Assessment of Moyamoya Disease Using Multidetector Row Computed Tomography

Assessment of Moyamoya Di

Toshiya Sugino, MD,* Takeshi Mikami, MD,* Shunya Ohtaki, MD,*

Tohru Hirano, MS,† Satoshi Iihoshi, MD,* Kiyohiro Houkin, MD,‡

and Nobuhiro Mikuni, MD*

From the *Department

Sapporo, Japan; †Division

Hospital, Sapporo, Jap

Hokkaido University Gra

Received October 4, 2

accepted January 18, 201

Address corresponden

Neurosurgery, Sapporo M

ku, Sapporo 060-8543, Ja

1052-3057/$ - see front

� 2013 by National Str

doi:10.1016/j.jstrokece

644

The recent introduction of multidetector row computed tomography (MDCT) scan-

ners has enabled high-resolution 3-dimensional reconstruction. The purpose of this

study was to establish a method to evaluate moyamoya disease using computed

tomography angiography (CTA), specifically MDCT. Twenty-four patients (48 sides

total) with moyamoya disease diagnosed by magnetic resonance angiography

(MRA) were evaluated by means of CTA using MDCT by 3 independent observers,

and the resulting 144 sideswere analyzed. CTA andMRAwere compared in terms of

the steno-occlusive changes exhibited in each vessel. CTA andMRA scores were as-

signed on the basis of the severity of occlusive changes in the internal carotid artery,

middle cerebral artery, anterior cerebral artery, and posterior cerebral artery. CTA

scores were significantly correlated with MRA scores (P , .0001), and the 2 scores

were in complete agreement in 57 sides (39.6%). The mean CTA score was signifi-

cantly lower than themeanMRA score (P,.0001). Comparedwith CTA,MRA over-

estimated occlusion in 115 of the 576 vessels assessed. The mean MRA score was

significantly higher in the overestimation group than in the good correlation group

(P , .0001). CTA had a significantly higher rate of detection of moyamoya-affected

vessels (P 5 .0001). Our data indicate that CTA using MDCT is a more reliable tech-

nique than MRA for diagnosing moyamoya disease. The ability to perform CTA

quickly is a significant benefit for patients with moyamoya disease, particularly in

pediatric and emergency cases. Key Words: Magnetic resonance angiography—

scoring system—cerebrovascular disease.

� 2013 by National Stroke Association

Magnetic resonance angiography (MRA) is a widely

used alternative to conventional angiography for diagnos-

ingmoyamoya disease.1,2 In fact, since 1994, the diagnostic

guidelines for moyamoya disease have specified MRA as

the definitive diagnostic technique.3-5 In this context,

ofNeurosurgery, SapporoMedicalUniversity,

of Radiology, Sapporo Medical University

an; and ‡Department of Neurosurgery,

duate School of Medicine, Hokkaido, Japan.

011; revision received December 12, 2011;

2.

ce to Takeshi Mikami, MD, Department of

edical University, South 1, West 16, Chuo-

pan. E-mail: [email protected].

matter

oke Association

rebrovasdis.2012.01.014

Journal of Stroke and Ce

MRA has been acknowledged as a reliable diagnostic

tool, with high sensitivity and specificity made possible

by the remarkable development of magnetic resonance

imaging (MRI) technology.6-12 Similarly, the clinical

relevance of computed tomography angiography (CTA)

has been demonstrated since 1996.13 A particularly useful

form of CTA is multidetector row computed tomography

(MDCT), the rapid spread ofwhich has enhanced the diag-

nosis of moyamoya disease by enabling the use of CTA

to evaluate vascular lesions.14-18 This diagnostic tool is

especially helpful in patients presenting to the emergency

room with suspected vascular abnormalities.19 Both of

these newer imaging techniques are in widespread use to-

day; in a recent randomized trial, carotid artery stenosis

was diagnosed using MRA or CTA only, without the use

of conventional angiography.20

rebrovascular Diseases, Vol. 22, No. 5 (July), 2013: pp 644-649

Table 1. Summary of the CTA and MRA scoring system

ICA

Normal 0

Stenosis of C1 1

Discontinuity of C1 signal 2

Invisible 3

MCA

ASSESSMENT OF MOYAMOYA DISEASE USING MDCT 645

MRA has several drawbacks, including the need for

specialized equipment and the relatively long scanning

time. For these reasons, it would be useful if CTA could

be used in its place. In this study, we investigated the pos-

sibility of evaluating moyamoya disease through CTA

with MDCT by comparing detailed scores obtained by

this technique with those obtained by MRA.

