INTRODUCTION
Moyamoya disease (MMD) is a unique cerebrovascular disorder characterized by chronic progressive occlusion of the terminal portion of the internal carotid artery (ICA) and is characterized by a hazy network of basal perforating arteries [19]. Leptomeningeal collaterals play an important role in the preservation of cerebral perfusion in patients with MMD [8]. Diffuse leptomeningeal high signal intensity (SI) along the cortical sulci on contrast-enhanced T1 magnetic resonance imaging (CEMR) and fluid-attenuated inversion-recovery (FLAIR) imaging has been reported in patients with MMD [9,14]. This imaging characteristic is termed the ivy sign as it resembles ivy creeping on stone.
While gadolinium (Gd)-enhanced three dimensional (3D) high-resolution vessel wall magnetic resonance imaging (VWMR) has recently proven to be valuable in understanding the pathogenic mechanisms of MMD, its application has primarily been directed towards proximal segments of the major intracranial arteries [16,17]. However, our recent observations have demonstrated a distinctive linear enhancement along the cortical sulci on Gd-enhanced VWMR in patients with MMD, introducing the concept of the “VWMR ivy sign” (VIS). In this preliminary report, we investigate VIS in 18 patients and propose its clinical implications.
MATERIALS AND METHODS
This study was approved by the Institutional Review Board of Kyung Hee University Hospital (IRB approval No. 2024-01-045).
Study participants
A total of 18 patients with MMD were included in the study (Table 1). Demographic data and clinical presentations were obtained from medical records, and imaging data were analyzed. Patients with MMD were categorized into ischemic, hemorrhagic, and non-stroke groups according to the stroke presentation. The ischemic group was defined by the presentation of neurological symptoms with restricted diffusion lesions on diffusion-weighted imaging and apparent diffusion coefficient mapping at the time of stroke. The hemorrhagic group included patients with intracranial or intraventricular hemorrhage on computed tomography (CT) and susceptibility-weighted imaging at the time of stroke presentation. The non-stroke group was defined as patients with no symptoms or no hemorrhagic or ischemic lesions on CT or magnetic resonance imaging (MRI).
VWMR evaluation
The primary point of interest in this study was the presentation of linear high SI signs along the cortical sulci on Gd-enhanced VWMR (Fig. 1). In the present study, this sign was prominently observed, particularly in high frontal and parietal areas of the patients with MMD. We termed the sulcal high SI signs on Gd-enhanced VWMR as the “VWMR ivy sign.” VIS was judged by comparing T1 black blood pre- and post-contrast sequences on VWMR, and was investigated in the precentral, central, and postcentral sulci. “Positive” and “negative” were defined as linear high SI and invisible on Gd-enhanced VWMR, respectively. The “VIS score” was calculated as the sum of VIS of the precentral, central, and postcentral sulci in each patient, ranging 0-3. VIS scores were evaluated in the symptomatic hemispheres of the ischemic and hemorrhagic groups of patients with MMD. For unilateral MMD in the non-stroke group, VIS scores were evaluated in the affected hemisphere. For definite MMD in the non-stroke group, the VIS scores were evaluated on the “dominant” hemisphere, which had the higher VIS score, and thus a more intense amount of ivy signs between the two hemispheres [18]. The conventional ivy sign in the corresponding hemispheres was also assessed in the precentral, central, and postcentral sulci using both FLAIR and CEMR imaging. All of the images were reviewed by one neurosurgeon (J.R.) and one neuroradiologist (K.M.L.) who were completely blinded to the patients’ clinical information, and a consensus was reached between two raters.
VWMR protocol
High-resolution VWMR was performed using a VIDA 3.0-Tesla scanner with a 128-channel head and neck coil (Siemens, Erlangen, Germany). The VWMR protocol included time of flight magnetic resonance angiography (TOF MRA) with maximum intensity projection, T1 weighted imaging (T1WI), proton density (PD), black blood image, and contrast-enhanced T1WI. MRI sequences were performed as follows : T1W spin echo (in-plane saturation, 0.5×0.5 mm; slice thickness, 0.5 mm; repetition time (TR)/echo time (TE), 700/13 ms; field of view, 180×180 mm; time, 7 minutes 5 seconds), PD imaging (in-plane saturation, 0.5×0.5 mm2; slice thickness, 0.5 mm; TR/TE, 1200/23 ms; field of view, 180×180 mm; time, 7 minutes 5 seconds), and contrast-enhanced T1WI after intravenous Gd administration (Prohance; Bracco, Milan, Italy). The total acquisition time was approximately 22 minutes. Three-dimensional images were reconstructed into coronal, sagittal, and axial images and displayed with an isovoxel of 0.5×0.5×0.5 mm3.
