Correlation of left ventricular angiographie casts and biplane left ventricular volumetry in infants and children

Correlation of left Ventricular Angiographic Casts and Biplane Left Ventricular Volumetry in Infants and Children TOSHIHIRO INO, MD, LEE N. BENSON, MD, HAVERJ MIKAILIAN, RRT, ROBERT M. FREEDOM, MD, and RICHARD D. ROWE, MD

To calculate left ventricular (LV) volumes from biplane cineangiography obtained in nonstandard views, regression equations were developed from LV casts of known volume. Volumes were calculated by the area-length method from casts ranging from 1.4 to 48.9 ml obtained from 30 postmortem cases with heart dlsease. The casts were divided into 2 groups: group I (n = 17) with abnormal and group II (n = 13) with normal right ventricular hemodynamics. Biplane cinegrams were taken in the anterolateral, anterior and long axial oblique, hepatoclavicular and sitting-up projections. The true volume of each cast was determined from its weight and spectfic gravity. In both groups, excel-

A

ngiocardiographic measurement of left ventricular (LV) volume is an important tool in assessing function in patients with heart disease.l-lo Commonly used approaches of volume determination are either the arealength method (anteroposterior and lateral projection) described by Dodge et aPJ2 or the single-plane modification (oblique projections] introduced by Kennedy et a1.13J4 The area-length method assumes the LV cavity is represented by a 3-dimensional ellipsoid. In using casts of known volumes, other investigators15J6 have found that the measured volume by the area-length method was affected by the shape of the cast, cardiac

From the Department of Pediatrics, Division of Cardiology, The Variety Club Cardiac Catheterization Laboratories and The Hospital for Sick Children, Toronto, Ontario, Canada. Dr. Ino is supported by a grant from the Research Institute in The Hospital for Sick Children, Toronto, Ontario, Canada. Manuscript received April 23, 1987; revised manuscript received September 14,1987, and accepted September 16. Address for reprints: Lee N. Benson, MD, The Hospital for Sick Children, Room 4515, 555 University Avenue, Toronto, Ontario, Canada M5G 1X8.

lent correlations were obtained between measured and true volumes (r = 0.92 to 0.99) in all projections, although each projection overestimated the true volume (slope value
phase or spatial orientation and resulted in an overestimation of the true volume. Consequently, regression equations were required to adjust for these variables to bring the measured volume within the true volume range. In addition, the majority of previous investigations of volume determination were obtained using casts made from hearts without heart disease 3-12v15-18 which may also influence the regression analysis. Because of the complex nature of congenital heart disease, current diagnostic angiographic imaging relies less on traditional positioning (anterolateral, right and left anterior obliques] and more on various axial oblique positions as described by Bargeron and Elliot.22,23Therefore, appropriate regression formulae are also required to correct measured volume using these varied projections. We assessedthe accuracy of LV volume determinations using casts obtained from postmortem specimens of patients with congenital heart disease and evaluated the results of LV volume measurement performed in 5 biplane projections. Left ventricular casts: Thirty LV casts were obtained from postmortem hearts of patients with congenital heart disease. After removal of blood clots from

BIPLANE LEFT VENTRICULAR

442

0

10

LV

20 true

VOLUMETRY

30 volume

40

50

(ml)

FIGURE 1. Distribution of true volume of the original human left and right casts. Group I = casts with abnormal right ventricular (RV) hemodynamlcs; group II = casts with normal right ventricular hemodynamlcs. LV = left ventricular.