Normal 0

Stenosis of M1 1

Discontinuity of M1 signal 2

Invisible 3

ACA

Normal A2 and its distal 0

A2 and its distal signal decrease or loss 1

Invisible 2

PCA

Normal P2 and its distal 0

P2 and its distal signal decrease or loss 1

Invisible 2

Total 0�10

Patients and Methods

Subjects

Twenty-four patients (48 sides total) were preopera-

tively examined by both CTA with MDCT and MRA

over the course of 3 years between January 2008 and

December 2010. All 24 patients had been diagnosed

with moyamoya disease or unilateral moyamoya disease

by means of MRI and MRA. The mean patient age was

30.7 6 16.5 years (range, 3-60 years, with 7 children

aged ,15 years and 17 adults aged $15 years.

MDCT

CTA examination was performed using a LightSpeed

VCT 64-row MDCT scanner (GE Healthcare, Waukesha,

WI). The following 3-dimensional (3D) CTA scanning pa-

rameters were used: tube voltage, 120 kV; collimation,

0.625 mm 3 32; slice thickness, 0.625 mm; slice interval,

0.312 mm. A total of 1-1.2 mL/kg iopamidol (Iopamiron

300; Bayer Healthcare, Leverkusen, Germany), a low-

osmolar iodinated contrast material, was administered

i.v. using a bolus tracking method via an 18- to 20-

gauge catheter positioned in an antecubital vein. The in-

formation was transferred to a workstation (Ziostation;

Ziosoft, Tokyo, Japan), and 3D multicolored images

were produced. Volume rendering (VR) and maximum-

intensity projection (MIP) were applied as postprocessing

techniques to aid evaluation.

MRA

MRI examination was performed using a Signa Excite

11 MRI system (GE Healthcare). MRAwas obtained using

the 3D time-of-flight technique with a 3D spoiled

gradient-recalled echo sequence using 3.0-T magnet

machines. The imaging parameters were as follows: flip

angle, 15 degrees; repetition time, 26 ms; echo time, 3.2

ms; bandwidth, 22.73 kHz; field of view, 200 3 180 mm;

scan thickness, 0.8 mm; locations per slab, 124; number

of slabs, 1; matrix, 384 3 320; acquisition, 1; imaging

time, 7 minutes, 35 seconds. VR and MIP**3** were ap-

plied as postprocessing techniques to aid evaluation.

Scoring

Houkin et al21 defined the MRA score as the total of the

scores for all 4 main cerebral arteries and described the

scoring method in detail. This scoring system correlates

well with the grading system originally established by

Suzuki and coworkers22-24 for moyamoya disease,

which describes the clinical stages detectable through

conventional angiography. In the present study, an MRA

score based on the severity of steno-occlusive changes

was assigned to each of the following: C1 portion of the

internal carotid artery (ICA), M2 portion of the middle ce-

rebral artery (MCA), A1 portion of the anterior cerebral

artery (ACA), and P2 portion of the posterior cerebral

artery (PCA) (Table 1). The MRA score was defined as

the total points assigned to these 4 main cerebral arteries.

The minimum possible score was 0, and the highest was

10 (ICA3 1 MCA3 1 ACA2 1 PCA2 5 10). In this study,

the definition of MRA score was adapted for CTA, and

CTA scores thus calculated were compared with MRA

scores. The CTA and MRA scores were determined on

the 2 sides of each subject by 3 independent observers

(T.S., T.M., and S.O.). All of the data gathered by these 3

observers were used for the comparative analysis.

