Moyamoya disease is a rare cerebrovascular occlusive disorder

Transcription

1 ORIGINAL RESEARCH O. Togao F. Mihaa T. Yoshiua A. Tanaka T. Noguchi Y. Kuwabaa K. Kaneko T. Matsushima H. Honda Ceebal Hemodynamics in Moyamoya Disease: Coelation between Pefusion-Weighted MR Imaging and Ceebal Angiogaphy BACKGROUND AND PURPOSE: In Moyamoya disease, the elationship between ceebal hemodynamics and angiogaphic findings has not been fully evaluated. The pupose of this study is to evaluate hemodynamics in Moyamoya disease with pefusion-weighted MR imaging (PWI) and ceebal angiogaphy. METHODS: Twenty patients with Moyamoya disease wee the subjects. Mean tansit time (MTT) deived fom PWI was calculated in the medial fontal lobes, the posteio fontal lobes, the occipital lobes, and the basal ganglia. Fom the angiogaphies, we classified the degees of intenal caotid atey (ICA) and posteio ceebal atey (PCA) stenoses as well as the degees of Moyamoya vessels and leptomeningeal anastomosis (LMA). MTT in each egion was compaed with the angiogaphic findings. RESULTS: MTT positively coelated with the degee of ICA stenosis in the medial fontal (P.01), posteio fontal (P.001), and occipital (P.001) lobes, as well as in the basal ganglia (P.001). MTT coelated with the degee of PCA stenosis in the medial fontal (P.001), posteio fontal (P.001), and occipital (P.001) lobes, as well as in the basal ganglia (P.001). MTT coelated with the degee of Moyamoya vessels in the medial fontal (P.05) and posteio fontal (P.01) lobes. A multivaiate analysis evealed that ICA and PCA stenoses and Moyamoya vessels wee independent factos that polonged MTT. CONCLUSION: Both ICA and PCA stenoses may influence oveall ceebal pefusion in Moyamoya disease. The development of Moyamoya vessels may indicate hemodynamic impaiment. Received May 2, 2005; accepted afte evision July 8. Fom the Depatment of Clinical Radiology (O.T., F.M., T.Y., A.T., T.N., Y.K., K.K., H.H.), Gaduate School of Medical Sciences, Kyushu Univesity, and the Depatment of Neuosugey (T.M.), Hamanomachi Hospital, Fukuoka, Japan. Addess coespondence to Futoshi Mihaa, MD, PhD, Depatment of Clinical Radiology, Gaduate School of Medical Sciences, Kyushu Univesity, Maidashi 3-1-1, Higashi-ku, Fukuoka , Japan. Moyamoya disease is a ae ceebovascula occlusive disode most often found among the Japanese. 1-3 This disease is angiogaphically defined as pogessive steno-occlusion of the bilateal intenal caotid ateies (ICAs) with chaacteistic abnomal vascula netwoks, so-called Moyamoya vessels, at the base of the bain. 1-3 Hemodynamics in Moyamoya disease ae athe complex. Steno-occlusive change can occu not only in the ICA but also in the posteio ceebal atey (PCA). 4,5 Leptomeningeal anastomosis (LMA) fom posteio ciculation and tansdual anastomosis fom the extenal caotid atey (ECA), as well as Moyamoya vessels, could develop to supply the ischemic bain. 1,2,5 The impotance of posteio ciculation in Moyamoya disease has ecently become ecognized. 6-8 Only a few epots have investigated the elationship between ceebal angiogaphy findings and hemodynamic status in Moyamoya disease. 6-8 By using dynamic susceptibility contast pefusion-weighted MR imaging (DSC- PWI), Yamada et al demonstated that PCA steno-occlusive change was significantly elated to ceebal hemodynamics in Moyamoya patients but ICA was not. 6 In Moyamoya disease, as in othe chonic steno-occlusive ceebovascula diseases, the steno-occlusive change in the main ceebal ateies deceases the ceebal pefusion pessue (CPP), and collateal ciculation maintains ceebal blood flow (CBF). 9,10 The mean tansit time (MTT) and its evese atio have been used as indices fo CPP in focal ischemia. 11,12 Quantitative measuement of the MTT by using a deconvolution algoithm can be a sensitive and eliable indicato of the ceebal pefusion eseve capacity, and it povides impotant infomation fo the management of patients with occlusive ceebovascula diseases We evaluated the elationship between angiogaphic findings and the egional MTT obtained by DSC-PWI in 20 patients with Moyamoya disease. Subjects and Methods Patients Fom Septembe 1995 to Mach 2004, a total of 29 patients with Moyamoya disease wee studied by both DSC-PWI and ceebal angiogaphy in ou adiology depatment. Of these 29 patients, we excluded 3 patients who had peviously undegone intacanial-extacanial evasculaization sugey