both ventricles, the mitral and tricuspid valves were sewn closed and the walls glued together with a cyanoacrylate compound (910 Adhesive, Permabond International Division]. A silicon rubber material (3110 RTV, Dow Corning Corp.] mixed with barium sulfate powder (E-Z-EM Co.) was injected simultaneously into both ventricles through the aortic and pulmonary valves. After the casts had hardened, the left and right ventricular free walls were incised and both casts carefully removed. Right ventricular hemodynamics: The clinical and hemodynamic data were reviewed if available. The hearts studied consisted of 5 with tetralogy of Fallot, 5 with secundum atria1 septal defect, 3 with complete transposition of great arteries with intact ventricular septum, 4 with myocarditis, 4 with endocardial fibroelastosis, 2 with total anomalous pulmonary venous return, 3 with ventricular septal defect, 2 with coarctation of aorta, 1 with congenital mitral stenosis with severe pulmonary hypertension and 1 with chronic car pulmonale. Mean age at autopsy was 3 f 5.4 years (range 1 day to 23 years), 7 had undergone cardiac. surgical procedures (3 open and 4 closed]. Preliminary data analysis (Figure 1) suggested that 2 groups could be identified according to right ventricular hemodynamics. For analysis, therefore, the casts so obtained were divided into 2 groups. In group I (n = l7), the shape of the left ventricle was not elliptical but prolate due to abnormal right ventricular hemodynamics (Figure 2). These casts included those with atria1 septal defect, tetralogy of Fallot, transposition of great arteries with intact ventricular septum, total anomalous pulmonary venous return, congenital mitral stenosis with severe pulmonary hypertension and car pulmonale. In group II (n = l3), the shape of the left ventricle was relatively elliptical, associated with normal right ventricular hemodynamics [Figure 2). These casts included those with endocardial fibroelastosis, myocar-

ditis, ventricular septal defect and coarctation of aorta. The true volume of each cast was obtained by weight in grams divided by the specific gravity of silicon rubber (1.17 g/cm3). Volume calculations: Casts were placed on a goniometer designed in our laboratory and positioned on the catheterization table for a biplane cinegram. The vertical and horizontal image intensifiers were fixed in the anteroposterior and lateral positions, respectively. The reference position for the cast was formed by a line connecting the apex with the most distal part of mitral valve forming a 35Oangle with vertical and 35’ angle with lateral planes (Figure 3). The casts were filmed in anteroposterior and lateral, right and left anterior oblique, long axial oblique, hepatoclavicular and sitting-up positions. In the long axial oblique projection, the casts were slanted 30° right anterior oblique relative to the vertical image intensifier and rotated approximately 15 to 20’ with the apex to the right side of the table. In the hepatoclavicular projec-

FIGURE 2. Left ventricular (LV) casts obtained from specimens with congenital heart disease. The shape of the LV casts observed from anteroposterlor (right) and lateral (left) views were prolate in group I (fop) and elllptlcal in group II (boffom).

vertical image intensifier ._.__-----__ - --.-__

--..I ---. :

FIGURE 3. The startlng position of the cast in anteroposterior and lateral view. The left ventricular reference axis connects the apex with most distal part of the mitral valve, and forms a 35’ angle with the frontal and lateral plane.

February 15, 1988

TABLE I

Regression

Equations and Associated

THE AMERICAN

Statistics

JOURNAL

for Left Ventricular

y=ai-bx

Anteroposterior and lateral Right and left anterior oblique Long axial oblique Hepatoclavicular Sitting up

y =

0.3 + 0.63x’

y = -0.9

+ 0.69x’

y = -0.3 + 0.69x+ y = -0.9 + 0.61x+ y = 0.7 t 0.64~’

443

Volumetry

SEE

R

+ 0.84x’

3.9

0.96

y = -0.3

+ 0.77x’

3.5

0.97

y = -1.0 y = -0.8 y = -1.1

+84x+ t 0.85x+ t 0.88x+

2.8 4.0 5.2

0.98 0.95 0.92

y=a+bx

SEE

R

2.1

0.98

y = -0.3

2.6

0.96

2.7 2.1 1.3

0.96 0.98 0.99

A = intercept; b = slope; R = correlation coefficient: corrected volume. * p <0.05, 7 p <0.025 (group I vs group II).