Evaluation of Steno-Occlusive Changes in Each Vessel

and Moyamoya Vessels

We evaluated each patient with respect to the following

2 points. First, the MRA and CTA scores reflecting steno-

occlusive changes in the same vessel of the same patient

were compared. Each vessel evaluated was then classified

according to whether using the MRA score as opposed to

the CTA score resulted in good correlation, overestima-

tion, or underestimation of steno-occlusive changes.

Each vessel (IC, MCA, ACA, and PCA) was examined

to assess whether the evaluation technique made a differ-

ence, and the mean MRA scores for the good correlation

group were compared with those for the overestimation

group or the underestimation group. Second, the

Figure 1. CTA and MRA in a 14-year-old male. This figure illustrates

a typical case of overestimation of stenosis and occlusion using MRA. The

steno-occlusive changes in the right ICA and MCA appear more severe on

MRA than on CTA (arrows).

Figure 2. Correlation between total CTA scores and total MRA scores.

T. SUGINO ET AL.646

visualization of moyamoya-affected vessels by MRA and

by CTA were compared. The presence of moyamoya-

affected vessels was evaluated by MIP imaging of MRA

and CTA at the level of the basal cistern and the basal gan-

glia. The evaluation was made based on the presence of at

least 2 small abnormal vessels. Evaluation of the moya-

moya vessels by MRA and CTA was also assessed by

the same blinded observers.

Statistical Analysis

Data are expressed as mean6 standard deviation. Sim-

ple regression analysis was used to confirm the correla-

tion between MRA and CTA scores, with a correlation

coefficient.0.7 indicating strong correlation. In addition,

the evaluation of moyamoya vessels by MRA and by CTA

was analyzed using the c2 test for independence. All sta-

tistical analyses were performed using the StatView soft-

ware package (SAS Institute Inc., NC). A P value , .05

was considered statistically significant.

Results

All of the lesions could be evaluated by means of CTA.

No patient exhibited systemic complication as a result of

infection caused by the contrast material used for CTA. In

one case diagnosed as unilateral moyamoya disease by

MRA, CTA revealed stenosis of C1 and M1 on the other

side; thus, this case was diagnosed as bilateral by means

of CTA. In one case diagnosed as bilateral moyamoya dis-

ease due to slight stenosis detected in the right C1 by

MRA, CTA showed that the right C1 was normal, and

thus the case was diagnosed as unilateral by means of

CTA (Fig 1).

Scoring

The mean CTA score was 3.79 6 2.01, and the mean

MRA score was 4.44 6 2.21, a statistically significant dif-

ference (P , .0001). Figure 2 depicts the correlation be-

tween the CTA and MRA scores in the 24 patients (48

sides) by the 3 observers. CTA scores were significantly

correlated with MRA scores (P , .0001; y 5 0.81x 1

0.17; r5 0.89), andwere in complete agreement in 57 sides

(39.6%), although CTA scores tended to be higher than

MRA scores.

Evaluation of Steno-Occlusive Changes in the Carotid

Fork and Moyamoya Vessels

Table 2 compares the severity of the steno-occlusive

changes of each vessel as evaluated by MRA and CTA

in each vessel of each patient. Of the 576 total vessels eval-

uated, 436 (75.7%) appeared to have the same extent of

steno-occlusive change when visualized through both

MRA and CTA (good correlation group). The remaining

vessels were divided between the 2 possible types of

poor correlation between MRA and CTA scores: overesti-

mation byMRA and underestimation byMRA. In cases of

overestimation, MRA suggested more severe steno-

occlusive changes than were detected by CTA; in cases

of underestimation, MRA suggested less severe changes

than CTA. Our series included 25 vessels (4.3%) in the un-

derestimation group and 115 vessels (20.0%) in the over-

estimation group. Compared with the good correlation

group, the mean MRA score was significantly higher in

the overestimation group (0.87 6 1.00 vs 2.18 6 0.71;

