sex-specific pediatric percentiles for ventricular …...are faster and provide improved image...

1942-0080American Heart Association. All rights reserved. Print ISSN: 1941-9651. Online ISSN: 2010 Copyright ©

Avenue, Dallas, TX 72514Circulation: Cardiovascular Imaging is published by the American Heart Association. 7272 Greenville

DOI: 10.1161/CIRCIMAGING.109.859074 2010;3;65-76; originally published online October 9, 2009;Circ Cardiovasc Imaging

Kelter-Kloepping, Hermann Koerperich, Titus Kuehne and Philipp BeerbaumSamir Sarikouch, Brigitte Peters, Matthias Gutberlet, Birte Leismann;, Andrea

MRI Flowfor Cardiac MRI : Assessment by Steady-State Free-Precession and Phase-Contrast Sex-Specific Pediatric Percentiles for Ventricular Size and Mass as Reference Values

http://circimaging.ahajournals.org/content/3/1/65.full

on the World Wide Web at: The online version of this article, along with updated information and services, is located

http://www.lww.com/reprintsReprints: Information about reprints can be found online at

[email protected]. E-mail:Health, 351 West Camden Street, Baltimore, MD 21201-2436. Phone: 410-528-4050. Fax: Permissions: Permissions & Rights Desk, Lippincott Williams & Wilkins, a division of Wolters Kluwer

http://circimaging.ahajournals.org/site/subscriptions/Subscriptions: Information about subscribing to Circulation: Cardiovascular Imaging is online at

by guest on June 18, 2012circimaging.ahajournals.orgDownloaded from

http://circimaging.ahajournals.org/content/3/1/65.full

http://circimaging.ahajournals.org/site/subscriptions/

mailto:[email protected]

http://www.lww.com/reprints

http://circimaging.ahajournals.org/

Sex-Specific Pediatric Percentiles for Ventricular Size andMass as Reference Values for Cardiac MRI

Assessment by Steady-State Free-Precession and Phase-ContrastMRI Flow

Samir Sarikouch, MD; Brigitte Peters, MD; Matthias Gutberlet, PhD; Birte Leismann;Andrea Kelter-Kloepping, RN; Hermann Koerperich, PhD; Titus Kuehne, PhD; Philipp Beerbaum, MD

Background—Cardiac MRI is important in the treatment of children with congenital heart disease, but sufficient normativedata are lacking. For ventricular volumes and mass, we sought to deliver reference centiles and to investigate sex effects.

Methods and Results—We included 114 healthy children and adolescents, uniformly distributed spanning an age range of4 to 20 years, as required by the Lambda-Mu-Sigma method to achieve a percentile distribution, thus avoiding arbitraryage categories. Subjects underwent axial volumetry (1.5-T scanner) using standardized 2D steady-state free-precessionand flow protocols. Percentiles were computed for age 8 to 20 years (99 subjects) because breath-holds were moreconsistent in this group. When indexed for body surface area or height, the centile curves of ventricular volumetricparameters showed allometric increase until adolescence, when a plateau was reached, with values comparable topublished adult reference data. In contrast, ventricular mass centiles increased without plateau. There was a significantsex difference, with centiles reflecting larger values in boys than in girls (P�0.05) when ventricular volumes wereindexed to body surface area or height but not when indexed to weight (exception: mass). There was excellent agreementof axial and short-axis volumetry and of volumetric and flow-derived stroke volumes.

Conclusions—Percentiles for ventricular volumes and mass in healthy children have been established to serve as referencevalues in pediatric heart disease. Significant sex differences were noted when indexing volumes to body surface areaor height. Unisex centiles related to weight may be considered for chamber volumes albeit not for mass. (CircCardiovasc Imaging. 2010;3:65-76.)

Key Words: heart disease � pediatrics � MRI � volumetry

The diagnosis of an enlarged or hypertrophied heart hasimportant implications for the treatment of children with

congenital or acquired heart disease. To this end, cardiac MRIhas evolved as an important noninvasive diagnostic tool, partic-ularly in the noninvasive quantitative assessment of the rightventricle.1 Cardiac MRI is generally considered the referencestandard for the assessment of ventricular dimensions, function,and mass in terms of accuracy and reproducibility,2 and this isreflected by class I–II recommendations for the clinical use ofcardiac MRI by recent consensus panels.3

Clinical Perspective on p 76Although these recommendations explicitly include deter-

mination of ventricular size and cardiac function, data are

sparse regarding these parameters in healthy children.4,5

Moreover, since these smaller series were published, spoiledgradient-echo cine (cine-GRE) sequences have been replacedby steady-state free-precessing (cine-SSFP) protocols, whichare faster and provide improved image contrast based on ahigher signal-to-noise ratio and reduced sensitivity to flowturbulence. As a result of the inherently improved myocardial–to–blood pool contrast, cardiac volumetric data obtained bySSFP sequences have been shown to differ significantly fromthose derived by cine-GRE sequences in that they providelarger end-diastolic volumes and smaller values for ventric-ular mass when directly compared with previous spoiledgradient-echo techniques.6,7 More recent work has usedcine-SSFP protocols to yield pediatric cardiac reference

Received February 19, 2009; accepted October 1, 2009.From the Department of Heart, Thoracic, Transplantation, and Vascular Surgery (S.S., A.K.-K.), Hannover Medical School, Hannover, Germany; the

Department of Congenital Heart Disease/Pediatric Cardiology (T.K.), Deutsches Herzzentrum Berlin, Germany; Institute for Biometry and MedicalInformatics (B.P.), University of Magdeburg, Germany; the Department of Radiology (M.G.), Heart Centre Leipzig, University of Leipzig, Germany; theDepartment of Congenital Heart Disease (S.S., B.L., A.K.-K.), Heart and Diabetes Centre Bad Oeynhausen, University of Bochum, Germany; Institutefor Radiology (H.K.), Nuclear Medicine and Molecular Imaging, Heart and Diabetes Centre Bad Oeynhausen, University of Bochum, Germany; and theDivision of Imaging Sciences (P.B.), King’s College London, Guy’s and St Thomas’ Hospital, London, United Kingdom.

Drs Kuehne and Beerbaum contributed equally to this work.Presented in part as an oral abstract at the 2009 12th Annual Scientific Sessions of the Society for Cardiovascular Magnetic Resonance in Orlando,

Fla, Jan 29 to Feb 2, 2009.Correspondence to Samir Sarikouch, MD, Hannover Medical School, Department of Cardiac, Thoracic, Transplantation, and Vascular Surgery (Dir.

Prof. Dr. A. Haverich), Carl-Neuberg-Str 1. 30625 Hannover, Germany. E-mail [email protected]© 2010 American Heart Association, Inc.