and 3 patients fo whom the inteval between PWI and ceebal angiogaphy was longe than 3 months. One patient who had an extensive infacted lesion ( 8cmin diamete) in the left hemisphee on MR imaging was also excluded. The emaining 20 patients wee included in this study. The diagnosis of Moyamoya disease was made by the angiogaphic appeaance of ICA stenosis o occlusion and chaacteistic collateal ateies, socalled Moyamoya vessels, at the base of the bain. The patients included 7 men and 13 women anging in age fom 5 to 48 yeas (mean SD, yeas). The subjects clinical types wee as follows: tansient ischemic attack (TIA) type, 15 patients; hemohagic type, 4 patients; asymptomatic type, one patient. The inteval between PWI and ceebal angiogaphy was days (ange, 0 89 days). In all patients, clinical status had been stable thoughout the peiod between PWI and ceebal angiogaphy, and no new stoke o exacebation of TIA had occued in that peiod. Infomed consent was obtained fom each patient and his o he family. This study was appoved by ou institutional eview boad. The clinical featues and adiologic findings of the subjects ae summaized in Table 1. BRAIN ORIGINAL RESEARCH AJNR Am J Neuoadiol 27: Feb

2 Table 1: Clinical data and image findings Patient No./ Age (y)/sex Clinical Type Majo Clinical Symptoms Duation fom Onset of Disease Duation fom Last TIA Infact o Hemohage on MRL Image 1/5/M TIA Weakness of bilateal limbs 11 mo 2 mo Negative 2/6/F TIA Weakness of left limbs 1 y 3 mo Negative 3/7/F TIA Involuntay movement of ight hand 10 mo 1 w Negative 4/9/F TIA Weakness of ight limbs 2 y 3 mo Bilateal fontopaietal cotex 5/10/M TIA Weakness of ight hand 5 mo 5 mo Negative 6/10/F TIA Weakness of left hand 5 y 4 mo Negative 7/11/F TIA Weakness of ight limbs 1 y 1 mo Left fontal cotex, left basal ganglia 8/13/M TIA Weakness of bilateal lowe limbs 10 y 4 mo Right fontal white matte 9/15/M TIA Weakness of bilateal limbs 7 y 1 y Right fontal cotex 10/17/M TIA Weakness of bilateal limbs 5 y 1 w Negative 11/17/F TIA Weakness of ight limbs 3 y 1 mo Left fontotempoopaietal white matte 12/26/M H Headache 21 y Bilateal fontopaietal white matte 13/28/F TIA Weakness of bilateal limbs 14 y 1 mo Negative 14/31/F H Weakness of ight hand 3 y Left basal ganglia 15/34/F None Hemohage of bilateal ocula fundus 7 y Negative 16/36/F TIA Weakness of ight limbs 2 mo 2 mo Left fontal cotex, left basal ganglia 17/37/F TIA Weakness of ight hand 6 mo 1 mo Left fontopaietal white matte 18/41/F H Headache 4 mo Right tempoal subcotex 19/45/F H Headache 3 mo Right thalamus 20/48/M TIA Weakness of bilateal limbs 22 y 1 wk Left fontal cotex, ight basal ganglia Note. TIA indicates tansient ischemic attack; H, hemohagic. Dynamic Susceptibility Contast-Enhanced MR Imaging Potocol Dynamic susceptibility contast-enhanced MR imaging studies wee pefomed with a 1.5T MR unit, Magnetom Vision (Siemens Medical Systems, Elangen, Gemany), by using a single-shot gadient-echo type of echo-plana sequence with a standad head coil. The imaging paametes wee as follows: TE 54 milliseconds; flip angle 90 ; acquisition matix ; field of view mm; and section thickness 5 mm. Gadopentetate dimeglumine at 0.05 mmol/kg body weight was administeed as a bolus though the anticubital vein at a ate of 2 ml/s, followed by a 20-mL saline flush. 12,14 The contast agent was injected manually. A single axial plane at the level of the stiatum was scanned evey 1 second, fom 10 seconds befoe to 50 seconds afte contast injection. To allow fo equilibation of image magnetization, the fist 3 images wee excluded fom the data analysis. T1-weighted images, T2-weighted images, and MR angiogaphies wee also obtained. MR Imaging Data Analysis All DSC MR imaging data wee tansfeed to a Linux wokstation, and images of the pefusion paametes wee geneated by using the image-analytical pogam of D. View/Linux (Asahi Kasei Joho System Co., Tokyo, Japan). Fist, smoothing of the MR images was pefomed with a 5 mm 5 mm unifom smoothing kenel befoe the data analysis. The maximum concentation was calculated pixel by pixel befoe ateial input function (AIF) sampling. 