P , .0001) and significantly lower in the underestimation

group (0.876 1.00 vs 0.406 0.71; P, .0001). A typical case

Table 2. Correlation of MRA score with CTA score in terms of severity of steno-occlusive changes

Location Assessment Sides, n (%)

Score (mean 6 SD)

P valueMRA CTA

ICA (n 5 144) Good correlation 104 (72.2) 1.52 6 1.00 1.52 6 1.00 -

Overestimation 32 (22.2) 2.09 6 0.59* 1.03 6 0.54 .0089

Underestimation 8 (5.6) 0.13 6 0.35* 1.13 6 0.35 .0148

MCA (n 5 144) Good correlation 80 (55.6) 1.63 6 1.01 1.63 6 1.01 -

Overestimation 58 (40.3) 2.50 6 0.66* 1.45 6 0.65 ,.0001

Underestimation 6 (4.2) 1.33 6 0.82 2.33 6 0.82 NS

ACA (n 5 144) Good correlation 121 (84.0) 0.60 6 0.76 0.60 6 0.76 -

Overestimation 19 (13.2) 1.73 6 0.45* 0.74 6 0.45 ,.0001

Underestimation 4 (2.8) 0.00 6 0.00 1.25 6 0.50 NS

PCA (n 5 144) Good correlation 131 (91.0) 0.15 6 0.38 0.15 6 0.38 -

Overestimation 6 (4.2) 1.00 6 0.00 0.00 6 0.00 NS

Underestimation 7 (4.9) 0.14 6 0.38 1.14 6 0.38 NS

Total (n 5 576) Good correlation 436 (75.7) 0.87 6 1.00 0.87 6 1.00 -

Overestimation 115 (20.0) 2.18 6 0.71* 1.14 6 0.69 ,.0001

Underestimation 25 (4.3) 0.40 6 0.71* 1.44 6 0.71 ,.0001

Abbreviation: NS, not significant.

In each vessel, mean MRA score was compared between the good correlation group and the overestimation group and the underestimation

group.

*Indicates significant difference.

ASSESSMENT OF MOYAMOYA DISEASE USING MDCT 647

of overestimation of stenosis and occlusion byMRA is de-

picted in Figure 1.

In terms of group distribution, evaluation of MRA and

CTA scores for the ICA revealed 104 vessels (72.2%) in the

good correlation group, 8 vessels (5.6%) in the underesti-

mation group, and 32 vessels (22.2%) in the overestima-

tion group. Compared with the good correlation group,

the mean MRA score was significantly higher in the over-

estimation group (1.52 6 1.00 vs 2.09 6 0.59; P , .0089)

and lower in the underestimation group (1.52 6 1.00 vs

0.13 6 0.35; P , .0148). For the MCA, evaluation revealed

vessels (55.6%) in the good correlation group, 6 vessels

(4.2%) in the underestimation group, and 58 vessels

(40.3%) in the overestimation group. The mean MRA

score was significantly higher in the overestimation

group than in the good correlation group (2.50 6 0.66 vs

1.631.73 6 0.45) 6 1.01; (P , .0001). The ACA included

121 vessels (84.0%) in the good correlation group, 4 ves-

sels (2.8%) in the underestimation group, and 19 vessels

(13.2%) in the overestimation group. The mean MRA

score was significantly higher in the overestimation

group compared with the good correlation group (1.73

6 0.45 vs 0.60 6 0.76; P , .0001). Finally, the PCA had

131 vessels (91.0%) in the good correlation group, 7 ves-

sels (4.9%) in the underestimation group, and 6 vessels

(4.2%) in the overestimation group.

In this series, moyamoya-affected vessels were ob-

served in 115 sides (79.9%) through CTA, but in only 85

sides (59.0%) through MRA. These data indicate that

CTA was significantly superior to MRA in the detection

of moyamoya vessels (P 5 .0001, c2 test).

