Collateral Predictors of Neovascularization After Indirect Revascularization in Adult Patients with Moyamoya Disease: A Single-Center Retrospective Study


 Objective.

The underlying factors of neovascularization after indirect bypass in the adult patients with moyamoya disease (MMD) remained unknow. The aim of this study was to explore potential predictors based on collateral characteristics for neovascularization after indirect bypass in adult MMD patients.
Methods.

The adult MMD patients treated by indirect bypass in a single-institution from August 2012 and January 2018 were retrospectively selected into our research. The collaterals based on cerebral angiography were classified into the following subtypes: intracerebral anastomosis, duro-cortical anastomosis, and leptomeningeal anastomosis. Neovascularization evaluation was based on Matsushima classification, with “good” collateral formation in level 2 and 3, and poor formation in level 0 and 1. Univariate and multivariate analyses were performed to identify neovascularization predictors after indirect bypass.
Results.

A total of 86 patients (97 hemispheres) (mean ± SD age 35.06 ± 15.09 years, range 18–61 years) were retrospectively included. Preoperative collateral circulation included intracerebral anastomosis in 49 (50.5%) cases, duro-cortical anastomosis in 19 (19.6%) cases, and leptomeningeal anastomosis in 29 (29.9%) cases, respectively. Postoperative good neovascularization was observed in 56 (57.7%) hemispheres. Multivariate analysis showed that intracerebral anastomosis (P<0.001, OR [95% CI] 2.984 [2.031-5.437]) was associated with favorable neovascularization, whereas older age (P<0.001, OR [95% CI] 0.827 [0.793-0.916]) and hemorrhagic onset (P<0.001, OR [95% CI] 0.138 [0.054–0.353]) were significantly associated with poor neovascularization. Hemispheres in the good neovascularization had lower modified Rankin scale score, and better long-term improvement than those in the poor neovascularization.
Conclusions.

Hemorrhagic onset and old age predict poor neovascularization after indirect bypass, while duro-cortical anastomosis and intracerebral anastomosis predict good neovascularization. Good neovascularization was associated with better long-term outcomes. The current study provides a basis for the selection of surgical procedure for MMD candidates.


Introduction
Moyamoya disease (MMD) characterized by a progressive stenosis or occlusion of terminal portion of bilateral internal carotid artery (ICA), proximal portion of middle cerebral artery (MCA) and/or proximal anterior cerebral artery (ACA), and formation of moyamoya vessels at the base of brain [1][2][3] . Cerebral revascularization was recommended as the mainstream of treatment for patients with MMD, including direct bypass, indirect bypass and combined bypass [4][5][6][7] . Direct revascularization may be considered as the rst-line treatment over indirect methods for adult patients with MMD, while indirect revascularization may be considered as an alternative to direct methods in adult MMD when the latter is not available 6, 8- 10 . Previous literature reported that indirect bypass was equivalent to direct bypass or combined bypass for the prevention of stroke recurrence in adult patients with MMD. 9,[11][12][13] Postoperative neovascularization was associated with favorable clinical outcome 14 15 . Identifying other potential risk factors of postoperative neovascularization after indirect bypass was bene cial to bene t adult patients with MMD from this procedure. The aim of this study was to explore potential predictors based on collateral characteristics for neovascularization after indirect bypass in adult MMD patients.

Patient Selection
The research was approved by the Ethics Committee of Beijing Tiantan Hospital and written informed consent was obtained. The participants selected into our research were from a single-center of adult MMD patients between August 2012 and January 2018. The inclusion criteria were following: 1) patients diagnosed with MMD according to guidelines published by the Research Committee on MMD 16 ; 2) patients treated by encephalo-duro-arterio-synangiosis (EDAS) alone; 3) patients performed with preoperative and postoperative cerebral angiography after revascularization; 4) patients older than 18 yrs.
Patients with age less than 18 years old, moyamoya syndrome (MMS), other bypass procedures, posterior cerebral artery (PCA) involvement and unwilling to participate in the research were excluded.
Finally, a total of 86 patients (97 hemispheres) were recruited into the research. Clinical data on gender, age, onset presentation, past medical history, modi ed Rankin Scale (mRS) on admission, and radiological ndings were collected.

