Cpaaen

M

TcacttFphiivawt

UTr

C

2

0d

Coronary Artery Diameters in Infants and Children WithCongenital Heart Disease as Determined by Computed

Tomography

Shyh-Jye Chen, MD, PhDa, Wen-Jeng Lee, MDa, Ming-Tai Lin, MDb, Kao-Lang Liu, MDa,Jou-Kou Wang, MD, PhDb,*, and Hung-Chi Lue, MDb

This study was conducted to establish reference curves and formulas for the diameters ofthe coronary arteries in infants and children using computed tomography. A total of 145children (57 female, 88 male) ranging in age from 2 days to 19 years, 11 months (mean 5years, 10 months), were retrospectively identified, and the diameters of their coronaryarteries were quantified. The measurability of the coronary arteries, coronary arterial sizedifferences between the genders, and relations of the diameters of the coronary arteries toage, height, weight, body surface area, and the diameter of the descending aorta wereexamined. Independent-samples Student’s t tests, 2-tailed Pearson’s correlations, andlinear regression were used in statistical analysis. The measurability of all coronary arterieswas 73.3%. No difference in coronary arterial size was found between the genders. Thediameter of the descending aorta correlated most strongly with coronary arterial size. Inconclusion, predictive formulas and reference curves for coronary arterial diameters ininfants and children were obtained. © 2007 Elsevier Inc. All rights reserved. (Am J

Cardiol 2007;100:1696–1701)

tbt

pcPaodhmmhn8(cTP

Gtdpn1sgd2pm

omputed tomography (CT) for coronary imaging hasroved reliable in demonstrating the coronary arteries indults1,2 and children3–6 with or without coronary arterialnomalies. The purpose of this study was to establish ref-rence curves and formulas for the diameters of the coro-ary arteries in infants and children, using CT.

ethods

he study protocol was approved by the hospital ethicsommittee, and written informed consent was obtained fromll patients or their guardians, as appropriate, after theomputed tomographic procedure was explained to the pa-ients or their family members. We conducted a retrospec-ive study at a 2,000-bed tertiary-referral general hospital.rom November 2003 to August 2006, 855 consecutiveatients aged �20 years who underwent cardiac CT at thisospital were included. Their indications for CT varied andncluded the identification of hypoplastic pulmonary arter-es, the determination of possible anomalous pulmonaryein configurations or stenoses, the clarification of associ-ted bronchopulmonary condition, and so on, associatedith underlying cardiac problems. Of these 855 patients,

hose with situs solitus, concordant atrioventricular connec-

aDepartments of Medical Imaging and bPediatrics, National Taiwanniversity Hospital and National Taiwan University College of Medicine,aipei, Taiwan. Manuscript received May 8, 2007; revised manuscript

eceived and accepted July 1, 2007.This study was partially supported by Grant CCF 06-05 from the

hildren’s Cardiac Foundation of the Republic of China, Taipei, Taiwan.*Corresponding author: Tel: 886-2-23123456 ext. 5137; fax: 886-2-

393-4749.

(E-mail address: jkw@ntumc.org (J.-K. Wang).

002-9149/07/$ – see front matter © 2007 Elsevier Inc. All rights reserved.oi:10.1016/j.amjcard.2007.07.016

ions, concordant ventriculoatrial connections, and normalranching patterns of the coronary arteries were enrolled inhis study.

Exclusion criteria were used to prevent the enrollment ofatients with any inherited or acquired abnormalities thatould cause size abnormalities of the coronary arteries.atients were excluded if they had (1) structural aorticbnormalities (such as hypoplasia, coarctation, interruption,r dilation) that likely had hemodynamic effects on theiameters of the coronary arteries; (2) cardiomegaly orypertrophy of the left ventricle or right ventricle, whichay cause coronary artery dilation by increasing flow de-and; and (3) fevers, because febrile conditions can cause

igh cardiac output, resulting in the enlargement of coro-ary arterial diameter.7 A total of 145 children (57 female,8 male) ranging in age from 2 days to 19 years, 11 monthsmean 5 years, 10 months), met the inclusion and exclusionriteria for this study. Their main diagnoses are listed inable 1, and their age distribution is shown in Figure 1.atients’ heart rates ranged from 86 to 148 beats/min.