Circ Cardiovasc Imaging is available at http://circimaging.ahajournals.org DOI: 10.1161/CIRCIMAGING.109.859074

65 by guest on June 18, 2012circimaging.ahajournals.orgDownloaded from


values in 60 children, divided into age groups of 8 to 11 years,12 to 14 years, and 15 to 17 years.8 However, a generalproblem in the interpretation of such data is the fact thatcardiac growth is known to be allometric, that is, the relationbetween the size of the heart and the body changes frominfancy to adulthood.9–12 Moreover, such allometric growth islikely to be different between the sexes.13 This renders anyage group definition somewhat arbitrary because sex differ-ences in the speed of maturation and cardiac growth are notaccounted for, thus limiting the scientific and clinical value ofage group–based reference data.

What is required is a definition based on the calculation ofpercentiles because this approach avoids the disadvantages ofsubjective age group definitions. Therefore, within the Ger-man Competence Network for Congenital Heart Disease, wesought to create a larger database for healthy children usingstandardized contemporary cine-SSFP and flow imagingtechniques. To compute reference centile curves, we used theLMS method described in detail by Cole (Lambda for theskew, Mu for the median, and Sigma for the generalizedcoefficient of variation).

This method uses a penalized likelihood technique to deriveage-dependent percentiles of target variables and thereby usesthe sample elements more efficiently than the use of a classifi-cation into more or less artificial age groups.14,15

Age and sex distribution of the study group was calculatedbefore start of the study, intentionally including even youngerchildren (4 to 7 years) for mathematical calculation of theinitial slope of the percentile curves. The final percentilecurves, however, were restricted to cover the age span of 8 to20 years because only in this age group, consistent breath-hold performance can be expected.

The goal was to provide a statistical calculation of age- andsex-dependent percentiles for ventricular volumes and massto serve as reference values for cardiac MRI.

MethodsStudy Population and DesignDuring a period of 26 months, we prospectively enrolled 114children and adolescents (59 female and 55 male; mean age,12.4�4.1 years; median age, 11.9 years; age range, 4 and 20 years),with a statistically prespecified age and sex pattern as based oncalculations using the LMS method. The subjects were in sinusrhythm, did not have any acquired or congenital heart disease orchronic illnesses, and did not participate in any competitive sportsactivities. Scans were performed only for study purposes, with noadditional examination of any other organ. The study was approvedby the local institutional review committee, and written informedconsent was obtained from the children’s parents or legal guardians(local registration No. 25/2005).

Body weight and height were measured, and the procedure wasexplained using a scaled-down model of the MRI scanner. Thechildren were then placed supine on the examination table and hadvector ECG leads, an abdominal belt for detection of respiratorymotion, and a 5-element cardiac receive coil attached.

Body mass index (BMI) was calculated as body weight (kg/heightin m2) and body surface area (BSA) by the Dubois formula: bodyweight (kg)0.425�height (cm)0.725�0.007184.

MRI TechniqueAll examinations were performed on a 1.5-T whole-body MRscanner (Philips Medical Systems, Intera, R11; maximum gradient

performance, 30 mT/m; slew rate, 150 T � m�1 � s�1). A 5-elementcardiac phased-array coil was used for signal acquisition, and thebody coil was used for signal transmission. A VCG-gated balancedgradient-echo sequence (SSFP) was applied to cover the whole heart.Therefore, 25 to 35 axial slices with 1 to 2 slices per breath-hold, nogap, with repetition time/echo time/flip angle, 2.8 ms/1.4 ms/60°;acquisition matrix, 160�168; number of averages, 1; SENSE reduc-tion factor, 2 (phase-encode); and 5 to 13 k-space lines per phasecorresponding to a gating width of 14 to 36 ms with an acquisitionresolution of 2.0 to 2.5�1.5 to 1.8 mm were collected. Depending onthe patient’s heart rate and physical constitution 25 to 35 phases (noheart phase interpolation) were acquired. A 5-mm slice thicknesswas used in children weighing �30 kg, whereas a 6-mm slicethickness was used for children weighing �30 kg. For youngerchildren, end-inspiratory breath-holds and for older children (�7years) end-expiratory breath-holds were used, with breath-holdsbetween 5 and 10 seconds for the VCG-gated SSFP. The procedurefollowed the standardized protocol of the MRI project of the GermanCompetence Network for Congenital Heart Defects published on thenetwork web site (www. kompetenznetz-ahf.de/en/research/mri).

In a subgroup of 29 children (mean age, 12.58�3.81 years),additional short-axis acquisition was performed using the samesequence, covering the ventricles with 12 to 14 slices, and analyzedin comparison to axial volumetry.

Additionally, quantitative phase-contrast flow measurements wereconducted in all subjects to ensure consistency by cross-checking tovolumetry. Orthogonal angulated flow measurements in the pulmo-nary artery (in the middle position between the valve and thebifurcation) and the ascending aorta (1 to 2 cm distal of the aorticbulbus) were performed using a standard, retrospectively gatedgradient-echo sequence with 30 to 40 heart phases corresponding toa temporal resolution ranging between 17 to 25 ms; 1 k-space line perphase; repetition time/echo time/flip angle, 15 ms/6.5 ms/30°;velocity-encoded value, 150 cm/s; number of signal averages, 2;slice thickness, 6 mm; acquisition matrix, 144�128; acquired spatialresolution, 2.1�2.1�6 mm; and through-plane, free-breathingphase-contrast-MRI protocol, as evaluated in children with left-to-right shunt.16

MRI AnalysisSemiautomatic threshold-based volumetric analysis using validatedcustomized software and analysis of stroke volumes in the mainpulmonary artery and the ascending aorta was performed by 1experienced observer. For interobserver variances, 10% of the datawere analyzed by a second experienced observer within the cardiacMRI core laboratory of the Competence Network for CongenitalHeart Defects. Both observers undergo regular joint consensustraining on image reading for improved interobserver and intraob-server performance.17 Furthermore, intraobserver variance was de-termined in 24 children, and interexamination variance was deter-mined in 8 children (2 separate examinations). The subgroups forinterobserver, intraobserver, and interexamination variability werechosen independently from each other.

After defining the end-systolic and end-diastolic heart phases, anauto-level function was applied for automatic endocardial contourdetection (threshold�mean value of the signal intensity defined in aregion-of-interest in the ventricular lumen and myocardial septum).If necessary, a manual adaptation of the window/level values wasallowed. Manual correction of crucial areas could be performed. Aphase-to phase cine modus was used to avoid partial volume effectsat valve planes and the inferior heart. For axial volumetry, particularattention was paid to the semilunar and for short-axis volumetry tothe atrioventricular valve level.17 Papillary muscles and trabecularstructures were excluded from the ventricular lumen but added to themyocardial muscle mass. The epicardial borders were drawn manu-ally (Figure 1).

We determined volumes of the left and right ventricles at end-systoleand end-diastole, thus measuring end-diastolic and end-systolic vol-umes. Stroke volume was calculated as the difference between thediastolic and systolic volumes. Ejection fraction, expressed as apercentage, was calculated as the ratio of stroke volume to end-di-

66 Circ Cardiovasc Imaging January 2010



astolic volumes. The volumes were indexed by dividing them byBSA, height, or body weight. Muscular mass was calculated sepa-rately for the free wall of the right ventricle, the interventricularseptum, and the left ventricle including the septum, by determiningthe volume of the particular structure in end-diastole and thenmultiplying by a factor of 1.05 g/mL.