16 To educe the patial volume effect, only pixels with a signal intensity intensity eduction of 99.5% of the maximum value wee extacted automatically in each egion of inteest. The mean concentation time cuve of these pixels was then used to detemine the AIF, which was obtained fom cicula egions of inteest (2 mm in diamete) placed manually on the ipsilateal middle ceebal atey (MCA) o PCA banch at the level of the basal ganglia. The MTT calculation is highly dependent on the choice of AIF. We theefoe measued AIF fom the MCA to calculate MTT in the anteio ciculation teitoy, and AIF fom the PCA to calculate MTT in the posteio ciculation teitoy (Fig 1A). 17 The signal intensity in each voxel in the bain paenchyma was conveted to tansvese elaxation ate. The signal intensity time cuves S(t) wee then conveted into concentation time cuves (C m (t)) on a pixel-by-pixel basis by using the following equation: 1) C m t In S t S 0 /TE, whee S 0 denotes the pecontast baseline signal intensity, S(t)isthe signal intensity at time t afte injection of the contast agent, and TE is the echo time. C m (t) was then deconvoluted with AIF[C a (t)] by using a fast Fouie tansfomation, and C(t) was obtained. This appoach has been descibed in detail by Wiestam et al 18 and Ostegaad et al. 19 The ceebal blood volume (CBV) was calculated by the following equation: 2) CBV k H 0 0 C m t dt C a t dt whee is the attenuation of the bain tissue ( 1.04 g/cm 3 ) and k H (0.73) is a facto coecting fo the diffeence in hematocit between the capillaies (H cap 0.25) and the lage ateial vessels (H at 0.45). 20 The CBF was calculated as 3) CBF k H C max, whee C max is the peak height of C(t). In addition to the AIF, the venous output function (VOF) was obtained fom egions of inteest placed on the supeio sagittal sinus, which is lage enough to allow fo VOF sampling without a patial volume effect. 21,22 The VOF is used to coect fo patial volume effects in the AIF due to the elatively small calibe of the MCA o PCA. This VOF coection was pefomed by multiplying the AIF with AIF/ VOF. 22 The MTT was calculated based on the atio of CBV/CBF in the PWI. 4) MTT CBV CBF., 392 Togao AJNR 27 Feb

3 Fig 1. Regions of inteest dawn in a PWI (A) and MTT map (B). The egions of inteest wee placed on cotical egions in the medial fontal lobe in the ACA distibution, in the posteio fontal lobe in the MCA distibution and in the occipital lobes in the PCA distibution as well as in the putamen in each ceebal hemisphee. To calculate the MTTs in the fontal lobes, as well as in the basal ganglia, an AIF (blue dot) was obtained fom an ipsilateal MCA banch to epesent anteio ciculation. In calculating MTT in the occipital lobe, an AIF (yellow dot) was obtained fom an ipsilateal PCA banch. This MTT map (B) was calculated with the AIF obtained fom the ight MCA banch. Kaneko et al demonstated that CBF and CBV obtained by the same method as ous wee significantly highe than those obtained by 15 O- positon-emission tomogaphy (PET). 14 In thei study, the nomalization factos between PWI and 15 O-PET wee estimated as 0.73 fo CBF and 0.49 fo CBV in the gay matte. 14 Theefoe, we also apply this nomalization facto in the calculation of MTT. The egions of inteest wee placed on cotical egions in the medial fontal lobes in the anteio ceebal atey (ACA) distibution, in the posteio fontal lobes in MCA distibution, and in the occipital lobes in PCA distibution as well as in the putamen in each ceebal hemisphee (Fig 1). The egions of inteest in the cotical egions wee mm in diamete, and those in the putamen wee mm. Aeas of infaction wee not included in these egions of inteest. Ceebal Angiogaphy The ceebal angiogaphy was pefomed by using a system of steeoscopic biplane digital subtaction angiogaphy (DFP-200A; Toshiba Medical Systems, Tokyo, Japan). In all 20 patients, angiogams of bilateal ICAs and ECAs, as well as unilateal o bilateal vetebal ateies (VAs), wee obtained by using the tansfemoal Seldinge s catheteization technique. A dose of about 10 ml of ioxaglate was injected into each vessel. Two boad-cetified adiologists (T.Y. and O.T.), who wee blinded to the patients clinical infomation and MR findings, met and discussed thei findings of the ceebal angiogaphies on the pinted films and eached a consensus fo each film. Accoding to the esulting consensus findings of ceebal angiogaphy of 40 hemisphees in the 20 patients, we applied a staging system to detemine the degees of steno-occlusive vascula change fo both anteio and posteio ciculation. The degee of the development of Moyamoya vessels at the base of the bain and that of leptomeningeal collateal vessels wee also