All subjects underwent 16-detector CT (LightSpeed 16;E Medical Systems, Milwaukee, Wisconsin), with elec-

rocardiographic gating. Images were obtained at the end-iastolic phase of the cardiac cycle. Data acquisition waserformed in the craniocaudal direction with a section thick-ess of 0.625 mm, without gaps. The x-ray tube current was20 kV at 250 to 550 mA, adjusted on the basis of patientize. The gantry rotation time was 500 ms. Electrocardio-raphically gated segment reconstruction was used to pro-uce diagnostic images. The temporal resolution time was27 ms. The matrix size in the XY plane was 512 � 512ixels. The calculated total radiation dose was approxi-ately 4 to 11 mSv.8,9 Nonionic iodinated contrast medium

2 ml/kg Ultravist 370; Schering AG, Berlin, Germany) was

www.AJConline.org

dtbstistsda5

pCip1

tabafiaa(r

FstctR

TP

D

NVAAPTAA

LPPPVIPPPT

1697Congenital Heart Disease/Coronary Size in Children by CT

elivered through a power injector in all patients at 80% ofhe maximally allowable injection rate. The delay timeetween the start of contrast medium administration and thetart of imaging was determined using a bolus-trackingechnique.10 The region of measurement used in bolus track-ng was in the ascending aorta, and the threshold level waset at 150 Hounsfield units. If the patients were old enougho cooperate, we instructed them to hold their breath duringcanning. Otherwise, quiet respiration without crying waseemed acceptable in small children and infants. Patientsged �5 years were routinely sedated with chloral hydrate0 mg/kg during their examinations.

Image reconstruction by multiplanar reformatting waserformed on an independent Hewlett-Packard (Palo Alto,alifornia) workstation (Advantage Windows 4.1; GE Med-

cal Systems). All computed tomographic images were dis-layed using a fixed window (level � 200, window �

Figure 1. Histogram of age distribution.

able 1atients’ cardiac-related diagnoses

iagnosis No. ofCases

ormal intracardiac structure 39entricular septal defect 22trial septal defect 16trial septal defect and ventricular septal defect 13artial anomalous pulmonary venous return 9otal anomalous pulmonary venous return 6berrant subclavian artery 6trial septal defect and partial anomalous pulmonaryvenous return

5

eft pulmonary artery sling 5atent foramen ovale and ventricular septal defect 4ulmonary venous obstruction 3ericardial effusion 3egetation on valves 3

ntracardiac tumor 3atent ductus arteriosus 3ulmonary stenosis 3atent foramen ovale 2otal 145

,000) for viewing. Thin-slab reformatted images were used t

o identify the coronary arterial segments. The readers werellowed to scroll through the volume until they found theest image for the visualization of a particular coronaryrterial segment. The left main coronary artery was identi-ed first. After that, the left anterior descending coronaryrtery, left circumflex coronary artery, and right coronaryrtery were sequentially evaluated.11 Two readers, with 12SJC) and 5 (WJL) years of experience in cardiac CT,etrospectively assessed the computed tomographic images

igure 2. Measurement of the caliber of the (A) right coronary artery. Themall diagram (lower left) represents the magnified Hounsfield profile ofhe index line that is perpendicular to the long axis of the proximal rightoronary artery. Measurement of the diameter of the DA (B) at the level ofhe left main coronary artery origin (at sign). PIR � posterior inferior right;IA � right inferior arterior.

ogether to confirm identification of the coronary arterial

sonau

Hacd

Fla

Fpca

TCts

V

AHWBD

N

ca

1698 The American Journal of Cardiology (www.AJConline.org)

egments of interest and their measurements. Measurementf the internal contrast-filled luminal caliber of these coro-ary arterial segments was done on the workstation directlynd electronically. Quantification of the calibers was done

igure 3. Change in the measurability of coronary arteries by age. LAD �eft anterior descending coronary artery; LC � left circumflex coronaryrtery; LM � left main coronary artery; R � right coronary artery.

igure 4. Reference curves for coronary arterial caliber versus diameter oarallel lines are the 95% confidence limits for the regression line. The biloronary artery (R0.5 � 0.79); (B) left anterior descending coronary artery (rtery (R0.5 � 0.63).

sing the principle of the full-width half-maximum (

ounsfield unit system by measuring from 1 inner edge tonother inner edge, excluding points of branching.12 Theoronary arteries were measured approximately at the mid-le portion of each segment before their first major branches

A. The line in the center represents the mean value. The bilateral inneruter lines are the 95% confidence limits for the data points. (A) Left main0.80); (C) left circumflex coronary artery (R0.5 � 0.62); (D) right coronary

able 2orrelation coefficients of age, height, weight, body surface area, and

he diameter of the descending aorta with the sizes of the differentegments of the coronary arteries

ariable LM LAD LCx R

ge (d) 0.85 0.83 0.73 0.76eight (cm) 0.88 0.88 0.76 0.77eight (kg) 0.80 0.80 0.68 0.75

ody surface area (m2) 0.79 0.82 0.71 0.77iameter of the DA(mm)