Evaluation of collateral vasculature
Preoperative collateral vasculature was evaluated by two independent neurosurgeons who were blinded to surgical information based on the criteria proposed by Gerasimos Baltsavias et al. 17 . The collateral circulation system was de ned as follows: (1) intracerebral anastomosis, including the direct connections of striate arteries with the medullary arteries of the hemispheres; (2) durocortical anastomosis, consisting of collaterals between major intracranial and extracranial arteries (including ophthalmic artery (OphA), middle meningeal artery (MMA), super cial temporal artery (STA), and occipital artery (OcciA)); (3) leptomeningeal anastomosis, consisting of the anastomoses of the PCA-ACA branches and PCA-MCA branches at the watershed zones. This system also included a leptomeningeal network fed by the uncal artery, a branch of the proximal anterior choroidal artery (AchA) or distal ICA ( Figure 1).

Surgical Treatment
Patients who met the following criteria were considered as unsuitable for direct revascularization: 1) those with no available recipient vessel for direct anastomosis during operation; 2) those who experienced frequent transient ischemic attacks (TIAs) during preoperative period (more than 3 times within 3 months before surgery) and/or those who experienced recent cerebral infarction (within 3 months before surgery). EDAS is the preferred indirect bypass revascularization procedure in our institution, with branch of the super cial temporal artery (STA) as the main donor artery.

Evaluation of postoperative neovascularization
Postoperative collateral grading was evaluated by reviewing the lateral view of the external carotid artery (ECA) based on cerebral angiography at 6 months after surgery. Collateral evaluations were performed by two independent neurosurgeons who were not involved in the surgery and were blinded to the clinical characteristics of MMD patients. The newly developed collateral circulation was evaluated by using the Matsushima gradings 5, 10 . Matsushima grading system was de ned as follows: grade 0, no collaterals present in the target revascularization area; grade 1, collaterals neovascularization presented in less than 1/3 MCA territory; grade 2, collaterals presented between 1/3 and 2/3 MCA territory; grade 3, collaterals presented in more than 2/3 territory of the MCA. We arti cially determined levels 2 and 3 as "good" collateral formation, whereas levels 0 and 1 were determined as "poor" collateral formation ( Figure 2).

Clinical follow-up
Score of mRS was used to evaluate neurological status on admission and at nal follow-up conducted by clinic visit or telephone up to minimal 6 months after surgery. Functional improvement and deterioration were de ned as decrease or increase of mRS scores, respectively. During the clinic follow-up, ischemic events, hemorrhage, epilepsy and death were included.

Statistical Analysis
Statistical analysis was performed with the use of SPSS (version 22.0., IBM Corp.). Continuous variables with normal distribution were compared using a t-test, and skew distribution using with Mann-Whitney test. Categorical variables were compared with the Pearson chi-square test or Fisher's exact test as appropriate. Multivariate logistic regression analysis was conducted, including factors with P < 0.05 in the univariate analysis and factors reported in previous literature or according to clinical experience. Odds ratios (ORs) and 95% con dence intervals (CIs) for good neovascularization were calculated. A p value < 0.05 was set as statistically signi cant.