evaluated. We classified steno-occlusive lesions of the distal ICA as well as those of the poximal egions of both ACA and MCA into 6 angiogaphic stages, adapted fom Suzuki s classification 1,2 : stage 1, mild naowing of the caotid atey bifucation only ( 50% eduction in diamete); stage 2, modeate to sevee stenosis of the caotid atey bifucation ( 50% eduction in diamete); stage 3, patial disappeaance of the teminal segment of the ICA and of the poximal egions of the ACA and MCA; stage 4, occlusion of the ICA bifucation, the main tunk of the ACA, and the MCA left faint taces within the Moyamoya vessels at the base of the bain; stage 5, occlusion of the ICA bifucation and absence of the ACA and MCA, but subtle anteogade blood flow to ACA and MCA banches though the basal Moyamoya vessels emained; stage 6, complete occlusion of the ICA and disappeaance of all ACA and MCA banches. We also staged PCA steno-occlusive lesions into 4 stages: stage 1, no occlusive change in the PCA; stage 2, mild stenosis ( 50% eduction in diamete) of the PCA with good visualization of distal banches; stage 3, modeate to sevee ( 50% eduction in diamete) stenosis of the PCA with poo visualization of distal banches; stage 4, occlusion of the PCA with almost no visualization of distal banches. The development of basal Moyamoya vessels was also gaded into 4 stages: stage 1, no Moyamoya vessels wee seen; stage 2, Moyamoya vessels wee localized in the aea aound the ICA bifucation, and each vessel was fine and had little contast; stage 3, Moyamoya vessels had intemediate extension and thickness; stage 4, Moyamoya vessels extended a geat deal and each one was thick and stongly opacified. The degee of development of the leptomeningeal collateal ciculation to ICA distibution fom the PCA and fom the posteio peicallosal ateies was classified into 4 gades: stage 1, no leptomeningeal collateal ciculation; stage 2, leptomeningeal cotical banches wee found in one of 3 lobes; stage 3, collateal ateies extended to 2 of 3 lobes; stage 4, leptomeningeal collateal ateies extended to all 3 lobes. Statistical Analysis We evaluated the elationship between each egion of inteest s MTT and each angiogaphic finding by using univaiate and multivaiate analyses. The univaiate analyses of the elationship between each egion of inteest s MTT and the angiogaphic stages of ICA, PCA, development of Moyamoya vessels, and LMA wee pefomed by using the Speaman ank coelation test. Multiple linea egession was used in the multivaiate analysis. The multivaiate analysis model included all angiogaphic factos (stages of ICA, PCA, Moyamoya vessels, and LMA), patient age, and clinical type (TIA type 1, hemohagic type 2). A P value.05 was consideed statistically significant in this study. All statistical analyses wee pefomed with StatView vesion 5.0 (SAS Institute Inc., Cay, NC). Results The stage distibution of stenotic o occlusive lesions in the ICA system of the 40 hemisphees was as follows: stage 1, 6 hemisphees; stage 2, 3 hemisphees; stage 3, 9 hemisphees; stage 4, 13 hemisphees; stage 5, 6 hemisphees; stage 6, 3 hemisphees. Of the 20 patients, 5 (25%) wee found to have steno-occlusive lesions in one o both PCAs. Eight PCAs (20%) of the 40 sides showed steno-occlusive change. The AJNR Am J Neuoadiol 27: Feb

4 Table 2: Mean MTT in each egion and ICA stage ICA Stage 1(n 6) 2 (n 3) 3 (n 9) 4 (n 13) 5 (n 6) 6 (n 3) Medial fontal Posteio fontal Occipital Basal ganglia Note. MTT indicates mean tansit time; ICA, intenal caotid atey. *, P.05;, P.01;, P.001; based on Speaman ank coelation test. Table 3: Mean MTT in each egion and PCA stage PCA Stage 1(n 32) 2 (n 1) 3 (n 5) 4 (n 2) Medial fontal Posteio fontal Occipital Basal ganglia Note. MTT indicates mean tansit time; PCA, posteio ceebal atey. *, P.05;, P.01;, P 0.001; based on Speaman ank coelation test. Table 4: Mean MTT in each egion and stage of Moyamoya vessels Moyamoya Vessels 1(n 9) 2 (n 22) 3 (n 5) 4 (n 4) Medial fontal * Posteio fontal Occipital Basal ganglia Note. MTT indicates mean tansit time. *, P.05;, P.01,, P.001; based on Speaman ank coelation test. Table 5: Mean MTT in each egion and stage of leptomeningeal anastomosis LMA 1(n 4) 2 (n 8) 3 (n 8) 4 (n 20) Medial fontal Posteio fontal Occipital Basal ganglia Note. MTT indicates mean tansit time; LMA, leptomeningeal anastomosis. *, P.05;, P.01,, P.001; based on Speaman ank