0.89 0.90 0.79 0.79

o. of cases 101 128 89 102

LAD � left anterior descending coronary artery; LCx � left circumflexoronary artery; LM � left main coronary artery; R � right coronaryrtery.

f the Dateral oR0.5 �

Figure 2).

tcAtmssaoem

ehwtptnDtc

wlqdstrhDhwcsuctpoefv

Ft(�

1699Congenital Heart Disease/Coronary Size in Children by CT

The measurability of the coronary arteries was defined ashe percentage of well-visualized and successfully measuredoronary arterial segments to that of the studied cases.dditionally, we stratified patients into different age groups

o explore any possible effect of age on coronary arterialeasurability. Gender is reported to have no effect on the

izes of coronary arteries. Using the 2-tailed, independent-amples Student’s t test, we also explored any differences inge, the diameter of the descending aorta (DA) at the levelf the left main coronary artery origin, and the caliber ofach coronary arterial segment between our male and fe-ale subjects.The diameters of the coronary arteries measured by

chocardiography are reportedly related to a patient’seight and body surface area.7,13 Nonetheless, if thereere any independent internal indicator on computed

omographic images that could be used to represent aatient’s physique, a diagnosis could be made of whetherhe concerned coronary arterial segments were dilated orot, solely from the computed tomographic images. TheA was proposed to be this internal indicator, because

he DA has similar hydrophilic pressure related to the

igure 5. Reference curves for coronary arterial caliber versus height. Thehe 95% confidence limits for the regression line. The bilateral outer linesR0.5 � 0.77); (B) left anterior descending coronary artery (R0.5 � 0.76); (C

0.60).

oronary arteries. For consistent quantification, the DA A

as measured on the transverse sectional image at theevel of the origin of the left main coronary artery. It wasuantified using 2 perpendicular measurements of theiameter on the transverse section (Figure 2). The repre-ented caliber of the DA was the root of the multiple ofhese 2 perpendicular measurements. We tested the cor-elation between the size of the coronary arteries and age,eight, weight, body surface area, and the diameter of theA using 2-tailed Pearson’s correlations. Using the mostighly correlated parameter as an independent variable,e created formulas to predict the diameter of each

oronary arterial segment using linear regression analy-is. For more convenience in clinical practice, we furthersed curve fitting in the linear regression analyses for alloronary arterial segments. Because age and height arehe most available information representing a patient’shysique, we also created formulas and reference curvesn the basis of age or height to predict the diameter ofach coronary artery. All statistical analyses were per-ormed using commercially available software (SSPSersion 11.0 for Windows; SPSS, Inc., Chicago, Illinois).

he center represents the mean value. The bilateral inner parallel lines are95% confidence limits for the data points. (A) Left main coronary artery

ircumflex coronary artery R0.5 � 0.58); and (D) right coronary artery (R0.5

line in tare the) left c

p value �0.01 was considered significant.

R

Atalalti

�o�d

fam(

Dr

wd(cca

f(e(0��aml�(�

Fla(

1700 The American Journal of Cardiology (www.AJConline.org)

esults

total of 420 coronary arterial segments were analyzed inhe 145 patients. The percentages of measurable coronaryrterial segments for all ages were 70% (101 of 145) for theeft main coronary artery, 88% (128 of 145) for the leftnterior descending coronary artery, 61% (89 of 145) for theeft circumflex coronary artery, and 70% (102 of 145) forhe right coronary artery. As age increased, the measurabil-ty of the coronary arteries mostly improved (Figure 3).

The mean age of the subjects did not differ by gender (p0.99). The diameter of the DA at the level of the origin

f the left main coronary artery did not differ by gender (p0.46). Furthermore, the diameters of all coronary arteries

id not differ by gender (p �0.25).The correlations between age, height, weight, body sur-

ace area, and the diameter of the DA and the coronaryrteries are listed in Table 2. The diameter of the DA wasost highly correlated with the sizes of all coronary arteries

all p �0.001).Using the most correlated parameter, the diameter of the

A (in millimeters), as the single independent variable

igure 6. Reference curves for coronary artery caliber against to the patientines are the 95% confidence limits for the regression line. The bilateral ourtery (R0.5 � 0.72); (B) left anterior descending coronary artery (R0.5 � 0.6R0.5 � 0.53).

epresenting patient physique, the linear regression models m

ere all significant in predicting coronary arterial segmentiameters (all p �0.001): (1) left main coronary artery �0.18 � DA � 0.54) � 0.11; (2) left anterior descendingoronary artery � (0.14 � DA � 0.55) � 0.07; (3) leftircumflex coronary artery � (0.095 � DA � 0.64) � 0.09;nd (4) right coronary artery � (0.12 � DA � 0.69) � 0.11.

