Quantitative Optical Coherence Tomography Angiography in Patients with Moyamoya Vasculopathy: A Pilot Study

ABSTRACT

Moyamoya (MM) disease is a chronic cerebrovascular disease that can lead to progressive stenosis of the terminal portions of the internal carotid arteries and their proximal branches. We sought to investigate and quantify retinal vascular changes in patients with MM vasculopathy (MMV) using optical coherence tomography angiography (OCTA) compared to healthy controls. Our findings reveal retinal microvascular changes in patients with MMV and highlights the potential of OCTA imaging for the detection of subclinical retinal pathology.

Introduction

Moyamoya vasculopathy (MMV) refers to collateral blood vessels at the base of the brain that resemble a ‘puff of smoke’ on angiography studies.¹ This pattern is seen as a result of vascular occlusions from moyamoya (MM) disease or MM syndrome. MM disease is a chronic cerebrovascular disease that typically leads to progressive stenosis of the terminal portions of the internal carotid arteries (ICAs) and their proximal branches, which in turn leads to limited blood flow to the major blood vessels of the anterior circulation. Rarely, the posterior circulation may also be affected. MM disease is, by definition, bilateral even though the severity may differ between sides.² The exact aetiology of the disease remains unknown and patients tend to present with transient ischaemic attacks or ischaemic stroke early in life.³ Some researchers hypothesise that different types of inflammatory responses are involved in hyperplasia of intimal vascular smooth muscle cells and neovascularisation, that ultimately lead to vascular lumen stenosis and collateral formation.⁴ Through genome-wide linkage analysis it has been identified that RNF213 is a strong susceptibility gene for MM disease in people of East Asian descent.⁵ MM syndrome, on the other hand, can be linked to previous radiotherapy, Down’s syndrome, neurofibromatosis type 1 and sickle-cell disease inter alia.¹ Ocular manifestations of MMV are not common, however, there have been a number of cases describing: morning glory disc anomaly and chorioretinal colobomas, likely as a consequence of neuroectodermal dysgenesis; amaurosis fugax, ocular ischaemic syndrome and rarely retinal vascular occlusion, secondary to anterior circulation insufficiency; homonymous hemianopia, as a result of posterior circulation insufficiency; and diplopia following intracranial haemorrhage.^6–10 The exact pathophysiology of retinal artery and vein occlusions in MMV patients remains controversial, as the chronic progressive course normally leads to collateral formation, which usually prevents acute vascular events.^11–13

In the past, a single study using optical coherence tomography (OCT) has demonstrated a significant reduction in optic nerve head volume and retinal nerve fibre layer thickness in patients with MMV compared to healthy controls.¹⁴ As far as we know, there has not been a study that has investigated retinal vascular changes associated with MM disease or syndrome using OCT angiography (OCTA). OCTA is a non-invasive imaging technique that rapidly generates volumetric angiography images without the need for fluorescein dye. This imaging modality allows for qualitative and quantitative analysis of the retinal vasculature.¹⁵ As such, with limited research on this topic and the wide range of potential, OCTA may possibly serve as a method to identify retinal vascular abnormalities in MMV. Thus, our study sought to investigate and quantify retinal vascular changes in patients with MMV compared to healthy age- and sex-matched controls.

Patients and methods

This cross-sectional cohort study was conducted at the Eye Care Centre at Vancouver General Hospital, Vancouver, Canada. The Clinical Research Ethics Board at the University of British Columbia (H18-02488) approved the research project and the study was conducted in accordance with the tenants of the Declaration of Helsinki. All volunteer subjects were required to sign a consent form.

Patients who had received a diagnosis of MMV from their neurologist were mailed an invitation to participate in the study. Interested patients were subsequently contacted by telephone to discuss details of the project and their involvement. Diagnosis of MMV was based on the characteristic appearance of stenosis in specific arteries and branches that are commonly affected in the disease process. A thorough evaluation was completed for each patient to establish a diagnosis, which included a neurological examination and brain magnetic resonance (MR) imaging including MR angiography. Diagnosis was made by a single neurologist in Vancouver, Canada with uniform parameters. Patients over the age of 18 with MMV were enrolled in the study and matched with age-similar healthy controls. Eyes with evident optic nerve head abnormalities, glaucoma, high refractive error (spherical equivalent <−6,6 or >+6 dioptres) or significant media opacity and patients with other vascular disorders affecting the central nervous system (CNS) or retina, including diabetes and neurodegenerative disorders were excluded from the study. Retrospective analysis was done on clinical and imaging records (MR and formal angiography) of baseline and follow up exams. The recruitment of patients began in April 2019 and was discontinued in March 2020, due to the COVID-19 pandemic.