coelation test. stage distibution of the 40 PCAs of the 20 patients was as follows: stage 1, 32 ateies; stage 2, 1 atey; stage 3, 5 ateies; and stage 4, 2 ateies. The gade distibution of the Moyamoya vessels was as follows: stage 1, 9 hemisphees; stage 2, 22 hemisphees; stage 3, 5 hemisphees; and stage 4, 4 hemisphees. The gade distibution of the leptomeningeal vessels was as follows: stage 1, 4 hemisphees; stage 2, 8 hemisphees; stage 3, 8 hemisphees; stage 4, 20 hemisphees. The esults of the univaiate analysis of the PWI data and angiogaphic findings wee as follows: MTT significantly coelated with the degee of ICA stenosis in the medial fontal ( 0.52; P.01), posteio fontal ( 0.64; P.001), and occipital ( 0.58; P.001) lobes as well as in the basal ganglia ( 0.58; P.001) (Table 2). MTT also coelated with the degee of PCA stenosis in the medial fontal ( 0.56; P.001), posteio fontal ( 0.60; P.001), and occipital ( 0.64; P.001) lobes as well as in the basal ganglia ( 0.53; P.001) (Table 3). The highe the degee of ICA and PCA stenosis, the longe MTT was. MTT obtained in the medial fontal ( 0.35; P.05) and posteio fontal ( 0.47; Table 6: Standad egession coefficient elating to MTT in each egion Medial Fontal Posteio Fontal Occipital Basal Ganglia ICA PCA Moyamoya vessels * 0.28 LMA * 0.23 Age * 0.38* Clinical type R Note. MTT indicates mean tansit time; ICA, intenal caotid atey; PCA, posteio ceebal atey; LMA, leptomeningeal anastomosis. *, P.05;, P.01;, P.001; based on multiple linea egession analysis. P.01) lobes was significantly positively coelated with the degee of Moyamoya vessels (Table 4). Thee was no coelation between the degee of LMA and MTT obtained in each egion (Table 5). The esults of the multiple linea egession analysis ae summaized in Table 6. It was evealed that both the ICA and PCA 394 Togao AJNR 27 Feb

5 Fig 2. A 10-yea-old boy with Moyamoya disease who pesented with weakness of the ight uppe extemity. Right intenal caotid ateiogam in posteio-anteio pojection (A) shows that the ight poximal anteio ceebal atey is mildly stenotic (black aow). No steno-occlusive change is found in the ICA o MCA. The ight ICA is in stage 1. No Moyamoya vessels ae seen. Left intenal caotid ateiogams in posteio-anteio pojection (B) show that the ight distal ICA and the anteio and MCAs ae occluded (white aow). The left ICA is in stage 3. Maked Moyamoya vessels ae seen, and peipheal banches of the anteio ceebal atey and MCA ae opacified via the collateal vessels (black aowhead). On the left vetebal ateiogam in posteio-anteio pojection (C), no steno-occlusive change is found in the bilateal posteio ceebal atey. The bilateal posteio ceebal atey is in stage 1. Leptomeningeal collateal vessels fom the left posteio ceebal atey to the anteio ciculation ae seen (white aowhead). MTT map (D) shows the aeas of polonged mean MTT in the left fontal and tempoal lobes. This map was calculated with the AIF obtained fom a ight MCA banch. stages wee independent significant factos that influenced MTT in each of the 4 egions. The degee of Moyamoya vessels was a significant facto fo MTT in the medial and posteio fontal lobes, and occipital lobes. The degee of LMA was a negative significant facto fo MTT in the occipital lobe. The patient s age was a positive significant facto fo MTT in the occipital lobe and basal ganglia. Clinical type was not significant in any egion. Multicollineaity was not obseved in any multivaiate analysis. Two epesentative cases ae shown in Figs 2 and 3. Discussion Ceebal hemodynamics in Moyamoya disease have been studied by using PET. Accoding to those epots, patients with Moyamoya disease have educed CBF and CO 2 esponse, and inceased CBV, MTT, and oxygen extaction faction (OEF). 23,24 These hemodynamic changes in Moyamoya disease wee explained by deceases in CPP and associated vasodilation. Schumann et al 11 and Gibbs et al 25 demonstated that, in patients with chonic ceebovascula occlusive disease, CBF/CBV, an invese equation of MTT, eflects the local CPP and is coelated with OEF. Kuwabaa et al epoted that MTT was polonged in both pediatic and adult patients with Moyamoya disease and that MTT was elated to CO 2 esponse. 23,24,26 By using PET, Taki et al found deceased CBF/CBV in patients with Moyamoya disease. 