Flow data analysis was performed offline on a computer worksta-tion using a customized computer algorithm for semiautomaticvessel border detection as described in detail and validated byBeerbaum et al16 to optimize measurement reproducibility andaccelerate image analysis.

Statistical AnalysisDescriptive statistical analysis was performed for all relevant data.Pearson correlation coefficients and Bland-Altman analysis werecalculated to compare volumetry and quantitative flow data as wellas axial versus short-axis volumetry. The coefficient of variation wascalculated to study the percentage of variability of the measurements.Significance of the observed differences between groups was testedwith a 2-sided Student t test when appropriate. A probability value of�0.05 was considered statistically significant.

Reference centile curves were computed with the LMS method byCole, using the software LMS, version 1.27, from the Institute ofChild Health (London, United Kingdom).14,15 This method describesthe distribution of the target parameter at a given age by normalapproximation after a Box-Cox transformation. The (age-dependent)estimates L, M, and S for the 3 parameters lambda (Box-Cox power),mu, and sigma (mean and coefficient of variation after transforma-tion) provide the name for this method. The estimation of theseparameters as function of the age uses a penalized maximumlikelihood approach, in which the penalty term is composed fromintegrals over the squared second derivatives expressing the smooth-

ness of the curves. Cubic splines are used to assess the function inbetween the considered distinct values of the age (full years). Thisprocess is controlled by 1 smoothing parameter for each of the L, M,and S curves. The authors give recommendations for the choice ofthe smoothing parameters dependent on the range of age, on therange of the observed parameter, and on the sample size. Based onthe L, M, and S curves, the centile curves can be derived. Age andsex distribution of the study group was calculated before start of thestudy, intentionally including children as young as possible. Becauseof a possibly reduced reproducibility for very young children, theresults are presented only in the age span from 8 to 20 years(represented by 99 children), using the children �8 years of age foran improved estimation of the initial slope (Figure 2).

ResultsSubject CharacteristicsAll 114 children were able to fully cooperate during theexaminations. No sedation or anesthesia was used. Examina-tions lasted between 25 and 40 minutes, and MRI quality wasdeemed sufficient for study purposes in all cases. Bodyweight, height, and BMI (body weight in kilograms; height incentimeters) are summarized for the total population and forage subgroups (8 to 15 years, 16 to 20 years) in Table 1.

Percentile Analysis Versus AgeGroup–Based AnalysisUsing the LMS method, reference centile curves were com-puted for ventricular volumes and mass separately for boysand girls from 8 to 20 years because of the observed

Figure 1. Volumetric analysis. A, Axial images for volumetric analysis of end-diastolic volumes of the left and right ventricles. B, Short-axis images for volumetric analysis of end-diastolic volumes of the left and right ventricles. C, Axial volumetry of the free walls of theright and left ventricles and the interventricular septum. D, Short-axis volumetry of the free walls of the right and left ventricles and theinterventricular septum.

Sarikouch et al Sex-Specific Pediatric CMR Percentiles 67



significant differences in the population when indexed forBSA and height (Table 2).

Interestingly, when indexing for weight, such sex differ-ence was only noted for ventricular mass parameters but notfor ventricular volumes (see below). We present percentilecurves for BSA and weight in this report. Percentile curvesrelated to height were not included because the noted sexdifferences and curve forms were similar to the BSA-relatedpercentiles. When analyzing pediatric patients (age 8 to 15years) versus adolescents/young adults (age 16 to 20 years),sex differences were noted in both age groups (see Tables 2and 3) but only became significant in the older group.

Effect of Sex on Relation Between VentricularVolumes and Body SizeVentricular volumes indexed to the BSA were, in the wholegroup, �10% to 15% higher in boys than in girls; this sexdifference was statistically significant for most volumetricparameters—only the left ventricular end-systolic volumewas not statistically significant. Ejection fractions of the rightand left ventricles showed no relevant sex difference. Thissex difference for ventricular volumes indexed to the bodysurface was notable but smaller (4% to 7%) when only theage group between 8 to 15 years was analyzed, and the sexdifferences were not statistically significant. For the young

adult group (age, 16 to 20 years), the sex significance wassubstantive and significant for most parameters; only the rightventricular end-systolic volume was not statistically signifi-cant (Table 2).

The sex difference in general for the ventricles was alsostatistically significant when volumes were indexed forheight. For this form of indexing ventricular volumes of boyswere �15% to 20% higher than in girls (Table 3). Again, thissex difference was notable but smaller (0% to 6%) when onlythe age group between 8 to 15 years was considered, and thesex differences were not statistically significant. For theyoung adult group, the sex difference was substantial andhighly significant (Table 3).

In contrast, when indexing volumes to body weight, the sexdifferences were much smaller, 1% to 5%, and the differenceswere not significant, neither for the total population nor forthe 2 age groups (Table 3).

Effect of Sex on Relation Between VentricularMass and Body SizeMuscle mass of the left ventricle, the right ventricle, and theinterventricular septum were in general significantly higherfor boys than girls when indexed for BSA or height. For theage group from 8 to 15 years, a significant sex difference wasonly observed for the interventricular septum indexed to

036 9921 6 123

10

5

0

15

20

25

10

5

0

15

20

25boys girls

No

age

[yea

rs] age [years]

Data of children < 8 years was onlyused for the starting slope of percentiles

Figure 2. Sex and age distribution. Study groupcomposition differentiated between boys and girls.

Table 1. Subject Characteristics

Age, y Weight, kg Height, cm BMI, kg/m2 Heart Rate, bpm

Total

Girls (n�59) 11.9�4.0 41.8�15.6 149�19 18.1�3.3 81.4�11.6

Boys (n�55) 12.9�4.1 51.8�24.0 156�22 19.9�4.3 79.4�13.2

8 to 15 y

Girls (n�35) 11.7�2.3 42.4�14.2 152�15 17.9�3.4 80.2�9.9

Boys (n�35) 11.2�2.3 42.0�16.0 148�16 18.5�3.3 82.6�11.9

16 to 20 y

Girls (n�13) 17.2�0.9 56.3�6.6 166�6 20.4�2.7 74.2�10.9

Boys (n�16) 18.1�1.3 80.1�13.9 182�6 24.2�3.6 69.9�12.2




BSA; masses indexed for height showed no significant sexdifference. In contrast, for the young adult group, the sexdifference was highly significant when muscle masses wereindexed for BSA or height (Table 4).

When ventricular masses were indexed for body weight,the general sex difference appeared smaller but still signifi-cant for the interventricular septum and the left ventricle,whereas there was no significant difference between boysand girls for the right ventricle. For the age group 8 to 15years, only the interventricular septum showed a signifi-cant sex difference, whereas the young adult group showedsignificant differences for the interventricular septum andthe left ventricle (Table 4).

