High-Degree Middle Cerebral Artery Stenosis

Intracranial atherosclerotic stenosis (ICAS) is a major cause of ischemic stroke worldwide and the middle cerebral artery (MCA) is the most common site of ICAS among Asian populations [1, 2]. Patients with high-grade (≥70%) MCA stenosis face a high risk of stroke occurrence and recurrence, despite best medical treatment [3]. The stenting and aggressive medical management for preventing recurrent stroke in intracranial arterial stenosis (SAMMPRIS) and the Vitesse intracranial stent study for ischemic stroke therapy (VISSIT) trials showed an increased risk of stroke with medical treatment plus percutaneous transluminal angioplasty and stenting (PTAS) compared with medical treatment alone, among patients with high-grade symptomatic ICAS, mostly due to the rough patient selection criteria (based solely on the severity of luminal narrowing) and high perioperative complications [4‐6].

Recent registry studies indicated promising safety and efficacy of PTAS, if performed in appropriately selected patients in high-volume centers [7]. It is still challenging to determine the benefit and risk of PTAS in individual patients with symptomatic ICAS, which are probably closely related to the stroke mechanisms.

Lenticulostriate arteries (LSAs), commonly originating from the MCA M1 segment, are particularly susceptible to MCA stroke often with severe motor weakness as they supply blood to the basal ganglia and parts of the internal capsule [8]. The most common periprocedural complication of PTAS is occlusion of perforating vessel(s) near the site of stenosis, when the plaque is possibly snow-ploughed over the perforator(s) [9, 10]. In the post hoc analysis of the SAMMPRIS trial, the rate of 30-day ischemic stroke after PTAS was decreased from 14.7% to 9.4%, when excluding patients with perforator infarction [11]. Therefore, understanding the spatial relationships between LSAs and MCA plaque is important in understanding the stroke mechanisms, guiding clinical decisions regarding PTAS versus medical treatment, and determining risks of periprocedural complications of PTAS; however, so far evidence regarding their relationships is mostly from post-mortem studies rather than in vivo studies, which was investigated in the current study with coregistration of magnetic resonance imaging (MRI) and three-dimensional digital subtraction angiography (3D DSA), along with their relationships with patterns of lenticulostriate infarctions, among patients with atherosclerotic stenosis of MCA.

Fusion or coregistration of 2 stand-alone imaging modalities have been used in pretreatment or posttreatment evaluations or cerebrovascular neuronavigation in various intracranial conditions, such as arteriovenous malformations (AVMs) and dissecting aneurysms [12‐17]. Previous studies showed the efficacy of DSA-MR fusion to understand the anatomical relationships between small arteries and surrounding structures [15, 18]. With DSA as the gold standard for evaluation of intracranial arteries, and MRI using the fast spin-echo techniques with isotropic 3D imaging (FSE-Cube) providing crucial and complementary anatomical information about MCA plaques, the fusion of the two could illustrate the morphologic characteristics of LSAs and the spatial relationships of LSA orifices with MCA plaques.

Consecutive ischemic stroke or transient ischemic attack (TIA) patients with severe MCA stenosis admitted to the department of neurology at a teaching hospital from September 2015 to September 2018, who received MRI (with diffusion-weighted imaging, DWI, and FSE-Cube) and 3D DSA within 2 weeks after symptom onset, were retrospectively screened and analyzed. The study was approved by the local institutional review board for retrospective data collection and review. Subjects fulfilling the following criteria were analyzed in the current study: 1) aged 18–80 years, 2) having focal stenosis ≥70% in MCA-M1 confirmed by DSA and 3) having 1 or more vascular risk factors including hypertension, diabetes mellitus, hyperlipidemia, obesity (body mass index ≥27) and smoking. Patients were excluded if they had any of the following conditions: 1) contraindications for MRI and DSA, such as allergy to contrast agents, ferromagnetic implants or claustrophobia, 2) insufficient image quality for analysis, 3) non-atherosclerotic arterial stenosis, such as vasculitis, dissection or moyamoya disease, 4) coexistent ipsilateral extracranial or intracranial internal carotid artery stenosis >50%, 5) bilateral MCA-M1 stenosis ≥70% and 6) evidence of possible cardioembolism.

Demographic features and vascular risk factors were recorded. We divided the patients into two groups, according to whether the severe MCA stenosis was relevant to the index ischemic stroke or TIA: culprit MCA stenosis and non-culprit MCA stenosis groups. Relevance of MCA stenosis with the index ischemic event was determined by experienced neurologists; uncertainties were resolved by consulting a senior neurologist.

In this study, we used the 3D DSA-MRI fusion technique to investigate the morphological features of MCA trunk, MCA plaque and LSAs in patients with high-grade MCA stenosis responsible or not responsible for an index ischemic stroke or TIA. Fused 3D DSA and MRI images accurately depicted the miniature structure and spatial relationships between these arteries/lesions. We found no difference in the degree of MCA stenosis or the numbers of LSA stem(s) and branches between those with culprit and non-culprit MCA stenosis; however, patients with culprit MCA stenosis had more LSAs originating from the stenotic MCA segment and more presence of MCA plaques growing over the orifice of the LSA(s), than those with non-culprit MCA stenosis. Moreover, we proposed a classification method for the spatial relationships between MCA-M1 trunk, the orifice of LSAs and MCA plaques, which can be used in further studies to better understand the role of such relationships in understanding the mechanisms of ischemic stroke in the presence of MCA stenosis.