10 PET-obtained MTT has been consideed a eliable paamete fo evaluating the hemodynamic status of patients with Moyamoya disease. MTT obtained by PWI has also been used to evaluate the hemodynamic status of patients with occlusive ceebovascula diseases. 12,14 Thee ae 2 majo methods of measuing MTT by using PWI. One is the deconvolution method, 13-16,18,19 and the othe is the nomalized fist moment method. 12,14,15,27 MTT values obtained by deconvolution analysis wee epoted to be eliable paametes of ceebal hemodynamics. Kaneko et al epoted that MTT obtained by the deconvolution method coelated well with the MTT obtained by 15 O-PET. 14 The deconvolution method equies the AIF obtained fom the majo ceebal ateies. 19 A cucial poblem in Moyamoya disease, howeve, is the difficulty of obtaining the AIF because of its extensive obstuction aound the cicle of Willis. The nomalized fist moment method has been the classic technique to measue MTT by using PWI without AIF; howeve, MTT detemined by the nomalized fist moment method is distinct fom the ideal MTT, because this method is influenced by both the ideal MTT and the fist moment of the AIF. 12 In the pesent study, the univaiate analysis showed that MTT in the medial and posteio fontal lobes, and occipital lobes, as well as that in the basal ganglia, significantly positively coelated with the degees of both ICA and PCA steno-occlusion. Multiple linea egession analysis evealed that the degees of both ICA and PCA stages wee the independent significant factos that detemined the local hemodynamic status in all 4 egions of inteest. We speculate that, in Moyamoya disease, ceebal pefusion deceases as both ICA and PCA AJNR Am J Neuoadiol 27: Feb

6 Fig 3. A 45-yea-old woman with Moyamoya disease who pesented with ight thalamic hemohage. Right intenal caotid ateiogams in posteio-anteio pojection (A) demonstated that the ight ICA, anteio ceebal atey, and poximal MCA ae occluded (lage black aow). The ight ICA is in stage 4. Slightly developed Moyamoya vessels ae seen aound the distal ICA (black aowhead). The distal MCA is visualized though the Moyamoya vessels. Left common caotid ateiogam in lateal pojection (B) shows that the left ICA is completely occluded in its poximal potion (white aowhead). The left intenal ICA is in stage 6. Left vetebal ateiogam in posteio-anteio pojection (C) shows sevee stenosis in the bilateal posteio ceebal ateies (white aows). Moyamoya vessels ae seen in the posteio egion of the bain base. Peipheal banches of bilateal posteio ceebal ateies ae visualized though Moyamoya vessels. The bilateal posteio ceebal atey is in stage 3. MTT map (D) demonstates extensive aeas of polonged MTT in bilateal ceebal hemisphees. This map was calculated with the AIF obtained fom a ight MCA banch. stenoses pogess. The impotance of posteio ciculation in Moyamoya disease has been epoted. 6-8 Posteio ciculation is significant because PCA is thought to play an impotant collateal ole in patients with ICA steno-occlusion, including Moyamoya patients. 6-8 In Moyamoya disease, PCA involvement would occu at the elatively high ICA stages. 6-8 Theefoe, as the disease pogesses, oveall ceebal pefusion would decease because both anteio and posteio ciculation deteioates. Ou esults exemplify this theoy. Using PWI, Yamada et al demonstated that the R2* peak value atio in the ceebal hemisphee deceased and that the R2* peak time atio inceased significantly with the pogession of PCA steno-occlusion. 6 Thei esults ae in accodance with ou study. In the studies of Yamada et al, howeve, no coelation was obseved between these pefusion paametes and the degee of ICA steno-occlusion, which contadicts ou esults. One of the easons fo the discepancy between ou study and theis could be the diffeence in the paametes used in the evaluation. MTT obtained by the deconvolution method, which we used in ou study, would be moe sensitive to hemodynamic status than the simple paametes used in thei study. Seveal epots have evaluated the elationships among angiogaphic findings and ceebal hemodynamics in Moyamoya disease. Yamada et al also epoted by using single-photon emission-ct that elative CBF in Moyamoya patients deceased with the pogession of steno-occlusive lesions of the PCA. 7 Using the xenon CT-CBF method, Nogawa et al demonstated that CBF was deceased in the anteio pat of the bain at stage 6, the most advanced stage in Suzuki s classification, and that the CO 2 esponse in the anteio pat of the bain tended to diminish with pogession though Suzuki s stages. 