Effect of Age on Ventricular Volumes and MassWhen indexing to BSA and height, there was a steady risewith increasing age of left and right ventricular end-diastolicvolumes in girls from 8 to 14 years, when the volumesreached a plateau. In boys, a plateau was reached later, at�17 years (Figure 3). End-systolic volumes demonstrated a

similar, almost parallel increase (Figure 3). Stroke volumes ofthe left and right ventricles, as derived parameters, displayedthe same slope from childhood to adolescence.

Masses of the right ventricle, the interventricular septum,and the left ventricle indexed to the BSA and height rosecontinuously with age and showed no clear plateau, with asteeper slope in boys (Figure 4).

As indexing ventricular volumes to body weight eliminatedlargely any sex difference (see above), the resulting unisexpercentile curves for right and left ventricular end-diastolicand end-systolic volumes are shown in Figure 5. End-diastol-ic and end-systolic volumes of both ventricles indexed tobody weight exhibited a slight decrease from childhood toadolescence where a plateau at �14 years was recognized(Figure 5).

Volumetric Versus Flow-Derived Stroke VolumesThere was a highly significant correlation between leftventricular stroke volume derived from volumetry and strokevolume in the ascending aorta established by flow measure-

Table 2. Effect of Sex and Age Group on Ventricular Volumes Indexed for BSA

RV-EDVi, mL/m2 RV-ESVi, mL/m2 RV-SVi, mL/m2 LV-EDVi, mL/m2 LV-ESVi, mL/m2 LV-SVi, mL/m2 Cardiac Index, L/min/m2* RV-EF, % LV-EF, %

Total

Girls (n�59) 76.9�12.7 28.6�5.4 48.2�6.8 77.9�10.8 28.5�6.2 49.3�8.1 3.98�0.7 62.8�4.3 63.4�6.1

Boys (n�55) 84.5�12.7 32.5�6.4 52.0�8.4 85.1�13.8 30.5�7.6 54.6�8.2 4.30�0.85 61.6�4.5 64.4�4.9

Difference, % 10 14 8 9 7 11 8 (�)2 2

Significance, P 0.001 0.001 0.01 0.002 0.132 0.001 0.029 0.126 0.331

8 to 15 y

Girls (n�35) 78.3�9.7 29.2�4.6 49.1�6.8 78.7�10.7 28.9�6.1 49.7�8.3 3.94�0.62 62.6�3.6 63.2�6.3

Boys (n�35) 82.9�12.6 31.3�6.1 51.6�8.4 81.9�12.9 28.2�6.7 53.7�8.5 4.42�0.85 62.3�4.3 65.7�4.9

Difference, % 6 7 5 4 (�)2 8 12 (�)1 4

Significance, P 0.090 0.117 0.165 0.254 0.613 0.048 0.010 0.704 0.061

16 to 20 y

Girls (n�13) 79.7�10.3 30.4�6.9 49.3�6.3 80.7�9.5 29.0�7.5 51.7�6.8 3.85�0.86 62.1�6.0 64.3�7.3

Boys (n�16) 90.2�10.9 36.1�6.3 54.1�7.3 93.5�10.9 35.9�7.1 57.6�5.9 4.02�0.79 60.0�4.4 61.8�4.3

Difference, % 13 19 10 16 24 11 4 (�)3 (�)4

Significance, P 0.014 0.029 0.075 0.003 0.018 0.019 0.571 0.294 0.265

*Cardiac index was determined by ventricular stroke volume multiplied by mean heart rate.

Table 3. Effect of Sex and Age Group on Ventricular Volumes Indexed for Weight and Height

RV-EDVw, mL/kg RV-ESVw, mL/kg LV-EDVw, mL/kg LV-ESVw, mL/kg RV-EDVs, mL/cm RV-ESVs, mL/cm LV-EDVs, mL/cm LV-ESVs, mL/cm

Total

Girls (n�59) 2.52�0.37 0.94�0.17 2.55�0.39 0.94�0.23 0.67�0.15 0.25�0.07 0.68�0.16 0.25�0.07

Boys (n�55) 2.59�0.45 0.99�0.20 2.60�0.45 0.93�0.21 0.79�0.20 0.31�0.09 0.80�0.22 0.29�0.01

Difference, % 3 5 2 (�)1 18 24 18 16

Significance, P 0.363 0.099 0.551 0.768 0.001 �0.001 0.001 0.016

8 to 15 y

Girls (n�35) 2.56�0.35 0.96�0.16 2.57�0.36 0.95�0.23 0.69�0.15 0.26�0.06 0.69�0.15 0.25�0.07

Boys (n�35) 2.69�0.42 1.02�0.20 2.66�0.43 0.91�0.20 0.73�0.17 0.27�0.07 0.72�0.17 0.25�0.07

Difference, % 5 6 4 (�)4 6 4 4 0

Significance, P 0.160 0.166 0.351 0.466 0.294 0.276 0.478 0.730

16 to 20 y

Girls (n�13) 2.31�0.35 0.88�0.22 2.34�0.32 0.85�0.24 0.78�0.10 0.29�0.06 0.79�0.10 0.28�0.07

Boys (n�16) 2.29�0.36 0.92�0.19 2.38�0.37 0.91�0.21 0.99�0.13 0.39�0.07 1.03�0.13 0.39�0.08

Difference, % (�)1 3 2 7 14 34 30 39

Significance, P 0.868 0.666 0.774 0.424 �0.001 �0.001 �0.001 �0.001




ments using phase-contrast MRI (r�0.95, P�0.001). Thecorrelation between right ventricular stroke volume and mainpulmonary flow was accordingly significant (r�0.94,P�0.001). The coefficient of variation was 6.47 for the leftventricular stroke volume and 6.67 for the right ventricularstroke volume. The limits of agreement (Bland-Altman anal-ysis) are shown in Table 5.

Axial Volumetry Versus Short-Axis VolumetryComparison between axial and short-axis volumetry revealedhigh and significant correlations with low coefficients ofvariation (CV) for end-diastolic volumes of the ventricles(CV-RV, 2.23; CV-LV, 2.46). End-systolic volumes dis-played more variation than end-diastolic volumes, with aslightly higher variation for the right ventricle (CV-RV, 6.02;CV-LV, 4.68). Muscle mass of the left ventricle and theinterventricular septum revealed the same variation as end-systolic ventricular volumes (CV-LV-MM, 5.93; CV-IVS-MM, 6.02). Right ventricular muscle mass demonstrated thehighest variation (CV-RV-MM, 9.78). The limits of agree-ment (Bland-Altman analysis) are shown in Table 6.

Intraobserver, Interobserver, andInterexamination VariancesThere was good intraobserver agreement, with the highestvariation in the end-systolic volumes of the ventricles andmasses of the ventricles.17 Interobserver comparison showedmore variation, with the highest variation also in end-systolicvolumes and masses of the ventricles. Interexaminationvariances demonstrated the greatest variation in the leftventricular end-systolic volume, right ventricular stroke vol-ume, and mass of the right ventricle. Mean differences, limitsof agreement, coefficients of variation (each in percent), andthe correlation coefficients are shown in Table 7.

