The Preoperative and Postoperative Cerebral Blood Flow and Vasoreactivity in Childhood Moyamoya Disease

Hon-ManLiu, Steven S. F. Peng and Yiu-WahLi DepartmentofRadiology, Medical College and Hospital, National Taiwan University, Taipei, Taiwan

Abstract. Purpose: to report the hemodynamic pattern in childhood moyamoya disease before and after vascular reconstruction. Methods: We performed the xenon cerebral blood flow (CBF) study in 10 cases of moyamoya disease in last 3 year. In 6 of them, study before and after operation was obtained for comparison. Preoperative study was obtained in 2 cases and postoperative study was obtained in 2 cases. They were 5 males and 5 females with age ranging from 1 to 12 years old. We studied their regional CBF before and after acetazolamide (CCBF) using stable xenon computed tomography. The angiography study was obtained for comparison. Result: In the routine study, the CBF and CCBF were poor in the ACA and MCA territory and well correlated to the angiography finding. However, in case of basal ganglion and posterior cerebral artery involvement, xenon CBF study was more sensitive than the angiography. On the postoperative cases, the CBF could increase to a level beyond the level of CCBF shown on the preoperative study. Conclusion: Hemodynamic study using stable xenon computed tomography can provide an objective reference for the treatment of moyamoya disease especially in case that surgical management is considered. It is a better modality for the evaluation of surgical outcome than the conventional contrast angiography study.

Key words: cerebral blood flow, xenon, moyamoya, operation

available because they had either a poor preoperative study, or the preoperative study was Introduction done with another machine. So, totally, there were Moyamoya disease is an intracranial 16 cerebral hemispheres in both preoperative and progressive occlusive lesion of the circle of Willis postoperative studies. with abnormal vascular network in the basal All the study was performed with a spiral CT ganglion region. The disease is far more common scanner (Picker, PQ 6000, Highland, Cleveland, in those Asian races. There are a number of USA) in a conventional mode. A xenon ventilator surgical techniques for revascularization for and computed equipment (Anzai, AZ-7000W, moyamoya disease. The purpose of this paper is Japan) was connected to the CT scanner so they to report the hemodynamics pattern in 10 cases of could work simultaneously and measure the childhood moyamoya disease before and after resting CBF (RCBF). We took 4 sections from vascular reconstruction. suprasellar region to the level of lateral ventricles in an interval of 1 cm. We used the protocol of 3 minutes wash-in and 5 minutes washout with a Materials and Methods xenon concentration of 30%. The measurement From January 1997 to April 1999, we were taken in both sides in all the vascular encountered 10 cases of childhood moyamoya territories including the anterior cerebral artery disease referred from the neurosurgical (ACA), the (MCA), the department for the cerebral blood flow study posterior cerebral artery (PCA), the striatum (CBF). (including the basal ganglion and head of caudate They were 5 males and 5 females. Their age nucleus), and the thalamus. Acetazolamide was was ranging from 1 to 12 years old. Six of them given intravenously in a dosage of 15mg/kg for had both the studies before and after operation. the vasoreactivity study. The acetazolamide Two cases had only the study before operation. challenge cerebral blood flow study (CCBF) was The other two had only the post-operative study taken 15 to 20 minutes after the injection of