Discussion

Moyamoya vessels (abnormally dilated perforators of

the basal ganglia) represent collateral circulation that de-

velops secondarily to the steno-occlusive changes in the

carotid fork.23 The detection of moyamoya vessels is the

most important factor in the diagnosis of this disease, as

well as in the differential diagnosis. Conventional angiog-

raphy remains the gold standard for the diagnosis of

moyamoya disease, but the Japanese Research Committee

recommends that cerebral angiography not be mandatory

if MRI and MRA show findings typical of moyamoya

disease.3 Recent studies comparing CTA using MDCT

with conventional angiography found that CTA was

very useful in determining the degree of stenosis and oc-

clusion in patients with atherosclerotic disease.14,15,17,18

Some studies have suggested that CTA may obviate the

need for conventional angiography in many clinical

circumstances.16 In the present study, stenosis and occlu-

sion were evaluated in patients with moyamoya disease

by CTA and MRA, and a quantitative comparison was

performed. Our findings suggest that CTA canmore accu-

rately evaluate the degree of stenosis and can detect

moyamoya vessels with greater sensitivity than MRA.

CTA is also considered a useful tool in preoperative plan-

ning.25 Moreover, the radiation dose received in a CTA

procedure (equivalent to �90-100 mSv) is 5- to 8-fold

lower than that received in conventional angiography.

Our findings reveal several key differences between

CTA andMRA that pertain to the diagnosis of moyamoya

disease. The first is the tendency of MRA to overestimate

Figure 3. VR imaging for preoperative plan-

ning (A-C) and postoperative evaluation (D-F).

(A and D) Surface imaging. (B and E) Translu-

cent imaging. (C and F) Vascular extracted imag-

ing. This postprocessing technique enabled

extraction of the required information. The arrow-

head indicates the middle meningeal artery.

T. SUGINO ET AL.648

the severity of steno-occlusive changes, especially in

vessels with higher MRA scores. The possibility of re-

duced accuracy and spatial resolution arising from the

relatively poor image quality of MRA should always be

taken into account when evaluating steno-occlusive

changes with this modality. In the present study, we

used 3.0-T MRI, which has superior detection capability

compared with 1.5-T MRI.26,27 Thus, our ability to detect

mild stenosis and small vessels, such as distal arteries

and moyamoya vessels, on CTA was not inferior to that

provided by conventional angiography. Probable causes

of the tendency of MRA to overestimate the severity of

lesions include laminar flow of the carotid siphon and

carotid fork, phase dispersion due to disturbed flow

at stenotic lesions, and degradation of flow contrast

medium due to the decreased flow rate at stenotic

lesions.28-32

As an example of overestimation in the present study,

one case that had been diagnosed as discontinuity of

the ICA signal by MRA was more accurately identified

as stenosis by CTA. Overall, stenosis (MRA/CTA score

of 1) is more frequently diagnosed through CTA than

throughMRA. Another case that was considered invisible

when visualized through MRAwas diagnosed as discon-

tinuity of signal when visualized through CTA. CTA’s

greater sensitivity is provided by its enhanced visualiza-

tion of distal flow. As result, CTA can readily identify

very small stenotic lesions in the ICA, MCA, and PCA,

whereas MRA tends to underestimate the presence of

such lesions. Therefore, especially in the earliest and

latest stages of moyamoya disease, diagnosis by MRA

should be done with caution, given that the typical find-

ings of steno-occlusive changes and the development of

moyamoya vessels are not always observed in these

stages.

Among the available postprocessing techniques that

can be applied to CTA, MIP is the most helpful in diag-

nosing moyamoya disease. In the present study, we ap-

plied both VR imaging and MIP imaging, but found

that VR imaging is not suitable for the evaluation of

steno-occlusive lesions as opposed to aneurysmal

lesions.14-17 However, VR imaging with CTA enables

detailed 3D analysis of areas including the skull, which

has proven helpful in guiding surgical strategies, as

depicted in Figure 3A-C.25 This imaging modality also

provides some information regarding transdural anasto-

mosis, especially in the middle meningeal artery, which

might help avoid unexpected injury of this spontaneous

collateral circulation during craniotomy and dural inci-

sion. CTA also has proven useful evaluating postopera-

tive status (especially bypass patency), other indirect

collateral circulation, and hyperperfusion.13,19 CTA is

not affected by the presence titanium plates after

surgery (Fig 3D-F).