DiscussionThe relation between the size of the heart and the body isknown to be allometric,18 which means that this relationchanges from infancy to adulthood. This makes the genera-tion of normative data challenging. However, the ability to

accurately measure volumetric and mass parameters of rightand left ventricular function, adequately scaled to body size,is of critical importance both for clinical care and forcardiovascular health research.18

Sex Differences in Cardiac Measures in theDeveloping ChildThe impact of sex on such normative data has always been amatter of concern in pediatric cardiology because sex influ-ence is (1) likely to be different before, during, and afterpuberty, and (2) the timing of the adolescent growth spurt inindividuals is different between the sexes. Although data arerelatively limited, a number of previous studies suggest theexistence of sex differences of cardiac measures even beforepuberty. In a pathoanatomic series (n�200, 0 to 19 years),variables such as heart weight, ventricular wall thickness, andcardiac valve circumferences were found to be usually largerin boys.11 Moreover, for echocardiographic parameters ofheart size, a small sex difference was identified in 7- to11-year-old children in the Bogalusa Heart Study.19 Zilber-man et al20 found in a large echocardiographic investigation(n�748, 434 boys) using standardized 2D measures that boyshad statistically larger valve diameters at all ages. Additionalevidence of prepubertal sex differences was identified fordiastolic left ventricular function parameters (Doppler echo-cardiography).21 These subtle prepubertal sex differencesonly became statistically significant in larger series.20,21 Incontrast, a large echocardiographic study in more than 2000healthy infants and children in central Europe revealed no sexdifferences, but the authors reasoned that this was probablydue to interobserver bias.22

Using the LMS Method for ReferenceCentile ComputationIt is mainly for this reason that we sought to deliversex-specific percentiles of cardiac MR volumetry in the firstplace, rather than presenting comparisons by age categories.Hence, we generated reference centile curves for children andadolescents from 8 to 20 years for ventricular volumes andmass using the LMS method described by Cole.14 This

Table 4. Effect of Sex and Age Group on Ventricular Mass Indexed for BSA, Weight, and Height

RV-MMi, g/m2 IVS-MMi, g/m2 LV tot-MMi, g/m2 RV-MMw, g/kg IVS-MMw, g/kg LV tot-MMw, g/kg RV-MMs, g/cm IVS-MMs, g/cm LV tot-MMs, g/cm

Total

Girls (n�59) 17.6�4.8 15.2�2.8 47.6�8.2 0.58�0.15 0.49�0.08 1.56�0.3 0.15�0.05 0.14�0.04 0.42�0.1

Boys (n�55) 20.5�7.5 18.5�4.2 55.9�12.3 0.61�0.17 0.56�0.09 1.69�0.3 0.19�0.09 0.18�0.06 0.53�0.2

Difference, % 16 22 17 5 14 8 26 29 26

Significance, P 0.015 �0.001 �0.001 0.267 �0.001 0.009 0.003 �0.001 �0.001

8 to 15 y

Girls (n�35) 17.3�4.7 15.2�2.8 48.2�8.2 0.56�0.15 0.49�0.07 1.58�0.27 0.15�0.05 0.13�0.04 0.42�0.10

Boys (n�35) 18.3�5.5 16.9�3.7 52.1�10.6 0.59�0.14 0.55�0.09 1.68�0.28 0.16�0.07 0.15�0.04 0.46�0.14

Difference, % 6 11 8 5 12 6 7 15 10

Significance, P 0.382 0.017 0.086 0.543 0.006 0.108 0.387 0.107 0.209

16 to 20 y

Girls (n�13) 20.4�5.5 17.1�1.9 51.2�7.3 0.59�0.17 0.49�0.07 1.49�0.25 0.20�0.05 0.17�0.02 0.50�0.07

Boys (n�16) 26.9�7.9 22.9�2.7 67.6�7.7 0.68�0.22 0.58�0.09 1.71�0.25 0.29�0.09 0.25�0.03 0.75�0.10

Difference, % 32 34 32 15 18 15 45 47 50

Significance, P 0.019 �0.001 �0.001 0.243 0.007 0.022 0.002 �0.001 �0.001




boys

20

40

60

80

100

120

140

8 10 12 14 16 18 20

age [yrs]

RV_

EDV

i[m

l/m2

]

P97P90P75P50P25P10P3

boys

0

10

20

30

40

50

60

8 10 12 14 16 18 20

age [yrs]

RV-

ESV

i[m

l/m2

]

P97P90P75

P50P25P10P3

girls

0

10

20

30

40

50

60

8 10 12 14 16 18 20

age [yrs]

RV-

ESV

i [m

l/m2

]

P97P90P75

P50P25P10P3

girls

20

40

60

80

100

120

140

8 10 12 14 16 18 20

age [yrs]

LV-E

DV

i [m

l/m2 ] P97

P90P75P50P25P10P3

boys

0

10

20

30

40

50

60

8 10 12 14 16 18 20

age [yrs]

LV-E

SVi [

ml/m

2]

P97P90P75P50P25P10P3

girls

0

10

20

30

40

50

60

8 10 12 14 16 18 20age [yrs]

LV-E

SVi [

ml/m

2] P97

P90P75P50P25P10P3

boys

20

40

60

80

100

120

140

8 10 12 14 16 18 20

age [yrs]

LV-E

DV

i [m

l/m2

]

P97P90P75P50P25P10P3

girls

20

40

60

80

100

120

140

8 10 12 14 16 18 20

age [yrs]

RV-

EDV

i [m

l/m2

]

P97P90P75P50P25P10P3

Figure 3. Reference percentile curves for ventricular volumes indexed for BSA for boys and girls. RV-EDVi [mL/m2] indicates right ven-tricular end-diastolic volume in mL per m2 BSA; RV-ESVi [mL/m2], right ventricular end-systolic volume in mL per m2 BSA; LV-EDVi[mL/m2], left ventricular end-diastolic volume in mL per m2 BSA; LV-ESVi [mL/m2], left ventricular end-systolic volume in mL per m2

BSA; P 50, percentile in which 50% of boys or girls of the respective age have smaller or equal ventricular volumes; P 97, percentile inwhich 97% of boys or girls of the respective age have smaller or equal ventricular volumes.




statistical technique has a number of advantages that render itparticularly useful in the pediatric age group. It avoids allmore or less artificial age categories, and, unlike regressionanalyses, LMS does not depend on linear relationships betweencardiovascular parameters and body size, and it is well knownthat these do not exist.15,18 An additional parameter also allowsthe fitting of nongaussian distributions. Moreover, LMS-derivedcentiles can display different ranges of distribution at differenttime points, whereas regression analyses assume constant rangesof distribution.14,15 This makes the LMS method generally suitedfor parameters of cardiac function with known or suspected sexinfluence.13 In children, this is particularly valuable as the

generation of normative data are further complicated by thedifference in onset and speed of pubertal growth spurt betweengirls and boys.13–15,23