SEE = standard error of estimate:

tion, the casts were slanted 20’ left anterior oblique relative to the vertical image intensifier, rotated 15 to 20” with the apex to the right of the table and the goniometer tilted forward 45’ (caudal cranial angulation). After each cinefilm a grid of l-cm squares was filmed at the cast center, parallel to each image intensifier. The cast outlines were projected and traced by hand to the outer margins of each silhouette. The cast volumes were calculated using the arealength method of Dodge et al,llvlz in which the LV cavity is represented as ellipsoid of revolution: V = (4?r/3)(Dvertical/2)(Dhorizontal/2)(Lmax/2), where Dvertical = minor axis in the anteroposterior view (cm), Dhorizontal = minor axis in the lateral view [cm), Lmax = major axis (cm) and V = volume (ml), and each is corrected for magnification distortion. The major axis is the longest measured length whether it was in the anteroposterior or lateral projection. The 2 minor axes, Dvertical and Dhorizontal, were calculated as: Dvertical = 4 Avertical/r Lvertical, and Dhorizontal = 4 Ahorizontal/a

Volume 61

Group II

Group I Projection

OF CARDIOLOGY

Lhorizontal,

where A (vertical, horizontal) = the projected areas (cm21obtained by planimeter, using an IBM PC computer with dedicated software supporting a graphics tablet (Trinity Computing Systems). L(vertica1 or horizontal) was the longest length in either view. With appropriate substitution: V = (8/3~)(Avertical)(Ahorizontal)/Lmin, where Lmin = shorter of the 2 long axes. Statistical methods: Calculated volumes for group I and II were compared to true volume for each group by standard linear regression analysis. To examine the degree of overestimation for each group, the ratios of LVtrue and calculated volume (LVtrue/calc) were measured and compared amongst casts, and a Student’s t test was done for comparisons between groups. A p <0.05 was considered significant. Data of regression analysis obtained from all 5 projections were compared between group I and II by analysis of covariants.

x = calculated volume; y =

Results The true volumes of LV and right ventricular casts ranged from 1.4 to 48.9 ml and 2.0 to 144 ml, respectively (Figure 1). In group I, mean values of the true volume of the LV and right ventricular were 10.7 f 9.1 and 24.8 f 32.9 ml; in group II, they were 21.3 f 12.8 and 16.8 f 10.7 ml, respectively. LV true volumes were significantly smaller in group I than group II (p
Discussion Baan et alzl have recently developed an on-line conductance catheter method to estimate cardiac chamber volumes in vivo; however, this technology has not received sufficient clinical validation. Other

TABLE II Mean Values and Statistical Analysis of Ratios of True and Calculated Left Ventricular (LV) Volume in Two Groups Mean Values of LV True/Calculated Ratio Projection

Group I

Group II

p Value

Anteroposterior and lateral Right and left anterior oblique Long axial oblique Hepatoclavicular Sitting up