A86 acetazolamide in the same sections. The 2, or 3 (the Suzuki classification) could have procedure was the same in both pre- and post normal to low RCBF and the difference between operative study. In all operative cases, the RCBF and CCBF could be more than operation procedures were 10ml/100g/minute in average. But in hemisphere encephalomyosynangiosis (EMS) and of stage 4, or 5, the RCBF and CCBF were much encephaloduroarteriosynangiosis (EDAS). decrease to the borderline level (20 All the angiography performed within 1 25ml/100g/min), and sometimes even came with month before operation and about 4 to 6 months negative challenge test, that mean the CCBF after operation. The examination included the became less than the RCBF. In the postoperative bilateral common carotid or selective internal study, the RCBF and CCBF increased in all the carotid (ICA), external carotid arteries (ECA) and cases in different ratio. vertebral artery injection before operation. In case of basal ganglion and posterior Selective ICA and ECA study would be cerebral artery territory involvement, the RCBF performed in all the operation side on the post study was more sensitive than the conventional - operative study. angiography. On the postoperative study, the resting RCBF could increase to a level beyond the measurement Results in the preoperative CCBF study. The average in In the pre-operative resting study, the RCBF preoperative CCBF study in ACA and MCA and CCBF were poor in the ACA and MCA regions were 28 and 53ml/100g/minute territory. The detail of the data was shown in respectively, while the postoperative RCBF study, figure 1 and figure 2. they were 44.1 and 59.5ml/100g/minute Regarding to the ACA and MCA territories, respectively. the hemisphere in with angiography of stage 1, or

Figure 1 Resting cerebral blood flow measurement in 16 hemispheres with moyamoya disease . Odd numbers represents preoperative study and even number represents the postoperative study. The arrowhead means the highest value, the bar means the average, and the square means the lowest value. ACA: anterior cerebral artery; MCA: middle cerebral artery; PCA: posterior

A87 Figure 2 Acetazolamide challenge cerebral blood flow measurement in 16 hemispheres with moyamoya disease. Odd numbers represents preoperative study and even number represents the postoperative study. The arrowhead means the highest value, the bar means the average, and the square means the lowest value. ACA: anterior cerebral artery; MCA: middle cerebral artery; PCA: posterior cerebral artery

vessels are not enough to perfuse the deep structures but we do have a good result even in Discussion the striatum area. We confirmed the finding that There are a lot of surgical techniques for the the CCBF decreased as the angiographic stage treatment of moyamoya disease. They include the advanced.7 indirect or direct revascularization procedures. There is uncertainty about the mechanism by Some reports that patients may refractory to the which acetazolamide dilates the cerebral vessels. indirect procedure, and sometimes the EDAS and From our study, we found that the postoperative EMS can not provide sufficient neoangiogenesis RCBF is much higher than the preoperative in the ACA and PCA territories. But for the CCBF in all the cases in ACA and MCA childhood moyamoya patients, it is generally territories. This indicates that preoperative CCBF considered the EDAS and EMS can form can not predict the postoperative RCBF level, the sufficient collateral pathways.1,2 increase of RCBF after revascularization is very In childhood moyamoya disease, the complex and further basic studies are needed to preoperative RCBF and CCBF were reduced. In clarify this. the postoperative study, most authors reported that the RCBF increase after revascularization. References However, the reports about CCBF were controversy, Nariai et al.3 found the CCBF were 1. Matsushima Y, Aoyagi M, Koumo Y, Takasato Y, increased, and Golby et al series, postoperative Yamaguchi T, Masaoka H, Suzuki R, Ohno K: studies were notable for improved baseline blood Effects of encephalo-duro-arterio-synangiosis on flow and increased reserve.4 Yamashita et al.5 childhood moyamoya disease: swift disappearance of ischemic attacks and maintenance of mental found postoperative RCBF increase in children capacity. Neurol Med Chir(Tokyo) 1991;31:708-14 with moyamoya disease but the postoperative 2. Kinugasa K, Mandai S, Kamata I, Sugiu K, CCBF did not significantly increase. In our series, Ohmoto T: Surgical treatment of moyamoya we found that both RCBF and CCBF increase in disease: operative technique for encephalo-duro the ACA, MCA and striatum region after -arterio-myo-synangiosis. Its follow-up, clinical operation but decrease in the PCA and thalamus. results, and angiograms. Suzuki et al.6 reported that neoangiogenesis 1993;32: 527-531