The 3D DSA-MRI fusion technology shows excellent diagnostic accuracy and therapeutic guidance for neurovascular diseases as indicated in previous studies [12‐15, 17]. In the current study, LSAs seemed to originate from MCA perpendicularly in MRI in some cases rather than in a zigzag course as revealed in the fusion images [20]. This would lead to misinterpretation of the true LSA orifice and its spatial relationship with MCA plaques, which is a limitation of MRI. The addition of 3D DSA could improve the accuracy in visualization of the extremely complex and numerous small vessels in the lateral fissure area, such as LSAs, recurrent artery of Heubner and parallel veins that frequently overlap and are difficult to delineate in MRI. Therefore, the 3D DSA-MRI fusion technology is promising in clinical practice and research scenarios when it is necessary to visualize these small vessels.

The path of LSAs shown in 3D DSA-MRI fusion images was consistent with previously published microanatomy findings, with LSAs usually turning medially in contact with MCA M1 after originating from MCA, and entering into basal ganglia in a curved shape afterward [16]. The total number of LSAs (median, 4) in our study was also similar to that reported in pathological studies, such as a study by Vuk Djulejic (mean 4) [21]. We also assessed features of LSA stems because the origin of the LSAs was directly related to features of MCA M1 and possibly MCA plaques. Previous pathological studies found that LSAs often arise by common stems [22] but the number of LSA stems has rarely been documented. We reported a smaller number of LSA stems (1 on average) in the current study, compared with previous imaging studies, including studies using 7.0 T time-of-flight MRA (mean, 4) and 3.0 T MRI (mean, 4) [23]. The difference may be partly attributed to the limited accuracy of MRI in illustrating LSA orifices as mentioned above, as the apparent “orifice” of LSAs might in fact be the intersections between LSA branches and MCA, which could lead to overestimation of the numbers of LSA orifices and LSA stems. It is also possible that LSAs might have been occluded by MCA plaques in the current cohort of patients with high-grade MCA stenosis. Further studies in patients with no MCA plaque or stenosis could verify such findings.

A more important finding of the current study was the differences in spatial relationships of LSA orifices with MCA trunk and plaques, between patients with the culprit and non-culprit MCA stenoses. Patients with culprit MCA stenosis had more LSA(s) originating from the stenotic MCA segment and more LSA(s) with MCA plaques growing over LSA(s) orifice, than those with non-culprit MCA stenosis. This may partly explain the mechanisms of ischemic stroke in patients with culprit MCA stenosis: first, LSAs derived from severely stenotic MCA segment maybe susceptible to hypoperfusion; second, MCA plaques partly or completely occluding the orifice of LSAs could not only lead to superficial infarction, but also deeper perforator infarctions [24] with microembolism.

The term branch atheromatous disease (BAD) had been proposed as a pathological diagnosis with atheromatous plaques at the orifice of LSAs, with the plaque from MCA or in situ LSAs [25].

However, BAD had not been visually observed in previous studies that mainly concentrated on atherosclerotic plaque characteristics and infarction features [26‐28]. The 3D DSA-MRI fusion technique, however, may help to accurately subclassify the pathogenic mechanism of perforator stroke based on the relationship between plaque location and the orifice of LSAs, when BAD was equivalent to type 1 and type 3 models in the proposed classification method (Fig. 5a,c). Further studies with this classification method could provide better understanding of BAD and its association with small subcortical infarct(s) [29].

The classification method proposed in this study for the spatial relationships between MCA-M1 trunk, the orifice of LSAs and MCA plaques may also provide valuable information in selecting appropriate patients with MCA stenosis for endovascular treatment. In patients with high-grade symptomatic MCA stenosis when perforator occlusion is a probable stroke mechanism, endovascular treatment may lead to no benefit but unnecessary risks for perioperative complications [4‐6]. For instance, patients with type 1 and 3 lesions might not be suitable for PTAS treatment, while type 2 and 4 lesions may benefit from PTAS with the restoration of cerebral perfusion (Fig. 5).

The current study was among the first to describe the relationships between lenticulostriate infarcts, distribution of LSAs and MCA plaques, using the 3D DSA-MRI fusion technique. This may provide insights for classification of stroke mechanisms and decision in treatment strategies in patients with symptomatic MCA stenosis; however, the study had limitations. First, there might be image distortion, such as pin cushion distortion and S distortion, in fused 3D DSA-MRI images [19] and the fusion image quality may be degraded by some artifacts from manual co-registration [30]. Second, the sample size of the study was relatively small from a single center retrospective analysis, while further longitudinal studies with larger cohorts are needed to extend applications of the 3D DSA-MRI fusion technique and explore clinical implications of our findings.

The current study demonstrated the feasibility of using 3D DSA-MRI fusion technique in visualization of the miniature structural features of LSAs and MCA plaques and their spatial relationships. In patients with MCA stenosis, LSA(s) originating from the stenotic MCA segment, or LSAs with plaque growing over from MCA plaques, are more likely to be associated with infarctions in distal areas. Classification of the types of spatial relationships between MCA-M1 trunk/plaques and the orifice of LSAs, such as the classification method proposed in the current study, could help understand the stroke mechanisms and possibly guide patient selection for endovascular treatment in patients with symptomatic MCA stenosis. Further studies are warranted to verify our findings.