For clinical convenience and usefulness, we also createdormulas and reference curves on the basis of patient heightin centimeters) or age (in days) to predict the diameter ofach coronary arterial segment (all p �0.001). For height,1) left main coronary artery � (0.018 � height � 0.59) �.11; (2) left anterior descending coronary artery � (0.014

height � 0.55) � 0.08; (3) left circumflex coronary artery(0.0093 � height � 0.65) � 0.11; and (4) right coronary

rtery � (0.012 � height � 0.68) � 0.12. For age, (1) leftain coronary artery � (0.00033 � age � 1.7) � 0.07; (2)

eft anterior descending coronary artery � (0.00024 � age1.4) � 0.06; (3) left circumflex coronary artery �

0.00016 � age � 1.3) � 0.06; and (4) right coronary artery(0.00021 � age � 1.5) � 0.06.The proposed reference curves of coronary arterial seg-

e line in the center represents the mean value. The bilateral inner parallels are the 95% confidence limits for the data points. (A) Left main coronaryleft circumflex coronary artery R0.5 � 0.58); and (D) right coronary artery

age. Thter line9); (C)

ents diameter together with their 95% confidence intervals

rs

D

TCcaeace

teHtaodCc

nwtowmcngoTss

FNhrcoegrtfcaoAsc

Ah

1

1

1

1

1

1

1

1

1

1701Congenital Heart Disease/Coronary Size in Children by CT

elative to the diameter of the DA, height, and age arehown in Figures 4, 5, and 6, respectively.

iscussion

here are no previous reports showing the successful use ofT for the quantification of coronary arterial diameters inhildren. In this study, we found that CT had good measur-bility for all coronary arteries. No significant size differ-nce in proximal coronary arterial diameters between malend female subjects has been reported.13,14 Our findings areonsistent with studies reporting that gender has little directffect on coronary arterial diameters.

The size criteria of the American Heart Association15 orhe Japanese Ministry of Health16 to classify coronary an-urysm are not based on an individual patient’s body size.owever, in this study, we confirmed that the diameters of

he coronary arteries in children increased with increasingge, as previously reported.7,17 We also found that the sizef the DA is more strongly correlated with coronary arteryiameter than age, weight, height, or body surface area.onsequently, we used the diameter of the DA to predictoronary arterial diameters.

Although computed tomographic coronary imaging doesot obviate the risks of contrast and radiation associatedith conventional angiography, it does not expose children

o the possible complications associated with invasive cor-nary angiography. Furthermore, CT can be performedithout the need for hospitalization, and its noninvasivenessake it is suitable for sequentially following patients. Be-

ause of the rapid advances in computed tomographic tech-ology and the widespread installation of computed tomo-raphic scanners, establishing criteria for the use of CT tobtain normal reference coronary arterial size is necessary.hese criteria could be used to confidently classify theubtle coronary arterial size abnormalities in subjects withuspected Kawasaki’s disease.

The following study limitations must be acknowledged.irst, not all of our subjects had normal intracardiac structure.evertheless, our patients with abnormal intracardiac structuread only small defects, with limited shunting, on echocardiog-aphy; they did not have dilation or hypertrophy of any cardiachambers (included among the study’s exclusion criteria). Sec-nd, not all coronary arteries could be clearly measured. Anarlier study suggested that image quality on computed tomo-raphic coronary angiography is inversely proportional to heartate.18 Our subjects (35.2% of whom were infants) had rela-ively rapid heart rates, so some images were not good enoughor coronary arterial quantification. Third, most of our subjectsould not hold their breath during CT because of their youngges (59.3% were aged �6 years). The respiratory movementf the chest wall also degraded coronary arterial image quality.

final limitation is the irradiation inherent in CT. In conclu-ion, we obtained predictive formulas and reference curves fororonary arterial diameters for infants and children using CT.

cknowledgment: We would like to thank Ritta Huang for

er assistance in the preparation of this report.

1. Kuettner A, Beck T, Drosch T, Kettering K, Heuschmid M, Burgs-tahler C, Claussen CD, Kopp AF, Schroeder S. Diagnostic accuracy ofnoninvasive coronary imaging using 16-detector slice spiral computedtomography with 188 ms temporal resolution. J Am Coll Cardiol2005;45:123–127.