A88 3. Nariai T, Suzuki R, Matsushima Y, Ichimura K. revascularization in patients with moyamoya Hirakawa K, Iskii K, Senda M: Surgical induced disease. Acta Neurol Scand 1996; 166 (s): 82-84 angiogenesis to compensate for hemodynamic 6. Suzuki R, Nariai T, Matsu shima Y, Hirakawa K: cerebral ischemia. 1994; 25: 1014-1021 Xe-CT in and moyamoya 4. Golby AJ, Marks MP. Thompson RC, Steinberg disease. Acta Neurol Scand 1996: 166 (S): 69-71 GK: Direct and combined revascularization in 7. Fukuuchi Y, Kohari M, Shinohara T, Nogawa S. pediatric moymoya disease. Neurosugery Watanabe S: Annual Report of 1993 on the 1999; 45: 50-60 Cooperative Study of Occlusion of the Circle of 5. Yamashita T, Kashiwagi S, Nakashima K, Ishihara H, Willis to the Ministry of Health and Welfare. Kitahara T, Nakano S, Ito H: Modulation of Fukuoka, Japan, Kyushu University, 1994;77-83 cerebral hemodynamics by surgical

A89 Local Cerebral Hemodynamic Changes through the Angiographic Stages of Moyamoya Disease

Shigeru Nogawa, Yasuo Fukuuchi, Masahiro Kobari, Katsuyuki Obara, Shigeru Watanabe, Keiji Yamaguchi and Tomohisa Dembo

Department of Neurology, School of Medicine, Keio University, Tokyo, Japan

Abstract. In order to elucidate the cerebral hemodynamic changes that occur in Suzuki's six angiographic stages of moyamoya disease, local cerebral blood flow (LCBF) during the stable state and CO2 responsiveness of LCBF (L-CO2R:•¢%LCBF/•¢PaCO2) were measured by the Xenon CT-CBF method. Nineteen patients with moyamoya disease (mean age: 36.8•}11.6 years) and 11 age-matched normal volunteers were studied. The LCBF during the steady state at all stages was not significantly different from that in normal volunteers. At stage 6, however, the LCBF was slightly decreased in the anterior part of the , resulting in loss of "hyperfrontality ." On the other hand, the L-CO2R in the anterior part of the brain tended to diminish with progression through the stages. Especially in the frontal cortex, the L-CO2R at stage 5 was significantly less than that in normal volunteers (p<0.01) or at stage 3 (p<0.05). In conclusion, the cerebrovascular reserve in the anterior circulation became insufficient after stage 4, although the posterior circulation was well maintained. Revascularization surgery involving the anterior circulation may be crucial to prevent ischemic events.

Key words: local cerebral blood flow, CO2 responsiveness, cerebrovascular reserve, Xenon CT CBF method

used in Japan to evaluate the extent and Introduction progression of moyamoya disease. However, the relationship between the angiographic stages Spontaneous occlusion of the circle of and hemodynamic status in moyamoya disease Willis,1 also called moyamoya disease,2 is a rare has not been fully elucidated.4 The purpose of cerebrovascular disorder characterized by slowly the present study is to examine how local progressive stenosis of bilateral terminal portions cerebral blood flow (LCBF) and cerebrovascular of the internal carotid arteries (ICAs). The reserve change through the angiographic stages. steno-occlusive change of the ICAs is followed by development of abnormal net-like "moyamoya" vessels in the basal ganglia and Subjects and Methods other cerebral regions ("moyamoya" represents Nineteen patients with moyamoya disease something hazy like a puff of smoke in Japanese). (mean age: 36.8•}11.6 years) and 11 age-matched The age distribution of the patients forms two normal volunteers (mean age: 38.5•}6.8 years) peaks with different clinical features. In the were studied. All patients underwent complete juvenile type (onset less than 15 years of age), conventional angiography to determine the the initial symptoms are usually the results of angiographic stage for each hemisphere. The cerebral ischemia, while in the adult type (onset criteria for determining the angiographic stage peaking at 35 years), is are as follows:2' In stage 1, the carotid fork is the most common initial presentation. Although narrowed with no other abnormalities. In stage the reason for these distinct clinical subtypes has 2, the intracerebral main arteries are dilated due not been clarified, it is thought to be related to to stenosis of the ICA and the moyamoya vessels hemodynamic changes elicited by advance of the in the basal ganglia are rarely seen. In stage 3 , illness1 or other aging factors including the middle and anterior cerebral arteries (MCA .3 and ACA) are defected, and the basal moyamoya More than thirty years ago, Suzuki et al.2 is intensified. In stage 4, the occlusion of the demonstrated that the pattern of moyamoya ICA extend to the junction of the posterior vessels, as seen on carotid arteriography, changes communicating artery (Pcom), and the basal with the degree of ICA stenosis and the moyamoya becomes rough and small, resulting in development of various collateral vessels, and enlargement of the intraorbital moyamoya and classified it into 6 angiographic stages. At the the collateral vessels from the external carotid present time, the angiographic stages are widely artery (ECA). In stage 5, the basal moyamoya