Study Findings: Sex and Age Effects onVentricular Volumes and MassFor the total population, we observed statistically highly signif-icant sex differences when indexing for BSA and for height forventricular volumes and mass parameters. In keeping with theliterature, not surprisingly, these differences were less clear forthe younger age group than for the adolescents (Table 2).

boys

0

10

20

30

40

50

60

8 10 12 14 16 18 20age [yrs]

RV-

MM

i [g/

m2

]

P97P90P75

P50P25P10P3

girls

0

10

20

30

40

50

60

8 10 12 14 16 18 20

age [yrs]

RV-

MM

i [g/

m2 ]

P97P90P75P50P25P10P3

boys

5

10

15

20

25

30

35

8 10 12 14 16 18 20

age [yrs]

IVS-

MM

i [g/

m2

]

P97P90P75P50P25P10P3

girls

5

10

15

20

25

30

35

8 10 12 14 16 18 20

age [yrs]

IVS-

MM

i [g/

m2

]

P97P90P75P50P25P10P3

boys

0

20

40

60

80

100

8 10 12 14 16 18 20age [yrs]

LV-to

t-MM

i [g/

m2 ]

P97P90P75P50P25P10P3

girls

0

20

40

60

80

100

8 10 12 14 16 18 20age [yrs]

LV-to

t-MM

i[g/

m2 ] P97

P90P75P50P25P10P3

Figure 4. Reference percentile curves for ventricular masses indexed for BSA for boys and girls. RV-MMi [g/m2] indicates right ventric-ular mass in g per m2 BSA, excluding the septum; IVSi [g/m2], interventricular septum mass in g per m2 BSA; LV-tot MMi [g/m2], totalleft ventricular mass in g per m2 BSA, including the septum; P 50, percentile in which 50% of boys or girls of the respective age havesmaller or equal ventricular mass; P 97, percentile in which 97% of boys or girls of the respective age have smaller or equal ventricularmass.




Interestingly, when indexing to body weight, there was nosignificant sex influence for volume parameters both beforeand after puberty, whereas there still was clear sex effect onmass parameters, particularly after puberty (Table 4). Hence,at least for cardiac volumes such as end-diastolic and end-systolic volume and derived parameters, weight-related uni-sex reference centiles may be considered for daily practice asshown in Figure 5, knowing, however, that BSA indexing ismore popular in pediatric cardiology. For mass parameters,relation to weight, however, does not provide this advantage,and sex-specific centiles are needed here as well (Figure 4).

More generally, ventricular volumes indexed for BSA andheight demonstrated allometric growth with steady gentle en-largement from childhood to adolescence, in which end-diastolic

and end-systolic volumes reached plateaus comparable to adultreference values.24 Masses of the ventricles and the interventric-ular septum, also indexed for BSA, exhibited a steady increasewithout a clear plateau. These results correspond to a recentlypublished cross-sectional study of sex- and age-specific normalvalues of the left ventricle from adolescence to late adulthood, inwhich the peak of the end-diastolic volume was found before 20years of age, but the peak for the left ventricular muscle massonly after 30 years.25

Volumetric Parameter Measures: Comparison tothe LiteratureWe observed small but notable differences to the volumetricMRI data for children as published so far. End-diastolic and

girls + boys

1,0

1,5

2,0

2,5

3,0

3,5

4,0

8 10 12 14 16 18 20

age [yrs]

RV-

EDV w

[ml/k

g]

P97P90P75P50P25P10P3

girls + boys

0,0

0,3

0,6

0,9

1,2

1,5

1,8

8 10 12 14 16 18 20age [yrs]

RV-

ESV w

[ml/k

g]

P97P90P75P50P25P10P3

girls + boys

1,0

1,5

2,0

2,5

3,0

3,5

4,0

8 10 12 14 16 18 20

age [yrs]

LV-E

DV w

[ml/k

g]

P97P90P75P50P25P10P3

girls + boys

0,0

0,3

0,6

0,9

1,2

1,5

1,8

8 10 12 14 16 18 20

age [yrs]

LV-E

SV [m

l/kg]

P97P90P75

P50P25P10P3

Figure 5. Reference unisex percentile curves for ventricular volumes indexed for weight. RV-EDVw [mL/kg] indicates right ventricularend-diastolic volume in mL per kg body weight (BW); RV-ESVw [mL/kg], right ventricular end-systolic volume in mL per kg BW;LV-EDVw [mL/kg], left ventricular end-diastolic volume in mL per kg BW; LV-ESVw [mL/kg], left ventricular end-systolic volume in mLper kg BW; P 50, percentile in which 50% of boys and girls of the respective age have smaller or equal ventricular volumes; P 97, per-centile in which 97% of boys and girls of the respective age have smaller or equal ventricular volumes.

Table 5. Phase-Contrast Flow Versus Volumetric Measurements

Variables Mean�SD

Correlations

CV, %

Limits of Agreement, �2 SD/�2 SD

R P Lower (�95% CI) Upper (�95% CI)

LV-SV PC flow, mL 72.65�25.4

LV-SV vol, mL 74.23�27.8 0.950 �0.001 6.47 �19.14 (�22.0; �16.26) 15.96 (13.09; 18.83)

RV-SV PC flow, mL 73.4�25.7

RV-SV vol, mL 71.7�26.1 0.942 �0.001 6.67 �15.90 (�18.78; �13.01) 19.27 (16.38; 22.16)

CV indicates coefficient of variation; LV-SV, left ventricular stroke volume; PC, phase-contrast flow; vol, volumetric.




end-systolic volumes were slightly higher than described byLorenz5 and Helbing.4 These differences may be explained bythe use of contemporary SSFP sequences with better blood-myocardium differentiation for the older reference data6 andby threshold-based semiautomatic segmentation versus man-ual segmentation for newer pediatric data. No clinicallysignificant difference for volumetric parameters was recog-nizable by the use of short-axis volumetry by Robbers-Visseret al.8 Contemporary SSFP sequences and new segmentationtools may account for the relatively lower myocardial massesof the ventricles observed in this study because computer-assisted segmentation is likely to be more precise thanmanual border tracing of the endocardium, particularly forthe right ventricle. There was a very good correlation torecently published echocardiographic data using 2D and3D approaches.26

MR Internal Validation and Observer VarianceIn line with Jeltsch et al,27 there was a very high correlationbetween volumetric-derived and flow-derived stroke vol-umes, which allowed simple internal quality control ofvolumetric data. Comparison between axial and short-axisvolumetry demonstrated no relevant differences, as reportedby Jauhiainen et al28 and in contrast to Alfakih et al, whofound short-axis volumes of the right ventricle to be larger inadult volunteers.29 Axial acquisition had the advantage ofallowing for additional phasic volumetry of the atria forfurther studies and has also recently be reported to be morereproducible than short-axis slices for ventricular volumetryin patients with corrected tetralogy of Fallot.30 Intraobserverand interobserver variability demonstrated results comparableto other studies with good agreement for end-diastolic vol-umes and less agreement for end-systolic volumes and