Discussion
The effect of indirect revascularization on MMD relied on the neoangiogenesis from the vascularized donator into the cortical brain, which might act the same role in prefusion improvement and prevention in recurrence stroke as direct or combined bypass 18, 19 . Previous reports investigated that effect in neovascularization was associated with favorable long-term outcome for patients with MMD, and potential predictors for postoperative neovascularization was of great value to provide standard for elective surgical procedure before surgery 20 . Thus far, few studies had reported potential risk factors for neovascularization after indirect bypass. In this research, we attempted to investigate the relationship between postoperative neovascularization and a variety of possible factors.
Our ndings showed that good neovascularization was achieved in 56 (57.7%) hemispheres after indirect bypass. The presence of preoperative collateral circulation system acted as independent risk factor for postsurgical neovascularization (P<0.001, OR [95%CI], 2.984 [2.031-5.437]). Preoperative intracerebral anastomosis predicted good neovascularization after surgery, whereas leptomeningeal anastomosis was a predictor of poor neovascularization. As it is known that cerebral ischemia indicated a potential signal for triggering moyamoya vessels and collateral growth in the progress of MMD 21 . As intracerebral anastomosis directly connects striate arteries with the medullary arteries, the main source of blood perfusion in cerebral cortex was leptomeningeal collateral, such as the anastomosis of the PCA-ACA branches and PCA-MCA branches at the watershed zones. Piao reported the hemisphere with extensive basal moyamoya vessels exhibited impaired perfusion and metabolism in cerebral cortex 22 . Takahashi et al. also reported that severe ischemia was essential for the formation of new vessels 23 . The transdural anastomosis developed in the advanced Suzuki stage, such as meningeal middle artery, ethmoid moyamoya and vault moyamoya vessels [24][25][26] . Thus, we hypothesized that hemisphere with intracerebral anastomosis was more likely to suffer cerebral ischemia than those with leptomeningeal collateral, stimulating the development of collateral development after surgical revascularization. Our study was a practical interpretation of association between preoperative collateral anastomosis and postoperative neovascularization, and the mechanisms behind this phenomenon remain to be clari ed by future studies.
Previous studies demonstrated that indirect bypass was more effective for patients with young age and ischemic onset symptom 4,27  Previous literature reported that hemisphere in young age was associated with better cerebrovascular reactivity and higher secretion of angiogenic peptides, vascular endothelial growth factor (VEGF) stimulating collateral growth than in old age when suffering from cerebral ischemia 28-30 . However, Zhao reported that age was not an independent predictor for neovascularization after indirect bypass in the entire series, but the predictor of MMD patients in the hemorrhagic-onset type, which differed from our results 15 . As for age was not the most in uential predictor affecting surgical effect, it still needed to explore further.
In addition, in our series, 41.5% hemispheres in the hemorrhagic had poorer neovascularization than 7.1% in the ischemic (P<0.001, OR [95%CI], 0.138 [0.054-0.353]). Previous reports demonstrated that hemorrhagic onset type was a potential predictor for neovascularization after indirect bypass 15,31 . As is known the hemispheres with normal perfusion in the hemorrhagic were more common than the ischemic group. The MMD patients with hemorrhage tended to be diagnosed in early stage of the disease due to abrupt presence of brain hemorrhage, whereas the ischemic tended to be in the advanced stage 32 . We hypothesized that cerebral revascularization itself might be a destruction to the vulnerable blood supply from extracranial arteries for hemorrhagic MMD patient receiving surgery treatment too early, which could cause an adverse effect on postoperative neovascularization. Hence, the general superiority of direct bypass over indirect bypass for MMD patients should not be ignored, especially for the MMD patients with hemorrhage.
According to previous reports, good neovascularization was associated with better long-term outcomes 33,34 . Better neurological status and more improved outcomes were observed in hemispheres with good neovascularization, suggesting the effect of revascularization was crucial to the long-term outcome for adult MMD patients (improvement: 94.6% vs 80.5%, P=0.014; deterioration: 1.8% vs 12.2%, P=0.036). However, there was no signi cant difference in recurrent ischemia, rehemorrhage and seizure between hemispheres with poor and good neovascularization (P>0.05). This nding was different from results reported by Arias et al 35 ., due to the comparatively larger sample size of the study

Limitations
The current study had a few limitations as follow. First, our study was a retrospective research in a single center, and selection bias might exist. The conclusion needed a multicenter research for further veri cation. Second, sample size of the study was small, which need more patients enrolled to con rm the conclusion. Third, the current study did not elucidate the potential pathophysiology, and future studies were needed to further address this issue. Fourth, the main indirect bypass included in the study is EDAS, so there was no basis for the evaluation of the other indirect bypass surgery. Fifth, the study lacked data on cerebral perfusion, and radiological brain perfusion examination would be the next.

Conclusions
Hemorrhagic onset and old age predict poor neovascularization after indirect bypass, while duro-cortical anastomosis and intracerebral anastomosis predict good neovascularization. Good neovascularization was associated with better long-term outcomes. The current study provides a basis for the selection of surgical procedure for MMD candidates. Declarations