This is a prospective registry study that has been approved by the ethics committee of our hospital. Written informed consent was obtained from the patients or their relatives.
Enrollment of Patients
Patients suspected to have symptomatic intracranial atherosclerotic stenosis (ICAS) at admission were enrolled. All patients received thorough medical evaluations to determine the underlying cause using a range of techniques including carotid artery duplex scan, transcranial Doppler, echocardiography, electrocardiography, computer tomography (CT), magnetic resonance imaging (MRI), CT angiography (CTA), magnetic resonance angiography (MRA), and digital subtraction angiography (DSA). Patients were included in the BA study according to the following criteria: 1) age ≥18 years; 2) ischemic stroke or transient ischemic attack (TIA) in the BA regions within 90 days following enrollment; 3) basilar artery stenosis ≥70% and without coexistent ≥50% ipsilateral extracranial vertebral artery stenosis; 4) absence of potential sources of cardioaortic embolism based on the modified Trial of ORG 10172 in Acute Stroke Treatment (TOAST) classification [14]; 5) one or more risk factors for atherosclerosis; 6) all the patients received DSA examination. Other risk factors were recorded for comorbidities including hypertension, dyslipidemia, diabetes, smoking, and obesity.
Patients with the following conditions were excluded: 1) non-atherosclerotic cerebral vasculopathies such as vasculitis and artery dissection, diagnosed by comprehensive laboratory examinations (such as erythrocyte sedimentation rate or C-reactive protein elevation, antinuclear antibody, or antiphospholipid antibody positivity), vascular imaging, and clinical evaluation. 2) contraindication to MR examination, instable clinical state precluding MR examination.
HRMRI acquisition and analysis
All HRMRI studies were performed on a DISCOVERY MR750 3.0T (GE Healthcare, Waukesha, WI, USA) or a 3T Trio MRI scanner (Siemens Healthcare, Ehrlangen, Germany). More details can be found in the study protocol (see Supplemental Table 1). Image reconstruction was conducted using the Reformate tool in Advantage Workstation 4.5 (GE Healthcare) and 3D multiple planer reconstruction tool in Siemens workstation. MR images were then processed for all the identified plaques using commercially available software (VesselMass; Leiden University Medical Center, Leiden, The Netherlands).
A culprit plaque was defined as the single lesion present in the artery supplying the infarct zone, or as the most severe stenotic lesion when multiple plaques were present in the supplying artery [15-16].
The arterial remodeling index (RI) was calculated as the ratio of outer wall area (OWA) at the site of maximal lumen narrowing to that at the reference site (RI=OWA lesion/OWA reference) [17]. The reference site was selected based on the Warfarin-Aspirin Symptomatic Intracranial Disease (WASID) trial method [18]. Three remodeling categories have previously been described, with RI ≥1.05 defined as positive remodeling, 0.95 <RI< 1.05 as intermediate remodeling, and RI ≤0.95 as negative remodeling. Plaque distributions were dichotomized into diffuse and non-diffuse patterns at culprit lesion. The anatomical location of the plaque was recorded as ventral, dorsal, left, and right quadrants [19]. Plaques spreading across four quadrants were defined as diffuse and that involving ≤3 quadrants were defined as non-diffuse. Intraplaque hemorrhage (IPH) was defined as a signal intensity greater than 150% of T1 signal of adjacent muscle [20]. As for plaque enhancement, non-enhancement was defined as similar to, or less than, that of normal intracranial arterial walls nearby, while enhancement meant signal intensity greater than non-enhancement and less than, or greater than, that of the pituitary infundibulum [15].
We adopted the same principle when interpreting HRMRI imaging for arterial remodeling and vessel wall features as we published before with small intra-observer and inter-observer variability [10,17,21]. The intra- and inter-observer variability of the two HRMRI scanners and identified vessel wall features were good to excellent (weighted k =0.82, 95% CI: 0.46, 1.00 and 0.83, 95% CI: 0.41-1.00, respectively).
Definition of BA configurations
BA anatomy was defined as having an either complete or incomplete configuration (see Figure 1). The patients with normal bilateral vertebral arteries and posterior cerebral arteries were categorized as presenting complete BA configuration (see figure 1A-C). Patients with fPCA (see figure 1D-F) and/or dysplasia in one VA (see figure 1G-I) were identified by DSA and/or CTA, MRA as having incomplete BA configuration. The presence of posterior communicating arteries was also recorded. Hypoplasia was defined as VA having a diameter <2 mm, ending in the posterior inferior cerebellar artery (PICA), or having a lumen diameter more than 50% difference [4-5]. fPCA is a common anatomic variation that was defined as a posterior cerebral artery originating from the internal carotid artery, in the absence of the P1 segment of PCA or presence of PCA P1 segment dysplasia [6-7]. Two neurologists (Z.Q.X. and N.M) reviewed the DSA images independently and discrepancies were resolved by consensus.
Statistical analysis
Continuous variables were presented as means ± SD or median with interquartile range. Categorical variables were presented as percentages. All baseline characteristics, plaque enhancement, intraplaque hemorrhage, arterial remodeling patterns, and plaque distribution were compared with χ2 test for categorical variables and one-way analysis of variance or the Kruskal-Wallis test for continuous variables between the complete and incomplete configuration groups. The analyses were performed using SPSS 23.0 statistical software (IBM, Chicago, IL, USA). A two-tailed p value less than 0.05 was considered statistically significant.