A90 shrinks, and MCA and ACA disappear. In stage inhalation. Thirty minutes after the first LCBF 6, the entire cerebral circulation is maintained by measurement, some subjects inhaled 5% CO2 and transdural anastomosis from the ECA. 35% Xe gas, and the LCBF was measured in the According to these criteria, the cerebral same way. Then, the inhalation of Xe gas was hemispheres of the patients (n=38) were continued for a total of 15 minutes, and CT scans classified into the following stages; stage 1 (n=0), of the Xe-saturated brain were obtained to stage 2 (n=1), stage 3 (n=12), stage 4 (n=12), estimate the brain-blood partition coefficients (ă) stage 5 (n=10), and stage 6 (n=3). Of the 38, 37 of the Xe gas for each CT pixel. The changes in hemispheres in stages 3 to 6 (except for 1 in stage Xe gas concentration in the arterial blood were 2) were studied. In addition, 22 hemispheres of estimated by scanning the serial blood samples normal volunteers were analyzed. placed in a phantom. The LCBF in the patient group was measured The LCBF values for each CT pixel were at least three weeks after the last episode of calculated from the changes in Xe gas stroke to avoid any influence of the insult, such concentration in the brain and blood samples and as edema, using the Xenon (Xe) CT-CBF the k values, and LCBF map was reconstructed. autoradiographic method as described Then, mean LCBF values for various cerebral previously.5 The LCBF in the steady state was structures were evaluated. Regions with measured in all subjects, and the LCBF change infarction or hemorrhage on CT scan were during CO2 inhalation was tested in 13 patients excluded from the analysis. In the subjects who and 6 normal volunteers. In brief, after baseline inhaled CO, local cerebrovascular CO2 CT scans were obtained by a CT scanner (GE responsiveness (L-CO2R) was estimated as the

CT/T 8800), the subjects inhaled 35% stable Xe percentage change of LCBF per 1mmHg change gas for 3 minutes. During the inhalation of Xe of PaCO2 (•¢%LCBF/•¢PaCO2%/mmHg). An gas, arterial blood samples were drawn at 15 to analysis of variance (ANOVA) followed by 30-second intervals from a catheter placed in the Scheffe's test was employed to compare the left brachial artery, and second CT scans were values of LCBF or L-CO2R in these stages. obtained before the termination of the Xe gas

Figure 1 Mean local cerebral blood flow (LCBF) values (ml/100g/min) in 22 hemispheres of normal volunteers and in 37 hemispheres of patients with moyamoya disease. The hemispheres of the patients were classified as stages 3 to 6 according to Suzuki's criteria. The LCBF values in each region did not show any significant difference between the patients and normal volunteers. However, the LCBF at stage 6 was decreased in the anterior circulation, resulting in loss of "hyperfrontality." Abbreviation: FC, the frontal cortex; TC, the temporal cortex; OC, the occipital cortex; CAU, the caudate nucleus; PUT, the putamen; THA, the thalamus; FW, the frontal white matter; IC, the internal capsule; OW, the occipital white matter.