Each subject was imaged using the PLEX Elite 9000 Swept-Source OCTA instrument (Carl Zeiss Meditec Inc., Dublin, CA, USA). This instrument uses a swept laser source with a central wavelength of 1060 nm and a scanning rate of 100,000 A-scans per second. For each eye, three to five 6 × 6 mm images centred at the fovea were captured at the same visit. Poor quality images with considerable motion artefact were excluded. A minimum of three high-quality images of each eye were selected for analysis. The OCTA was solely used for acquisition of images, while analysis was subsequently done independently. As described by Lo et al., a convolutional neural network (CNN) was used to delineate the vasculature, followed by individual analysis of the inter-capillary regions.¹⁶ The CNN used a U-net architecture and was trained on 6×6 mm OCTA enface images acquired with the same system as used in this study but from subjects with diabetic retinopathy. Following vessel delineation, the inter-capillary areas were individually measured; the metric of interest was the maximum distance to nearest vessel within the region. Subsequently, each inter-capillary area was colour-coded and overlaid on the original OCTA image based on the number of standard deviations it exceeded a reference mean. Three vascular parameters were extracted from the vessel segmentation network results: vessel density (whole-image density, central density); skeleton density (SD) and fractal dimension (FD). The image was binarised and the whole image density was calculated as a proportion of measured area occupied by pixels that were classified by the CNN as a vessel. Central density (VDC) was calculated as the density within a 1 mm diameter from the centre, and SD as the density of the skeletonised vessels. FD represents the complexity of vessel branching by calculating the minimum number of ‘box’-shaped segments required to cover the vasculature.

In addition, two foveal avascular zone (FAZ) morphometric parameters were calculated from the vessel segmentation network results: area (FAZ) and axis ratio (AR). The FAZ was segmented by locating the largest non-vessel area within a 1 mm diameter from the centre. The centroid for this area was used to determine the maximum diameter, and minimum diameter. AR was calculated as a ratio of the maximum FAZ diameter to the minimum FAZ diameter.

Excel for Mac Version 14.5 and IBM SPSS Statistics Version 22.0 (SPSS Inc, Chicago, Illinois, USA) were used for data entry and statistical analysis. Descriptive statistics of mean, range, and standard deviation were calculated. The Shapiro Wilk test and Q–Q plots were used to test for normality of distribution. Levene’s test determined equality of variances. Comparisons were assessed with independent samples t-test for parametric data and Mann–Whitney U test for non-parametric data. A significance level of 0.05 was pre-determined. Data was reviewed for multiple comparisons, with Bonferroni and Benjamini–Hochberg analysis.

Results

Four MM patients and eight age- and sex-matched healthy controls had OCTA testing during a two-year period. Demographic and clinical features of the MM patients are presented in Table 1.

Table 1.

Clinical characteristics of patients with moyamoya vasculopathy

Patient Number	Age	Sex	Disease Duration (Years)	Ocular Diagnoses	MRI findings	Neurologic Findings at Presentation	Surgical Treatment
1	50	M	4	None	Bilateral moyamoya pattern, right MCA stenosis, left MCA occlusion	Left SAH, Right perforator stroke	ECA to ICA bypass 2016
2	53	M	4	None	Left ICA moyamoya pattern, Left M1/MCA occlusion	Left hemispheric stroke	None
3	19	F	8	Peter’s anomaly left eye, NF1	Left MCA and ACA occlusion with moyamoya pattern	Recurrent right TIA	EDAMS procedure
4	47	F	4	None	Right M1/MCA occlusion with moyamoya pattern	Right parietal SAH, seizure activity	Right STA to MCA bypass

ACA = anterior cerebral artery; ECA = external carotid artery; EDAMS = encephalo-duro-arterio-synangiosis; ICA = internal carotid artery; MCA = middle cerebral artery; NF1 = neurofibromatosis type 1; SAH = subarachnoid haemorrhage; STA = superficial temporal artery; TIA = transient ischaemic attack.

One eye (left eye of M3) was excluded due to Peter’s anomaly, resulting in 23 total eyes included in analysis. The mean patient age was 42 ± 16 years (range: 18 to 53 years) and the mean age of healthy controls was 42 ± 15 years (range: 20 to 63 years). The male-to-female ratio was 1:1 for patients and healthy controls. Two patients were diagnosed with MM disease and two with MM syndrome. The ethnicity of MM patients was determined as three Asian and one white Caucasian patient. Healthy controls were four white Caucasians, two Asians, and two individuals of mixed ethnicity. Uncorrected visual acuities of included patient eyes were 0.05 ± 0.06 logMAR (range: 0 to 0.1) and controls had visual acuities of ≤0.0 logMAR.

