Circle of Willis (CoW) is an important source of collateral blood flow to maintain adequate cerebral perfusion. But CoW has large anatomical variability and the prevalence of incomplete CoW is high in the general population [1]. When the cerebral artery occluded, incomplete CoW may not provide satisfactory collateral blood supply, which may influence the severity of ischemia stroke [2]. Thus, assessment of the CoW is significant in clinical practice. However, traditional method to evaluate the CoW majorly focused on the luminal conditions. Recently, black-blood MRI has been proposed to imaging cerebral artery vessel wall with high spatial resolution [3, 4], providing a good opportunity to further assess the CoW in not only luminal conditions but also the vessel wall atherosclerosis. Thus, the purpose of this study is to assess CoW using bright-blood and black-blood MRI and investigate its relationship with the severity of ischemic stroke.

Population: In this study, Thirty-seven patients (21 males; mean age=56±8.4 years) with ischemic stroke occurred within 3 months were recruited. The study protocol was approved by local institutional review board prior to the initiation of this study. The written consent forms were obtained from all subjects before data acquisition. All patient were scanned more than 2 weeks after stroke.

Image acquisition: The intracranial arteries were scanned for each case using a Philips 3T scanner (Achieva TX, Philips Medical System, Best, The Netherlands) with a 32-Head coil. The brain of each subject was scanned with T2W FLAIR for ischemic stroke identification. 3D TOF, 3D SNAP [3] and T1W VISTA (black-blood) [4] sequences were also scanned to obtain the bright-blood and black-blood intracranial artery images, including CoW. The imaging parameters for all the sequences were listed in Table 1.

Data Analysis: First, an experience radiologist reviewed the T2W FLAIR images to estimate the final ischemic infarction size for each case. The ischemic infarction size was divided into four levels to represent different severity of ischemic infarction [5]: level 1, no ischemic infarction ; level 2(Fig 1F), ischemic infarction size smaller than 1cm; level 3, ischemic infarction size larger than 1cm, but smaller than 3cm; level 4, ischemic infarction size larger than 3cm (Fig 1B). The CoW integrity was identified on 3D TOF, and classified into two kinds: complete or incomplete CoW (Fig 1A, 1E). The SNAP and VISTA sequences were used to identify atherosclerotic plaque on the vessel wall of CoW. Thicken vessel wall observed both on SNAP and VISTA images were considered as a plaque, as shown in Fig 1C, 1D. The review for ischemic infarction, integrity of CoW, and atherosclerotic plaque on CoW, were blinded to each other and patient information. The incidence rate of integrity and atherosclerosis of CoW were reported for each ischemic infarction size level. Chi-square test was used to analysis the association between ischemic infarction size level and integrity & atherosclerosis of CoW.

As shown in Table 2, only 5 (13.5%) patients in this population have complete CoW, all of which have atherosclerosis in CoW and relative large sichemic infarction size (level 3&4). In this population, most patients (32, 86.5%) have incomplete CoW, of which, most patients (26 of 32, 81.25%) have atherosclerosis in CoW. Among the patients with incomplete and atherosclerotic CoW, 11 of 26 (42.3%%) patients suffered from large infarction (ischemic infarction size level 4), and 5 of 26 (19.2%) patients has level 3 infarction. The chi-square test revealed that there was significant association between ischemic infarction size level and integrity & atherosclerosis of CoW (p=0.037).In this study, we found that most stroke patients have incomplete CoW, and patients with atherosclerotic CoW tends to have larger ischemic infraction. More importantly, ischemic infarction size in stroke patients was significantly associated with integrity & atherosclerosis of CoW, suggesting that not only the integrity, but also the atherosclerosis of CoW may be a risk factor to predict the severity of stroke. Thus, the high-resolution black-blood vessel wall MRI, such as T1W VISTA and SNAP, is a good tool for CoW evaluation, and may be a good predictor for stroke severity.