CTA is not without possible drawbacks, however. In

particular, the contrast medium sometimes causes com-

plications, and the clinician should have experience in

proper timing of contrast medium injection.

The optimum protocol for each particular case depends

on the resources available in the facility where the diagno-

sis is to be performed. The introduction of 320-section CT

scanners will make even more detailed evaluations possi-

ble. Overall, we conclude that CTA using modern MDCT

is at least equal to, and perhaps even superior to, MRA

for the diagnosis of moyamoya disease.

Conclusion

CTA by means of MDCT may serve as a reliable and

possible alternative method to MRA in the diagnosis

ASSESSMENT OF MOYAMOYA DISEASE USING MDCT 649

and assessment of moyamoya disease. CTA can be per-

formed quickly and without any specialized equipment.

The short duration of the procedure is a particular advan-

tage for patients with moyamoya disease, particularly in

pediatric and emergency cases.

References

1. Houkin K, Aoki T, Takahashi A, et al. Diagnosis of moya-moya disease with magnetic resonance angiography.Stroke 1994;25:2159-2164.

2. Yamada I, Matsushima Y, Suzuki S. Moyamoya disease:Diagnosis with three-dimensional time-of-flight MRangiography. Radiology 1992;184:773-778.

3. Fukui M. Research Committee on Spontaneous Occlusionof the Circle of Willis (Moyamoya Disease) of the Minis-try of Health and Welfare, Japan. Guidelines for the diag-nosis and treatment of spontaneous occlusion of the circleof Willis (‘‘moyamoya’’ isease). Clin Neurol Neurosurg1997;99(Suppl 2):S238-S240.

4. Numaguchi Y, Gonzalez CF, Davis PC, et al. Moyamoyadisease in the United States. Clin Neurol Neurosurg1997;99(Suppl 2):S26-S30.

5. Yamada I, Suzuki S, Matsushima Y. Moyamoya disease:Comparison of assessment with MR angiography andMR imaging versus conventional angiography. Radiol-ogy 1995;196:211-218.

6. Aoki S, Yoshikawa T, Hori M, et al. Two-dimensionalthick-slice MR digital subtraction angiography for assess-ment of cerebrovascular occlusive diseases. Eur Radiol2000;10:1858-1864.

7. Hasuo K, Mihara F, Matsushima T. MRI and MR angiog-raphy in moyamoya disease. J Magn Reson Imaging 1998;8:762-766.

8. Kuroda S, Houkin K. Moyamoya disease: Currentconcepts and future perspectives. Lancet Neurol 2008;7:1056-1066.

9. Shirane R, Mikawa S, Ebina T. A case of adult moyamoyadisease showing progressive angiopathy on cerebral an-giography. Clin Neurol Neurosurg 1999;101:210-214.

10. Takanashi JI, Sugita K, Niimi H. Evaluation of magneticresonance angiography with selective maximum inten-sity projection in patients with childhood moyamoyadisease. Eur J Paediatr Neurol 1998;2:83-89.

11. Thibaud C, Garnier-Viarouge MP, De Kersaint-Gilly A,et al. Moyamoya disease: Importance of the MRI-MRAcombination and difficulties in management andfollow-up in 7 cases. J Neuroradiol 2001;28:84-91.

12. Yamada I, Nakagawa T,Matsushima Y, et al. High-resolu-tion turbo magnetic resonance angiography for diagnosisof moyamoya disease. Stroke 2001;32:1825-1831.

13. Kikuchi M, Asato M, Sugahara S, et al. Evaluation of sur-gically formed collateral circulation in moyamoya dis-ease with 3D-CT angiography: Comparison with MRangiography and X-ray angiography. Neuropediatrics1996;27:45-49.