Statistical analysis
The Kruskal-Wallis and chi-squared tests were used to evaluate the differences in VIS scores among the 18 patients in the three MMD groups. If the Kruskal-Wallis and chi-squared test showed a significant difference among the three groups; a post-hoc multiple comparison using the Mann-Whitney U and chi-squared tests was performed using the Bonferroni correction. p<0.05 was deemed statistically significant, except when the Bonferroni correction was used, where statistical significance was set at p<0.0167. Logistic regression models were used to adjust for age and sex. Inter-modality agreement between VIS and the conventional ivy sign was assessed using the Cohen κ statistics with a 95% confidence interval (CI). The Kappa coefficient ranged from 0 to 1, with agreement strength categorized using the guidelines of Landis and Koch : <0, poor; 0-0.20, slight; 0.21-0.4, fair; 0.41-0.6, moderate; 0.61-0.80, substantial; and 0.81-1, almost perfect [6]. Statistical analyses were performed using the SPSS software (version 24.0; SPSS Inc. Chicago, IL, USA).
RESULTS
Patient demographic data and clinical characteristics of the three MMD groups are summarized in Tables 1 and 2, respectively. Fifty-four sulci were investigated in all 18 patients. The median values of VIS scores in the ischemic, hemorrhagic, and non-stroke groups were 3 (interquartile range [IQR], 1.75-3), 3 (IQR, 2.5-3) and 0 (IQR, 0-1.25), respectively. The VIS scores were significantly different among the three groups (p=0.004, Kruskal-Wallis test). The VIS scores in both the ischemic and hemorrhagic groups were significantly higher than those in the non-stroke group (ischemic to non-stroke, p=0.009; hemorrhagic to non-stroke, p=0.004; Mann-Whitney U test with Bonferroni correction). No significant difference was found between the ischemic and hemorrhagic groups (p=0.699). One control patient who was diagnosed with an unruptured middle cerebral artery (MCA) aneurysm and had no ischemic symptoms, showed no VIS (Supplementary Fig. 1).
The VIS scores were entered in the logistic regression model using the non-stroke group as the reference group (Table 3). Age- and sex-adjusted analyses were performed to compare VIS scores between the non-stroke-to-ischemic and non-stroke-to-hemorrhagic groups. The VIS scores remained significantly higher in both the ischemic (p=0.046; odds ratio [OR], 8.27; 95% CI, 1.03-66.19) and hemorrhagic (p=0.039; OR, 7.78; 95% CI, 1.11-54.48) groups, compared to the non-stroke group. Excluding one patient, VIS was found in at least one of the precentral, central, and postcentral sulci in all hemispheres affected by stroke, including ischemic or hemorrhagic stroke. Only two non-stroke patients with MMD showed VIS.
Identification of VIS and FLAIR/CEMR ivy signs was further compared (Table 4). In the identification of the ivy sign within the precentral sulcus, one and two discrepancies were observed between VIS and the FLAIR ivy sign, and between VIS and the CEMR ivy sign, respectively. In the detection of the ivy sign within the central sulcus, VIS and FLAIR ivy signs were consistent, while one discrepancy was observed between VIS and CEMR ivy signs. In the identification of the ivy sign within the precentral sulcus, two and three discrepancies were observed between VIS and FLAIR ivy signs, and between VIS and CEMR ivy signs, respectively. The intermodality agreement between VIS and FLAIR ivy sign was substantial in the precentral region, perfect in the central region, and substantial in the postcentral region (precentral sulcus : κ=0.609; 95% CI, 0.213-1; central sulcus : κ=1; and postcentral sulcus : κ=0.769; 95% CI, 0.475-1). Inter-modality agreement between the VIS and CEMR ivy sign were substantial in the precentral, central, and postcentral sulci, respectively (precentral sulcus : κ=0.727; 95% CI, 0.384-1; central sulcus : κ=0.609; 95% CI, 0.384-1; and postcentral sulcus : κ=0.649; 95% CI, 0.310-0.998).