28 Suzuki s staging system, howeve, focuses mainly on the ICA steno-occlusive change and so cannot be used to evaluate the steno-occlusive change of posteio ciculation. 1,2 Mugikua et al epoted that pogessive changes in both the anteio and posteio ciculation ae associated with the distibution and extent of ceebal infaction on CT o MR. 8 Although the elationship between the infaction and the ceebal hemodynamic status may not be simple, the esults of Mugikua et al seem to be consistent with ous. In the pesent study, the MTT was significantly polonged with the development of Moyamoya vessels in the medial and posteio fontal lobes, and occipital lobes. Piao et al epoted that the extensive development of basal Moyamoya vessels is a sign of sevee hemodynamic impaiment in adult patients with ischemic Moyamoya disease. 9 Moyamoya vessels would develop to compensate fo the misey ceebal pefusion. We suppose the eason why LMA is insignificant is that, as PCA stenosis advances, the LMA fom posteio to anteio ciculation deceases, and thus a linea elationship between LMA and MTT would not be obtainable. Pefusion-weighted MR imaging in Moyamoya disease has potential intinsic limitations. The pesence of steno-occlusion of the main ceebal ateies and collateal vessels always leads to the delay and dispesion of the bolus of the contast 396 Togao AJNR 27 Feb

7 agent. Calamante et al epoted that, in Moyamoya patients, the maps geneated with the use of deconvolution can be misleading, esulting in oveestimation of the MTT. 29 They mentioned that the main poblem is the impossibility of measuing the tue AIF, which is geneally estimated fom a lage atey that, in pactice, may be distant fom the tissue that is being studied. We used the AIF obtained fom an MCA o PCA banch at the level of the basal ganglia, which is fine and naow in patients with Moyamoya disease, and thus the AIF would be susceptible to signal intensity noise and technical eo. A new method that can compensate fo o esolve this poblem is desiable. The othe poblem is that, in chonic steno-occlusive disease, the eliability of MTT obtained by the deconvolution method has not been fully evaluated in Moyamoya disease, though it has been confimed to a cetain extent in chonic steno-occlusive disease. 13,14 We should have used the atio o the diffeence between affected egion and nomal tissue as a paamete that estimates egional ceebal pefusion. In Moymoya disease, howeve, because both ceebal hemisphees ae extensively affected, we could not obtain such values. Ou study has seveal othe limitations. Hemodynamic paametes ae known to change with age. 23 In Moyamoya disease, clinical pesentation diffes between adults and pediatic patients. 23 In the esults of ou multivaiate analysis, MTT tended to be polonged in olde patients (Table 6). It would have been bette if this study had discussed pediatic and adult patients in sepaate goups. Anothe limitation is that the injection of the contast agent fo PWI was pefomed manually, though at a elatively constant ate that was easily epoducible. Also, tansdual collateal vessels fom the extenal caotid atey system, which play an impotant ole in collateal ciculation in Moyamoya disease, wee not evaluated angiogaphically. This was because tansdual anastomosis is so multifaious that it was not easy to adapt a staging system to these angiogaphic findings. Conclusion The univaiate and multivaiate analyses demonstated that the degees of both ICA and PCA steno-occlusive change wee significant factos in the polongation of MTT in all 4 ceebal egions measued in patients with Moyamoya disease. These esults suggested that both ICA and PCA stenoses influence the global ceebal pefusion in Moyamoya disease. In the multivaiate analysis, the degee of Moyamoya vessels was a significant facto in the polongation of MTT in the medial and posteio fontal and occipital egions. The development of Moyamoya vessels would indicate hemodynamic impaiment. Acknowledgments This wok was suppoted in pat by a gant-in-aid fo scientific eseach (C ) fom the Japan Society fo the Pomotion of Science. Refeences 1. Suzuki J, Takaku A. Ceebovascula moyamoya disease: disease showing abnomal net-like vessels in base of bain. Ach Neuol 1969;20: Suzuki J, Kodama N. Moyamoya disease: a eview. Stoke 1983;14: Nishimoto A, Takeuchi S. Abnomal