14. Lell MM, Anders K, Uder M, et al. New techniques in CTangiography. Radiographics 2006;26(Suppl 1):S45-S62.

15. Malago R, D’Onofrio M, Tavella D, et al. Diagnostic accu-racy in coronary stenosis: comparison between visual

score and quantitative analysis (quantitative computedtomographic angiography) in coronary angiography bymultidetector computed tomography–coronary angiog-raphy and quantitative analysis (quantitative coronaryangiography) in conventional coronary angiography. JComput Assist Tomogr 2010;34:652-659.

16. Menke J, Larsen J, Kallenberg K. Diagnosing cerebral an-eurysms by computed tomographic angiography: Meta-analysis. Ann Neurol 2011;69:646-654.

17. MetR,BipatS,LegemateDA,et al.Diagnosticperformanceof computed tomography angiography in peripheral arte-rial disease:A systematic reviewandmeta-analysis. JAMA2009;301:415-424.

18. Tan JC, Dillon WP, Liu S, et al. Systematic comparison ofperfusion-CT and CT-angiography in acute stroke pa-tients. Ann Neurol 2007;61:533-543.

19. Bacigaluppi S, Dehdashti AR, Agid R, et al. The contribu-tion of imaging in diagnosis, preoperative assessment,and follow-up of moyamoya disease: a review. Neuro-surg Focus 2009;26:E3.

20. Brott TG, Hobson RW 2nd, Howard G, et al. Stenting ver-sus endarterectomy for treatment of carotid-artery steno-sis. N Engl J Med 2010;363:11-23.

21. Houkin K, Nakayama N, Kuroda S, et al. Novel magneticresonance angiography stage grading for moyamoya dis-ease. Cerebrovasc Dis 2005;20:347-354.

22. Suzuki J, Kodama N. Cerebrovascular ‘‘moyamoya’’ dis-ease, 2: Collateral routes to forebrain via ethmoid sinusand superior nasal meatus. Angiology 1971;22:223-236.

23. Suzuki J, Kodama N. Moyamoya disease: A review.Stroke 1983;14:104-109.

24. Suzuki J, Takaku A. Cerebrovascular ‘‘moyamoya’’ dis-ease: Disease showing abnormal net-like vessels in baseof brain. Arch Neurol 1969;20:288-299.

25. Coppens JR, Cantando JD, Abdulrauf SI. Minimally inva-sive superficial temporal artery to middle cerebral arterybypass through an enlarged bur hole: The use of com-puted tomography angiography neuronavigation in sur-gical planning. J Neurosurg 2008;109:553-558.

26. Jin Q, Noguchi T, Irie H, et al. Assessment of moyamoyadisease with 3.0-T magnetic resonance angiography andmagnetic resonance imaging versus conventional angiog-raphy. Neurol Med Chir (Tokyo) 2011;51:195-200.

27. Kikuta K. Experiences using 3-Tesla magnetic resonanceimaging in the treatment of moyamoya disease. ActaNeurochir Suppl 2008;103:123-126.

28. Anderson CM, Saloner D, Tsuruda JS, et al. Artifacts inmaximum-intensity projection display of MR angio-grams. AJR Am J Roentgenol 1990;154:623-629.

29. Creasy JL, Price RR, Presbrey T, et al. Gadolinium-enhanced MR angiography. Radiology 1990;175:280-283.

30. Marchal G, Bosmans H, Van Fraeyenhoven L, et al. Intra-cranial vascular lesions: Optimization and clinical evalua-tion of three-dimensional time-of-flight MR angiography.Radiology 1990;175:443-448.

31. Motomiya M, Karino T. Flow patterns in the human ca-rotid artery bifurcation. Stroke 1984;15:50-56.

32. Ruggieri PM, Laub GA, Masaryk TJ, et al. Intracranialcirculation: pulse-sequence considerations in three-dimensional (volume) MR angiography. Radiology1989;171:785-791.