2. Kuettner A, Trabold T, Schroeder S, Feyer A, Beck T, Brueckner A,Heuschmid M, Burgstahler C, Kopp AF, Claussen CD. Noninvasivedetection of coronary lesions using 16-detector multislice spiral com-puted tomography technology: initial clinical results. J Am Coll Car-diol 2004;44:1230–1237.

3. Goo HW, Park IS, Ko JK, Kim YH, Seo DM, Yun TJ, Park JJ.Visibility of the origin and proximal course of coronary arteries onnon-ECG-gated heart CT in patients with congenital heart disease.Pediatr Radiol 2005;35:792–798.

4. Chen SJ, Lin MT, Lee WJ, Wang JK, Chang CI, Li YW, Chiu IS.Coronary artery anatomy in children with congenital heart disease bycomputed tomography. Int J Cardiol 2007;120:363–370.

5. Goo HW, Park IS, Ko JK, Kim YH. Coronary CT angiography andMR angiography of Kawasaki disease. Pediatr Radiol 2006;36:697–705.

6. Chu WC, Mok GC, Lam WW, Yam MC, Sung RY. Assessment ofcoronary artery aneurysms in paediatric patients with Kawasaki dis-ease by multidetector row CT angiography: feasibility and comparisonwith 2D echocardiography. Pediatr Radiol 2006;36:1148–1153.

7. Kurotobi S, Nagai T, Kawakami N, Sano T. Coronary diameter innormal infants, children and patients with Kawasaki disease. PediatrInt 2002;44:1–4.

8. Nakonechny KD, Fallone BG, Rathee S. Novel methods of measuringsingle scan dose profiles and cumulative dose in CT. Med Phys2005;32:98–109.

9. DeMarco JJ, Cagnon CH, Cody DD, Stevens DM, McCollough CH,Zankl M, Angel E, McNitt-Gray MF. A Monte Carlo based method toestimate radiation dose from multidetector CT (MDCT): cylindricaland anthropomorphic phantoms. Phys Med Biol 2005;50:3989–4004.

0. Cademartiri F, van der Lugt A, Luccichenti G, Pavone P, Krestin GP.Parameters affecting bolus geometry in CTA: a review. J ComputAssist Tomogr 2002;26:598–607.

1. Scanlon PJ, Faxon DP, Audet AM, Carabello B, Dehmer GJ, EagleKA, Legako RD, Leon DF, Murray JA, Nissen SE, et al. ACC/AHAguidelines for coronary angiography: a report of the American Collegeof Cardiology/American Heart Association Task Force on PracticeGuidelines (Committee on Coronary Angiography)—developed incollaboration with the Society for Cardiac Angiography and Interven-tions. J Am Coll Cardiol 1999;33:1756–1824.

2. Liu Y, Hopper KD, Mauger DT, Addis KA. CT Angiographic mea-surement of the carotid artery: optimizing visualization by manipulat-ing window and level settings and contrast material attenuation. Ra-diology 2000;217:494–500.

3. Oberhoffer R, Lang D, Feilen K. The diameter of coronary arteries ininfants and children without heart disease. Eur J Pediatr 1989;148:389–392.

4. Arjunan K, Daniels SR, Meyer RA, Schwartz DC, Barron H, KaplanS. Coronary artery caliber in normal children and patients with Ka-wasaki disease but without aneurysms: an echocardiographic and an-giographic study. J Am Coll Cardiol 1986;8:1119–1124.

5. Dajani AS, Taubert KA, Takahashi M, Bierman FZ, Freed MD, Fer-rieri P, Gerber M, Shulman ST, Karchmer AW, Wilson W. Guidelinesfor long-term management of patients with Kawasaki disease. Reportfrom the Committee on Rheumatic Fever, Endocarditis, and KawasakiDisease, Council on Cardiovascular Disease in the Young, AmericanHeart Association. Circulation 1994;89:916–922.

6. Research Committee on Kawasaki Disease. Report of Subcommitteeon Standardization of Diagnostic Criteria and Reporting of CoronaryArtery Lesions in Kawasaki Disease. Tokyo, Japan: Ministry of Healthand Welfare, 1984.

7. de Zorzi A, Colan SD, Gauvreau K, Baker AL, Sundel RP, NewburgerJW. Coronary artery dimensions may be misclassified as normal inKawasaki disease. J Pediatr 1998;133:254–258.

8. Nieman K, Rensing BJ, van Geuns RJ, Vos J, Pattynama PM, KrestinGP, Serruys PW, de Feyter PJ. Non-invasive coronary angiographywith multislice spiral computed tomography: impact of heart rate.

Heart 2002;88:470–474.