0.622 0.636 0.669 0.698 0.677

0.842 0.785 0.820 0.820 0.812

<0.005
444

BIPLANE LEFT VENTRICULAR

VOLUMETRY

of biplane and single plane volumetry, it has been demonstrated that compression of LV in patients with an atria1 septal defect foreshortened the lateral axis resulting in a significantly greater single-plane overestimation of the LV volume compared with biplane estimation. Therefore, the use of an ellipsoid model becomes inaccurate in caseswith abnormal right ventricntal ular hemodynamics when the short minor axis is compressed by ventricular septum [Figure 4). Lange et all5 attributed the differences in regression slopes to differences in manufacture of the casts and to cardiac phase. In our study, the casts were produced by simultaneously filling silicon rubber into nom&heart both chambers to avoid shifting the ventricular septum abnormalRV hemodynamics during preparation. In their additional investigation, Lange et all6 found that systolic cross sections were not FIGURE 4. Transverse ventricular images obtalned from biplane as elliptical as in diastole, and prominent papillary clneanglography uslng perpendicular Image lntenslfler shows the muscle volumes were included within the systolic conIdentical sized silhouette In both groups. The use of an ellipsoid tours. Consequently, the volumes calculated in systole model becomes Inaccurate In hearts wlth abnormal right ventricular were larger than the true volumes, further accounting (RV) hemodynamlcs. for the reduced regression slopes. On the other hand, in patients with right ventricular enlargement the disparity in LV morphology is evident at end-diastole, but methodologies, such as echocardiographicz2 or radionot so at end-systole after ejection of a large right vennuclide angiocardiography,23 have had useful clinical tricular stroke volume.17 Small changes in LV shape application but are frequently obtained without simulduring the cardiac cycle in this setting, especially in taneous hemodynamic data; in addition, they are technically cumbersome and have large standard er- the lateral minor axis, may be caused by paradoxical or flat motion of ventricular septum compared with rors22-24in the volume estimate. Angiographic volume normal hearts. Our data indicate, therefore, that difestimation therefore remains an important clinical ferences between the 2 groups could be attributed methodology for assessment of ventricular performainly to the effect of abnormal right ventricular gemance. In this regard, postmortem casts of heart chamometry. The effect of right ventricular geometry must bers have been frequently used to determine true cartherefore be taken into account when calculating LV diac volumes and provide correlations to angiographic volume. estimations.3-12J5-17For LV volume determination the Effect of projections: Recently, Bargeon,lg Elliott20 area-length method for biplane cineangiography described by Dodge et a111J2assumed the LV cavity to be and their co-workers described the clinical imporan ellipsoid of revolution and has been widely accept- tance of axial cineangiography in outlining the anaed and applied. Previous investigations have demon- tomic detail in congenital heart disease. The long axial strated excellent correlations (r >0.90) between the oblique profiles the anterior ventricular septum, LV volumes of ventricular casts and those calculated from outflow tract and aortic valve-anterior mitral valve leaflet.31 The hepatoclavicular profiles the posterior biplane cinegrams using this approach.4-18.25*26 However, calculated volumes usually overestimate the true ventricular septum, atria1 septum and separates atriovolume, especially in small nonelliptic hearts.3J5J6 ventricular valves. The sitting-up position visualizes This reflects inability to estimate the volume occupied the bifurcation of pulmonary trunk. We could find few reports of LV volume determination in the long axial by the paillary muscles and trabeculae carnae within Morethe cavity, which is an applicable error especially at oblique and hepatoclavicular positions.18*25s26 over, LV volume in the sitting-up position has not been end-systole.15J6 Effect of right ventricular hemodynamics: In our evaluated. The regression equations obtained from our cast series are useful, allowing these improved study using casts derived from patients with congenital anatomic views to be used in performing quantitative heart disease, regression slopes in each projection were lower in group I compared with previously re- ventriculography. Of interest is the finding that the ported data. 3~18In our group II, the cross-sectional regression slope, correlation coefficient and standard shape of the LV cast was less elliptical and more pro- error of estimate obtained from long axial oblique prolate and the regression slopes were also lower. In ex- jection were the best of any projection in either group. perimental studies, ventricular size and shape were This may be so, as the major axis in long axial oblique projection reflects a more accurate long axis of the left shown to be dependent on the state of opposite ventricle.27In tetralogy of Fallot, transposition of great arte- ventricle because the outflow is not foreshortened as in other views. ries with intact ventricular septum, chronic pulmonary hypertension, atria1 septal defect or total anomalous Acknowledgment: We thank Cameron Finlay for pulmonary venous return, a shift of ventricular septum his assistance in figure production and data analysis, due to pressure or volume overload of the right ventriand Christine Viglatzis for the preparation of the mancle may compress the LV cavity.17*28-30 In a comparison uscript.

1

1l

February 15, 1988

References 1. Fisher EA. Eckner FAO, Dubrow IW, Hastreiter AR. Correction of angiographic and autopsy Jeff ventricular dimensions in children. Circulation 1978;58:739-746.