Mean central retinal vessel density of superficial and deep plexuses, as measured with OCTA, was 18 ± 10.0% in MMV and 21.0 ± 8.0% in healthy controls (p = .047). The mean overall vessel density measured 37.0 ± 6.0% in MMV and 39.0 ± 4.0% in healthy controls (p = .443). Mean skeleton density measurements resulted in 0.12 ± 0.01 for both groups (p = .112). FD was 1.88 ± 0.01 in MMV and 1.89 ± 0.01 in controls (p = .548). The FAZ measured 0.43 ± 0.17 mm² in MMV patients and 0.43 ± 0.19 mm² in healthy controls (p = .982). The axis ratio of the FAZ resulted in 1.54 ± 0.22 for the MMV group and 1.45 ± 0.15 in controls (p = .151). Both superficial and deep retinal plexuses were considered for the analysis of all parameters, which are depicted in Table 2. No retinal microaneurysms or collaterals were observed in the studied macular subfield of patients with MMV. Representative OCTA images of MM patients versus controls are shown in Figure 1.

Table 2.

Results of analysed parameters (mean ± standard deviation)

Parameter	Healthy controls (n = 8)	Moyamoya (n = 4)
Vessel Density Central (%)	21.0 ± 8.0*	18.0 ± 10.0*
Vessel Density Overall (%)	39.0 ± 4.0	37.0 ± 6.0
Skeleton Density	0.12 ± 0.01	0.12 ± 0.01
Fractal Dimension	1.89 ± 0.01	1.88 ± 0.01
Foveal Avascular Zone (mm2)	0.43 ± 0.19	0.43 ± 0.17
Axis Ratio	1.45 ± 0.15	1.54 ± 0.22

*p = <0.05. When comparing Vessel Density Central, the differences remained significant (p = 0.282) with Benjamini-Hochberg test for multiple comparisons.

Figure 1.

Vessel densities as measured by optical coherence tomography angiography of a moyamoya patient: (A) superficial plexus and (B) deep plexus compared to a healthy control: (C) superficial plexus and (D) deep plexus

Discussion

MM disease is defined as a cerebrovascular condition in which progressive stenosis of intracranial vessels and subsequent collateral formation may result in diverse ischaemic or haemorrhagic events. To date, there has been a single study reporting the OCT findings in patients with MM, but no data on OCTA exist so far. Albrecht et al. found decreased optic nerve head volume and macular retinal nerve fibre layer in patients with idiopathic MM disease, which they attributed to mesenchymal dysgenesis of the cerebral and retinal vasculature.¹⁰

Our report presents the results of various vascular parameters from OCTA imaging of four MMV patients and compared them to age- and sex-matched healthy controls. We found a trend towards poorer microvascular parameters, that is, reduced vessel density in patients with MMV, which indicates retinal microvascular involvement in the absence of clinical signs or symptoms, albeit the majority of parameters did not differ significantly. Research suggests that the underlying pathophysiology of MM disease is associated with a dysregulation of matrix metalloproteinases affecting blood vessels in genetically predisposed individuals.¹⁷ Histopathological studies of intracranial vessels have demonstrated a combination of hyperplasia of smooth-muscle cells with luminal thrombosis and dilated perforating arteries that serve as collaterals.¹⁸ The ICAs are commonly affected in MMV. Hence, one would expect changes at the level of the retina, as it is mainly supplied by the ophthalmic artery, the first branch of the ICA. Whether these microvascular changes mirror cerebrovascular abnormalities in MMV remains yet to be determined. The main advantage of retinal versus intracranial imaging is that it allows for non-invasive and rapid evaluation.

The Limitations of this study are its small sample size, extensive age range, lack of visual field testing and fluorescein angiography at baseline and inclusion of patients with both MM disease and MM syndrome. Age-related changes or subclinical vascular changes as a result of secondary diseases may confound our results. In addition, our patient group and healthy controls were not matched by ethnicity. Larger prospective studies and studies including optic nerve head data are required, yet OCTA imaging holds the potential to establish quantitative parameters for subclinical microvascular retinal pathology in MMV.

Funding Statement

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Declaration of interest statement

None of the authors have any financial interests or conflicts related to this research to disclose.

Ethical considerations

This study adhered to the Declaration of Helsinki and was approved by the Clinical Research Ethics Boards at the University of British Columbia (H18-02488).