Representative cases
Case 1
A 35-year-old female patient experienced a right internal border zone infarction (Fig. 2). The patient had no relevant medical history. Right internal carotid artery angiography (ICAG) revealed stenosis of the terminal ICA and marked development of lenticulostriate arteries. Left ICAG showed no abnormal angiographic vasculature. Gd-enhanced 3D VWMR allowed visualization of VIS in the precentral, central, and postcentral sulci of the right hemisphere, with no visualization of VIS in the left hemisphere. Images in TOF MRA showed no arterial flow in the central and postcentral sulci. The site of VIS was well-matched with that of the ivy signs on FLAIR and CEMR. The patient underwent superficial temporal artery-middle cerebral artery bypass of the right hemisphere. Gd-enhanced 3D VWMR at follow-up one year postoperatively showed the disappearance of VIS in the precentral, central, and postcentral sulci of the right hemisphere, which appeared more pronounced compared with the postoperative decrease of the ivy sign on FLAIR and CEMR. Furthermore, postoperative TOF MRA showed newly developed arterial flow in the central and postcentral sulci.
Case 7
A 40-year-old female patient presented with headache and mental deterioration (Fig. 3). The patient had no pre-existing medical history. Brain CT on admission revealed intraventricular hemorrhage and intracerebral hemorrhage at the posterior two-thirds of periventricular white matter. The patient underwent external ventricular drainage at both Kocher points. Right ICAG revealed stenosis of the terminal ICA and marked development of periventricular anastomosis. Left ICAG showed no involvement of MMD. Gd-enhanced 3D VWMR clearly depicted visualization of VIS in the precentral, central, and postcentral sulcus of the right hemisphere, with no visualization of VIS in the left hemisphere. The site of VIS in the precentral and central sulcus corresponded well with the ivy signs on FLAIR and CEMR.
Case 13
A 55-year-old female patient presented with dizziness (Supplementary Fig. 2). Brain CT and MRI revealed no evidence of stroke. Digital subtraction angiography demonstrated Suzuki grade 6 and grade 4 for the right and left hemispheres, respectively. Gd-enhanced 3D VWMR did not reveal VIS in the precentral, central, or postcentral sulcus of both hemispheres. However, FLAIR and CEMR demonstrated ivy signs in the central, and postcentral sulcus of the right hemispheres and in the central sulcus of the left hemisphere.
DISCUSSION
In this study, the presence of VIS in the precentral, central, and postcentral sulci was investigated in 18 MMD patients using Gd-enhanced VWMR. Compared to non-stroke patients with MMD, the VIS scores were significantly higher in patients with stroke symptoms. The inter-modality agreement between VIS and conventional ivy signs was mostly substantial. However, a few discrepancies between the two modalities suggest that the distribution of VIS does not completely match that of conventional ivy signs.
Although the precise mechanism underlying the visualization of VIS in MMD remains unclear, the similar topographic sites of the VIS and conventional ivy signs suggest that VIS may share a common pathophysiology with conventional ivy signs. In the present study, this sign was prominently observed particularly in the high frontal and parietal area of patients with MMD. It was also sometimes detected at the same sulcus where the ivy sign was also visible on CEMR or FLAIR imaging. The FLAIR ivy sign was also reported to be observed prominently in the high frontal and parietal lobes [1]. The primary mechanisms underlying the FLAIR ivy sign have been proposed as follows : 1) slow retrograde flow of engorged pial arteries via leptomeningeal anastomosis [9], 2) maximally dilated pial vasculature compensating for decreased perfusion pressure [4], and 3) congestive thickening of the leptomeninges [9]. One prior report also suggested that the CEMR ivy sign may correspond to angiographic cortical microvascularization via leptomeningeal collaterals [13]. While these observations state the conventional ivy signs in CEMR or FLAIR imaging to be seen in the pia, VIS on the other hand appeared to lie in the sulcus space rather than on the cortical surface. Furthermore, our representative case (case 1) showed that while the VIS in the central and postcentral sulci disappeared after revascularization surgery, postoperative TOF MRA showed new arterial flow in the same locations. Based on these observations, we speculated that VIS may signify slow and retrograde flow of the sulcal vessels via leptomeningeal collateral. However, as this study did not pathologically confirm these vessels, future studies should aim to observe the specific identities of the vessels.