A91 Figure 2 Mean values for the local CO2 responsiveness (L-CO2R, •¢%LCBF/•¢PaCO2%/mmHg) in 12 hemispheres

of normal volunteers and in 25 hemispheres of patients with moyamoya disease. Compared with the LCBF, the L - CO2R in moyamoya desease tended to diminish through the angiographic stages. The L-CO2R in the FC at stage 5 was significantly lower than that in normal volunteers (p<0.01) or at stage 3 (p<0.05). The mean value for the FC at stage 6 was not shown because 2 of 3 hemispheres had infarct areas in the FC.

Figure 3 LCBF maps of a 13-year-old woman who complained of transient left-sided hemiparesis especially after hyperventilation. The angiographic stages of her right and left hemispheres were stage 5 and stage 3, respectively. The LCBF in the steady state (left panel) was well preserved in both hemispheres. During CO2 inhalation (right panel), however, the LCBF was markedly reduced in the FC, FW, and CAU (arrows) on the right hemisphere, which stage was 5, although the LCBF values in the other regions were increased. This suggests a steal phenomenon to the other regions where CO2responsiveness was maintained.

A92 Table 1 Collateral channels observed in each angiographical stage (Suzuki et al., 1968)

LCBF, which was observed in the normal volunteers, was lost in the patients of advanced Results stages as reported previously.4,7,11This finding The LCBF values during the steady state are indicates that blood supply through the vertebro shown in Figurel. At all stages of moyamoya -basilar system is important to maintain adequate disease, LCBF were not significantly different cerebral circulation in moyamoya disease in the from that in normal volunteers. At stage 6 , late stages. however, a small decrease in LCBF was observed Cerebrovascular response in moyamoya in the anterior part of the brain, i.e. the frontal disease to hypercapnia has been investigated as a

cortex (FC), the temporal cortex (TC) , the hemodynamic reserve capacity.7,9,11,13 In the caudate nucleus (CAU), and the putamen (PUT), present study, the L-CO2R was decreased resulting in loss of "hyperfrontality." On the together with progression of the angiographic other hand, in the posterior cortex and white stage in contrast to the LCBF under the resting matter (PC and PW), LCBF was well maintained state. The collateral channels may be even in the late stages. maximally dilated in normocapnic conditions and The L-CO2R in moyamoya disease tended to not be able to dilate further in response to diminish with progression of the stage in the hypercapnia. Moreover, the reduction in L anterior part of the brain (Figure 2). Especially CO2R was prominent in the anterior circulation, in the FC, the L-CO2R at stage 5 especially in the FC. This finding indicates that

(0.5•}0.5%/mmHg, mean•}SE) was significantly cerebrovascular reserve was markedly reduced in less than that in normal volunteers these regions, corresponding to the fact that the

(6.8•}1.1 %/mmHg, p<0.01) or at stage 3 patients frequently suffer from ischemic events in the anterior part of the brain (8.5•}2.2%/mmHg, p<0.05). The same . tendency was found in the TC, CAU, and PW, The occipital dominance in cerebral but this was not significant. hemodynamics of moyamoya disease may be related to the development of collateral vessels with the occlusive change of the ICA (Table 1). Discussion In the earlier stages, the basal moyamoya and the

In accordance with earlier investigations,4,6-9 intraorbital moyamoya flourish in the anterior LCBF values during the steady state in the adult part of the brain, which successfully supply enough blood to maintain the LCBF in the steady patients with moyamoya disease were comparable to those in the age-matched normal state. After stage 4, in which the Pcom is volunteers in the present study. Obara et al .10 occluded, collateral vessels from the posterior circulation, i.e. the posterior choroidal artery reported that the LCBF values of the disease , the were significantly higher than those in patients posterior pericallosal artery and the posterior with atherosclerotic occlusion of the major moyamoya, are then enlarged. In the last stage, th cerebral artery. Although a mild LCBF e transdural anastomosis and the posterior decrease may exist in pediatric patients,11,12 these collateral vessels come to supply the whole brain. findings suggested that the developed moyamoya Therefore, as the angiographic stage progresses, th vessels function well as collateral channels in the e functioning collateral channels shift from resting state. However, "hyperfrontality" of anterior to posterior part of the brain, and thus the compensatory capacity to a decline of LCBF in