ceebovascula netwok elated to the intenal caotid ateies. J Neuosug 1968;29: Miyamoto S, Kikuchi H, Kaasawa J, et al. Study of the posteio ciculation in moyamoya disease: clinical and neuoadiological evaluation. J Neuosug 1984;61: Yamada I, Himeno Y, Suzuki S, et al. Posteio ciculation in moyamoya disease: angiogaphic study. Radiology 1995;197: Yamada I, Himeno Y, Nagaoka T, et al. Moyamoya disease: evaluation with diffusion-weighted and pefusion echo-plana MR imaging. Radiology 1999; 212: Yamada I, Muata Y, Umehaa I, et al. SPECT and MRI evaluations of the posteio ciculation in moyamoya disease. J Nucl Med 1996;37: Mugikua S, Takahashi S, Higano S, et al. The elationship between ceebal infaction and angiogaphic chaacteistics in childhood moyamoya disease. AJNR Am J Neuoadiol 1999;20: Piao R, Oku N, Kitagawa K, et al. Ceebal hemodynamics and metabolism in adult moyamoya disease: compaison of angiogaphic collateal ciculation. Ann Nucl Med 2004;18: Taki W, Yonekawa Y, Kobayashi A, et al. Ceebal ciculation and metabolism in adults moyamoya disease: PET study. Acta Neuochi (Wien) 1989;100: Schumann P, Touzani O, Young AR, et al. Evaluation of the atio of ceebal blood flow to ceebal blood volume as an index of local ceebal pefusion pessue. Bain 1998;121: Mihaa F, Kuwabaa Y, Tanaka A, et al. Reliability of mean tansit time obtained using pefusion-weighted MR imaging; compaison with positon emission tomogaphy. Magn Reson Imaging 2003;21: Kikuchi K, Muase K, Miki H, et al. Quantitative evaluation of mean tansit times obtained with dynamic susceptibility contast-enhanced MR imaging and with (133)Xe SPECT in occlusive ceebovascula disease. AJR Am J Roentgenol 2002;179: Kaneko K, Kuwabaa Y, Mihaa F, et al. Validation of the CBF, CBV, and MTT values by pefusion MRI in chonic occlusive ceebovascula disease: a compaison with 15 O-PET. Acad Radiol 2004;11: Lythgoe DJ, Ostegaad L, William SC, et al. Quantitative pefusion imaging in caotid atey stenosis using dynamic susceptibility contast-enhanced magnetic esonance imaging. Magn Reson Imaging 2000;18: Rempp KA, Bix G, Wenz F, et al. Quantification of egional ceebal blood flow and volume with dynamic susceptibility contast-enhanced MR imaging. Radiology 1994;193: Soensen AG, Reime P. Ceebal MR pefusion imaging: pinciples and cuent applications. Stuttgat and New Yok: Thieme;2000: Wiestam R, Andeson L, Ostegaad L, et al. Assessment of egional blood flow by dynamic susceptibility contast MRI using diffeent deconvolution techniques. Magn Reson Med 2000;43: Ostegaad L, Weiskoff RM, Cheste DA, et al. High esolution measuement of ceebal blood flow using intavascula tace bolus passages. Pat I. Mathematical appoach and statistical analysis. Magn Reson Med 1996; 36: Beeczki D, Wei L, Otsuka T, et al. Hypecapnia slightly aises blood volume and sizably elevates flow velocity in bain micovessels. Am J Physiol 1993;264: Lin W, Celik A, Dedeyn C, et al. Quantitative measuement of ceebal blood flow in patients with unilateal caotid atey occlusion: a PET and MR study. J Magn Reson Imaging 2001;14: Hagen T, Batylla K, Piepgas U. Coelation of egional ceebal blood flow measued by stable xenon CT and pefusion MRI. J Comput Assist Tomog 1999;23: Kuwabaa Y, Ichiya Y, Otsuka M, et al. Ceebal hemodynamic change in the child and the adult with moyamoya disease. Stoke 1990;21: Kuwabaa Y, Ichiya Y, Sasaki M, et al. Ceebal hemodynamics and metabolism in moyamoya disease: a positon emission tomogaphy study. Clin Neuol Neuosug 1997;99(suppl 2):S Gibbs JM, Wise RJ, Leendes KL, et al. Evaluation of ceebal pefusion eseve in patients with caotid-atey occlusion. Lancet 1984;11: Kuwabaa Y, Ichiya Y, Sasaki M, et al. Response to hypecapnia in moyamoya disease: ceebovascula esponse to hypecapnia in pediatic and adult patients with moyamoya disease. Stoke 1997;28: Nighoghossian N, Bethezene Y, Philippon B, et al. Hemodynamic paamete assessment with dynamic susceptibility contast magnetic esonance imaging in unilateal symptomatic intenal caotid atey occlusion. Stoke 1996;27: Nogawa S, Fukuuchi Y, Kobai M, et al. Local ceebal hemodynamic changes though the angiogaphic stages of moyamoya disease. Keio J Med 2000; 49(suppl 1):A Calamante F, Ganesan V, Kikham FJ. MR pefusion imaging in Moyamoya syndome: potential implications fo clinical evaluation of occlusive ceebovascula disease. Stoke 2001;32: AJNR Am J Neuoadiol 27: Feb