNAP was also demonstrated a potential alternative to HR-VWI in identifying stenosis and IPH.