Table 6. Axial Volumetry Versus Short-Axis Volumetry

Variables Mean�SD

Correlations

CV, %

Limits of Agreement, �2 SD/�2 SD

R P Lower (�95% CI) Upper (�95% CI)

RV_EDVi axial, mL/m2 77.11�8.89

RV_EDVi SA, mL/m2 74.88�10.22 0.920 �0.001 2.23 �5.8 (�8.50; �3.18) 10.31 (7.64; 12.96)

LV_EDVi axial, mL/m2 76.01�8.81

LV_EDVi SA, mL/m2 74.51�8.68 0.924 �0.001 2.46 �5.31 (�7.56; �2.07) 8.30 (6.06; 10.55)

RV_ESVi axial, mL/m2 29.28�5.22

RV_ESVi SA, mL/m2 27.58�6.162 0.867 �0.001 6.02 �4.44 (�6.46; �2.41) 7.84 (5.81; 9.87)

LV_ESVi axial, mL/m2 26.29�5.22

LV_ESVi SA, mL/m2 25.46�5.79 0.911 �0.001 4.68 �3.48 (�4.91; �2.06) 5.14 (3.72; 6.56)

RV_MMi axial, g/m2 21.31�6.96

RV_MMi SA, g/m2 22.94�6.29 0.880 �0.001 9.78 �8.26 (�10.45; �6.01) 4.99 (2.80; 7.17)

LV_FW-MMi axial, g/m2 35.87�8.23

LV_FW-MMi SA, g/m2 36.39�9.07 0.890 �0.001 5.93 �8.79 (�11.51; �6.06) 7.74 (5.02; 10.47)

IVS_MMi axial, g/m2 16.65�3.66

IVS_MMi SA, g/m2 17.27�3.99 0.862 �0.001 6.10 �4.69 (�6.03; �3.35) 3.44 (2.10; 4.78)

CV indicates coefficient of variation; SA, short-axis; LV_FW-MM, left ventricular free wall.

Table 7. Intraobserver, Interobserver, and Interexamination Variability

LV-EDV LV-ESV LV-SV RV-EDV RV-ESV RV-SV LV-MM IVS-MM RV-MM

Intraobserver variability

Mean difference 2.6 4.9 1.2 3.1 0.5 4.5 4.6 3.7 2.9

LOA �5.9 to 11.1 �8.6 to 18.5 �10.3 to 12.8 �7.6 to 13.7 �16.0 to 16.9 �7.5 to 16.5 �10.8 to 20.0 �14.3 to 21.8 �20.8 to 26.5

COV 2.6 4.2 2.9 3.5 4.8 4.3 4.6 5.4 6.9

R 0.994 0.994 0.984 0.990 0.979 0.986 0.990 0.991 0.973

Interobserver variability

Mean difference 3.4 2.5 3.8 8.3 13.1 5.6 �16.2 �6.5 �13.8


COV 3.1 5.4 4.1 6.5 10.8 6.4 14.5 8.4 22.3

R 0.992 0.972 0.985 0.976 0.965 0.957 0.854 0.962 0.658

Interexamination variability

Mean difference 1.3 �1.1 3.2 0.8 �1.5 2.7 3.7 �1.1 2.4


COV 5.7 11.0 3.1 5.0 5.2 6.8 6.7 8.3 10.5

R 0.684 0.351 0.933 0.737 0.861 0.577 0.790 0.638 0.770

Except for the correlation coefficient R, all values are given in percentages. LOA indicates limits of agreement; COV, coefficient of variation; R, correlation coefficient.




muscle mass with right ventricular variability higher than leftventricular variability.2,17 Interexamination agreement wassatisfactory, with the highest variability for left ventricularend-systolic volume and right ventricular mass, which corre-lates to results in adults and supports that MRI can adequatelyassess ventricular size and function in children.2

LimitationsThe age of the children studied allowed no calculation ofpercentile curves for infants and toddlers. Reference valuesfor neonates and infants would be of great interest, forexample, regarding thresholds for biventricular repair in thesetting of ventricular imbalance.

We used accepted modern statistical calculation models tocreate pediatric reference data for cardiac volumes and mass.Nevertheless, true normative, sex-specific data from birth toadulthood would be preferable, but this would have requiredincluding a much larger number of children (�1000) toideally also account for ethnic differences. The ideal bodysize measure to relate volumetric variables to may be neitherweight, height, nor BSA. It may well be that—once a muchlarger normative database is accumulated—other body sizederivates may prove better suited, and these may also bedifferent for different parameters of cardiac size and function.

ConclusionsAge- and sex-specific percentiles for ventricular volumes andmass of healthy children and adolescents from 8 to 20 yearshave been established by cardiac MRI to serve as referencevalues in the evaluation of acquired and congenital heartdisease. Significant sex differences were noted when index-ing volumes to BSA or height. In contrast, indexing ventric-ular volumes (but not mass) to body weight largely elimi-nated the sex differences and may therefore be attractive fordaily use in pediatric cardiology.

AcknowledgmentsWe thank Anne Gale, ELS, for editorial assistance.

Sources of FundingThis study was part of the Magnetic Resonance Imaging Project ofthe Competence Network for Congenital Heart Defects funded by theGerman Federal Ministry of Education and Research (FKZ01G10210).

DisclosuresNone.

References1. Haddad F, Hunt SA, Rosenthal DN, Murphy DJ. Right ventricular

function in cardiovascular disease, I: anatomy, physiology, aging, andfunctional assessment of the right ventricle. Circulation. 2008;117:1436–1448.

2. Keenan NG, Pennell DJ. CMR of ventricular function. Echocardiography.2007;24:185–193.

3. Pennell DJ, Sechtem UP, Higgins CB, Manning WJ, Pohost GM, Rade-makers FE, van Rossum AC, Shaw LJ, Yucel EK. Clinical indications forcardiovascular magnetic resonance (CMR): Consensus Panel report. EurHeart J. 2004;25:1940–1965.

4. Helbing WA, Rebergen SA, Maliepaard C, Hansen B, Ottenkamp J,Reiber JH, de Roos A. Quantification of right ventricular function withmagnetic resonance imaging in children with normal hearts and withcongenital heart disease. Am Heart J. 1995;130:828–837.

5. Lorenz CH. The range of normal values of cardiovascular structures ininfants, children, and adolescents measured by magnetic resonanceimaging. Pediatr Cardiol. 2000;21:37–46.

6. Moon JC, Lorenz CH, Francis JM, Smith GC, Pennell DJ. Breath-holdFLASH and FISP cardiovascular MR imaging: left ventricular volumedifferences and reproducibility. Radiology. 2002;223:789–797.