Another common observation shared between VIS and the conventional ivy sign is the postoperative changes following revascularization surgery. In one study, Kawashima et al. [3] demonstrated a postoperative decrease of the FLAIR ivy sign after revascularization surgery. A previous study using CEMR also showed a reduction in leptomeningeal contrast enhancement after bypass surgery [5]. The total disappearance of the precentral, central, and postcentral VIS after direct revascularization was further illustrated in case 1, presenting a more distinct delineation than the postoperative change of the FLAIR or CEMR ivy sign. Such postoperative changes may suggest the potential of this sign as a radiological marker during follow-up after bypass surgery in adult MMD patients. Future studies regarding VIS disappearance after revascularization surgery should be performed for more insight.
Some findings may suggest the underlying different pathophysiology between VIS and FLAIR ivy sign. A researcher assumed that the underlying pathophysiology of FLAIR and CEMR ivy sign may also differ [11]. The FLAIR ivy sign is an MMD-specific finding that can aid in diagnosing this condition [1,12]. However, VIS may not be exclusive to MMD, but rather associated with stroke symptoms. Indeed, the one MMD patient without stroke symptoms (case 13) exhibited no VIS on Gd-enhanced VWMR, although the ivy sign was visible on FLAIR and CEMR imaging. A patient with atherosclerotic intracranial artery disease who presented with cerebral infarction also showed sulcal hyperintense arteries on FLAIR and postcontrast VWMR (Supplementary Fig. 3). A previous study demonstrated that hyperintense arteries on FLAIR and postcontrast VWMR in acute MCA stroke were associated with slow collateral flows [7]. Furthermore, the FLAIR ivy sign is correlated not only with stroke presentation but also with hemodynamic status [4,12,18]. However, although this study showed that VIS scores were significantly higher in patients with ischemic or hemorrhagic stroke compared to those without stroke, it did not investigate the association between VIS and hemodynamic status. A previous study demonstrated that pial arteries dilate under hypoxic conditions, while parenchymal arterioles lack the response [15]. Accordingly, the size of the arterial structure responsible for VIS may be larger than that responsible for FLAIR ivy signs.
FLAIR and CEMR may have some limitations and challenges in identifying sulcal ivy sign in each sulcus. The slice thickness of FLAIR imaging is usually 5 or 6 mm in the axial plane. Further, subarachnoid hemorrhage in the cortical sulci appears as high SI on FLAIR images, posing a challenge in identifying the FLAIR ivy sign in patients with hemorrhagic MMD. The scoring system of the FLAIR ivy sign was also graded according to the degree the ivy signs on the brain surface rather than the cortical sulcus [2,12]. In terms of CEMR, diffuse leptomeningeal enhancement may sometimes interfere with the clear identification of sulcal ivy sign. In contrast, the technical characteristics of VWMR include 1) high spatial resolution using isotropic voxel dimensions in the 0.4 to 0.7 mm range, 2) multiplanar 3D acquisitions, 3) multiple tissue weightings, and 4) suppression of signals in the cerebrospinal fluid and brain parenchyma [10]. These strengths may allow for a simple and precise identification of VIS in each sulcus in patients with MMD.
Overall, the results of this study indicate that the spatial evaluation of VIS may contribute to our understanding of the status of leptomeningeal collateral circulation in patients with MMD. While this study focused on investigating VIS in the precentral, central, and postcentral sulci, further research is warranted to encompass VIS in other sulci of the frontoparietal area. This broader examination may facilitate the development of an optimal scoring system for VIS and establish its clinical correlation through additional studies.
The present study had some limitations. First, this study may have a selection bias owing to its retrospective nature. Second, the study population was small. Third, although we suggested that VIS may signify slow and retrograde flow of sulcal vessels, based on two of our findings—1) a correlation between VIS and stroke presentation, and 2) postoperative disappearance of VIS—the specific mechanism of VIS is still not fully understood. Fourth, although Table 4 of this study suggests there may be a difference between the underlying pathophysiology of VIS and conventional ivy signs, additional studies are required to gain a complete understanding of these differences. Fifth, since this study only suggests a correlation between VIS and stroke presentation, and we still do not fully understand the diagnostic capabilities of VIS in MMD patients or its relationship with hemodynamic parameters. Therefore, future studies that focus on these topics may provide further insight into the clinical significance of VIS.
CONCLUSION
Gd-enhanced VWMR can effectively characterize the status of sulcal vessels through leptomeningeal collateral circulation among patients with MMD. This preliminary series introduces the concept of VIS, suggesting slow and retrograde flow of sulcal vessels via the leptomeningeal collateral. Future studies are needed to develop an optimal scoring system for VIS and establish its clinical correlation with stroke presentations in MMD patients.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation 
Supplement1
Print 