A93 the anterior circulation may be decreased in the infarction due to moyamoya disease in young advanced stages. adults. Stroke 1988;19: 826-833 In conclusion, while the posterior circulation 7. Takashima S: Local cerebral blood flow and CO2 was well maintained until the advanced stages of responsiveness in patients with "Moyamoya Disease". Jpn. J. Stroke 1989; 11:407-415 moyamoya disease, the cerebrovascular reserve 8. Taki W, Yonekawa Y, Kobayashi A, et al: Cerebral in the anterior circulation became insufficient circulation and metabolism in adults' moyamoya although the collateral vessels from the vertebro disease PET study. Acta Neurochir -basilar system as well as the transdural 1989; 100:150-154 anastomosis develop. Common direct and 9. Kuwabara Y, Ichiya Y, Sasaki M, et al: Cerebral indirect bypass are effective in improving hemodynamics and metabolism in moyamoya cerebral hemodynamics in the territory of the disease a positron emission tomography study. MCA, but sometimes fail to prevent ischemic Clin Neurol Neurosurg 1997;99 Suppl 2:S74-78 events in the territory of the ACA. 10. Obara K, Fukuuchi Y, Kobari M, et al: Cerebral Revascularization surgery involving the frontal hemodynamics in patients with moyamoya disease lobe14,15 may be crucial to prevent stroke in and in patients with atherosclerotic occlusion of the major cerebral arterial trunks. Clin Neurol patients with moyamoya disease, particularly in Neurosurg 1997;99 Suppl 2:S86-89 the late stages. 11. Takeuchi S, Tanaka R, Ishii R, et al: Cerebral hemodynamics in patients with moyamoya disease. References A study of regional cerebral blood flow by the 133Xe inhalation method. Surg Neurol 1. Kudo T: Spontaneous occlusion of circle of Willis. 1985;23: 468-474 Neurology 1968; 18:485-496 12. Nariai T, Suzuki R, Matsushima Y, et al: Surgically 2. Suzuki J, Takaku A: Cerebrovascular "moyamoya" induced angiogenesis to compensate for disease. Disease showing abnormal net-like vessels hemodynamic cerebral ischemia. Stroke in base of brain. Arch Neurol 1968;20: 288-299 1994;25: 1014-1021 3. Houkin K, Yoshimoto T, Kuroda S, et al: 13. Ogawa A, Nakamura N, Yoshimoto T, et al: Angiographic analysis of moyamoya disease how Cerebral blood flow in moyamoya disease. Part 2: does moyamoya disease progress? Neurol Med Autoregulation and CO2 response. Acta Neurochir Chir (Tokyo) 1996;36: 783-787 1990;105: 107-111 4. Ogawa A, Yoshimoto T, Suzuki J, et al: Cerebral 14. Iwama T, Hashimoto N, Miyake H, et al: Direct blood flow in moyamoya disease. Part 1: revascularization to the anterior cerebral artery Correlation with age and regional distribution. territory in patients with moyamoya disease: report Acta Neurochir 1990; 105:30-34 of five cases. Neurosurgery 1998;42: 1157-1161 5. Kobari M, Fukuuchi Y, Shinohara T, et al: Local 15. Kuroda S, Houkin K, Kamiyama H, et al: cerebral blood flow and its response to intravenous Regional cerebral hemodynamics in levodopa in progressive supranuclear palsy. childhood moyamoya disease. Childs Nerv Comparison with Parkinson's disease. Arch Neurol Syst 1995; 11:584-590 1992;49: 725-730 6. Bruno A, Adams HP, Jr., Biller J, et al: Cerebral

A94