7. Alfakih K, Plein S, Thiele H, Jones T, Ridgway JP, Sivananthan MU.Normal human left and right ventricular dimensions for MRI as assessedby turbo gradient echo and steady-state free precession imagingsequences. J Magn Reson Imaging. 2003;17:323–329.

8. Robbers-Visser D, Boersma E, Helbing WA. Normal biventricularfunction, volumes, and mass in children aged 8 to 17 years. J Magn ResonImaging. 2009;29:552–559.

9. de Simone G, Daniels SR, Devereux RB, Meyer RA, Roman MJ, deDivitiis O, Alderman MH. Left ventricular mass and body size in nor-motensive children and adults: assessment of allometric relations andimpact of overweight. J Am Coll Cardiol. 1992;20:1251–1260.

10. de Simone G, Devereux RB, Daniels SR, Mureddu G, Roman MJ,Kimball TR, Greco R, Witt S, Contaldo F. Stroke volume and cardiacoutput in normotensive children and adults: assessment of relations withbody size and impact of overweight. Circulation. 1997;95:1837–1843.

11. Scholz DG, Kitzman DW, Hagen PT, Ilstrup DM, Edwards WD. Age-related changes in normal human hearts during the first 10 decades of life,I (Growth): a quantitative anatomic study of 200 specimens from subjectsfrom birth to 19 years old. Mayo Clin Proc. 1988;63:126–136.

12. Kitzman DW, Scholz DG, Hagen PT, Ilstrup DM, Edwards WD. Age-related changes in normal human hearts during the first 10 decades of life,II (Maturity): a quantitative anatomic study of 765 specimens fromsubjects 20 to 99 years old. Mayo Clin Proc. 1988;63:137–146.

13. de Simone G, Devereux RB, Daniels SR, Meyer RA. Gender differencesin left ventricular growth. Hypertension. 1995;26:979–983.

14. Cole TJ. The LMS method for constructing normalized growth standards.Eur J Clin Nutr. 1990;44:45–60.

15. Cole TJ, Green PJ. Smoothing reference centile curves: the LMS methodand penalized likelihood. Stat Med. 1992;11:1305–1319.

16. Beerbaum P, Korperich H, Barth P, Esdorn H, Gieseke J, Meyer H.Noninvasive quantification of left-to-right shunt in pediatric patients:phase-contrast cine magnetic resonance imaging compared with invasiveoximetry. Circulation. 2001;103:2476–2482.

17. Beerbaum P, Sarikouch S, Barth P, Kropf S, Kelter-Kloepping A, Gut-berlet M, Berger F, Kuehne T. Cardiac function by MRI in congenitalheart disease: impact of consensus training on inter-institutional variance.J Magn Reson Imaging. 2009;30:956–966.

18. Dewey FE, Rosenthal D, Murphy DJ Jr, Froelicher VF, Ashley EA. Doessize matter? Clinical applications of scaling cardiac size and function forbody size. Circulation. 2008;117:2279–2287.

19. Burke GL, Arcilla RA, Culpepper WS, Webber LS, Chiang YK,Berenson GS. Blood pressure and echocardiographic measures inchildren: the Bogalusa Heart Study. Circulation. 1987;75:106–114.

20. Zilberman MV, Khoury PR, Kimball RT. Two-dimensional echocardio-graphic valve measurements in healthy children: gender-specific dif-ferences. Pediatr Cardiol. 2005;26:356–360.

21. O’Leary PW, Durongpisitkul K, Cordes TM, Bailey KR, Hagler DJ, TajikJ, Seward JB. Diastolic ventricular function in children: a Doppler echo-cardiographic study establishing normal values and predictors ofincreased ventricular end-diastolic pressure. Mayo Clin Proc. 1998;73:616–628.

22. Kampmann C, Wiethoff CM, Wenzel A, Stolz G, Betancor M, Wip-permann CF, Huth RG, Habermehl P, Knuf M, Emschermann T,Stopfkuchen H. Normal values of M mode echocardiographic mea-surements of more than 2000 healthy infants and children in centralEurope. Heart. 2000;83:667–672.

23. Lopez-Siguero JP, Garcia JM, Castillo Jde D, Molina JA, Cosano CR,Ortiz AJ Cross-sectional study of height and weight in the populationof Andalusia from age 3 to adulthood. BMC Endocr Disord. 2008;8(Suppl 1):S1.

24. Hudsmith LE, Petersen SE, Francis JM, Robson MD, Neubauer S.Normal human left and right ventricular and left atrial dimensions usingsteady state free precession magnetic resonance imaging. J CardiovascMagn Reson. 2005;7:775–782.




25. Cain PA, Ahl R, Hedstrom E, Ugander M, Allansdotter-Johnsson A,Friberg P, Arheden H. Age and gender specific normal values of leftventricular mass, volume and function for gradient echo magneticresonance imaging: a cross sectional study. BMC Med Imaging.2009;9:2.

26. Poutanen T, Jokinen E. Left ventricular mass in 169 healthy children andyoung adults assessed by three-dimensional echocardiography. PediatrCardiol. 2007;28:201–207.

27. Jeltsch M, Ranft S, Klass O, Aschoff AJ, Hoffmann MH. Evaluation ofaccordance of magnetic resonance volumetric and flow measurements indetermining ventricular stroke volume in cardiac patients. Acta Radiol.2008;49:530–539.

28. Jauhiainen T, Jarvinen VM, Hekali PE. Evaluation of methods for MRimaging of human right ventricular heart volumes and mass. Acta Radiol.2002;43:587–592.

29. Alfakih K, Plein S, Bloomer T, Jones T, Ridgway J, Sivananthan M.Comparison of right ventricular volume measurements between axial andshort axis orientation using steady-state free precession magnetic res-onance imaging. J Magn Reson Imaging. 2003;18:25–32.

30. Fratz S, Schuhbaeck A, Buchner C, Busch R, Meierhofer C, Martinoff S,Hess J, Stern H. Comparison of accuracy of axial slices versus short-axisslices for measuring ventricular volumes by cardiac magnetic resonancein patients with corrected tetralogy of Fallot. Am J Cardiol.2009;103:1764–1769.

CLINICAL PERSPECTIVEThe present study examined a healthy pediatric group large enough to serve as a reference for volumetric cardiac MRI.Age- and sex-specific percentiles for right and left ventricular volumes and mass of healthy children and adolescents from8 to 20 years of age were established to serve as reference values in the evaluation of acquired and congenital heart disease.Significant sex differences were noted when indexing volumes to body surface area or height. In contrast, indexingventricular volumes (but not mass) to body weight largely eliminated the sex differences and may therefore be attractivefor daily use in pediatric cardiology. An individual patient’s progression of ventricular size in relation to these percentilecurves may inform the timing of or the response to medical or surgical treatment. This potentially will facilitate a moreaccurate diagnosis of ventricular dilation and ventricular hypertrophy by cardiac MRI in this age group, with potentialimpact on management decisions.




sex-specific pediatric percentiles for ventricular …...are faster and provide improved image...

Documents