2. Graham TP, Lewis BW, larmakani MM, Canent RV, Capp MP. Left heart volume and moss quantification in children with left ventricular pressure overload. Circulation 1970;41:203-212. 3. Graham TP, Jarmakani MM, Canent RV, Morrow MN. Left heart volume estimation in infancy and childhood. Circulation 1971;43:895-904. 4. Arcilla RA, Tsai P, Thilenius OG, Ranninger K. Angiographic method for volume estimation of right and left ventricles. Chest 1971;60:446-454. 5. Sandler H, Alderman E. Determination of left ventricular size and shape. Circ Res 1974;34:1-8. 6. Thilenius OG, Arcilla RA. Angiographic right and left ventricular volume determination in normal infants and children. Pediotr Res 1974;8:67-74. 7. Wynne J. Green LH. Mann T. Levi” D, Grossman W. Estimation of left ventricular volumes in man from biplane cineangiograms filmed in oblique projections. Am f Cardiol 1978;41:726-732, 8. Cohn PF. Gorlin R, Adams DF, Cha hine RA, Vokonas PS, Herman MV. Comparison of biplane and single plane left ventriculograms in patients with coronary artery disease. Am \ Cordial 1974;33:1-6. 9. Rogers WI, Smith LR. Hood WP, Mantle JA, Rackley CE. Russell RO. Effect of filming projection and interobserver variability on ongiographic biplane left ventricular volume determination. Circulation 1979;59:96-104. 10. Froelicher V. Angiographic assessment of left ventricular volume, mass and wall motion. Eur Heart [ 1985;6(suppJC]:29-31. 11. Dodge HT, Sandler H, Ballew DW. Lord JD. The use of biplane angiocardiogrophy for the measurement of left ventricular volume in man. Am Heart 1 1960;60:762-776. 12. Dodge HT. Sandler H. Baxley WA, Hawley RR. Usefulness and limitations of radiographic method for determining left ventricular volume. Am J Cordiol 1966;18:10-24. 13. Kennedy JW. Trenholme SE, Kasser IS. Left ventricular volume and mass from single plane cineangiogrom. A comparison of anteroposterior and right anterior oblique methods. Am Heart J 1970;80:343-352. 14. Hillis LD, Winniford MD, Dehmer GJ, Firth BG. Left ventricular volumes by single-plane cineangiography. In viva validation of Kennedy regression equation. Am / Cardiol 1984;54:1159-1163. 15. Lange PE, Onnasch D, Farr FL, Malerczyk V, Heintzen PH. Analysis of left and right ventricular size and shape, as determined from human casts. Description of method and its validation. Eur r Cardiol 1978;431-448. 16. Lange PE, Onnasch D, Farr FL, Heintzen PH. Angiocardiographic left ventricular volume determination. Accuracy, as determined from human casts, and clinical application. Eur r Cardiol 1978;449-476.

THE AMERICAN

JOURNAL

OF CARDIOLOGY

Volume 61

445

17. Gault JH. Angiogrophic estimation of left ventricular volume. Cothet Cardiovasc Diagn 19752:7-N 16. Formanek A, Schey HM, Ekstrand KE, Velasques G, D’Souza VJ. Glass TA. Single versus biplane right and left ventricular volumetry: a cast and clinical study. Cathet Cardiovosc Diagn 1984;10:137-156, 19. Bargeron LM, Elliott LP, Soto B, Bream PR, Curry GC. Axial cineangiography in congenital heart disease. Section 1. Concept, technical and anatomic considerations. Circulation 1977;56:1075-1083. 20. Elliott LP. Bargeron LM, Bream PR, Soto B, Curry GC. Axial cineangiogrophy in congenital heart disease. Section 2. Specific lesions. Circulation 1977:56:1084-1093.

21. Baan J, Van der Velde ET, De Bruin HG, Smeenk GJ. Koops J, Van Dijk AD, Temmerman D, Senden 1, Buis B. Continuous measurement of left ventricular volume in animals and humans by conductance catheter. Circulation 1984; 70:812-823. 22. Silverman NH, Hudson S. Evaluation of right ventricular volume and ejection fraction in children by two-dimensional echocardiography. Pediatr Cardiol 1983;4:197-204. 23. Hoilund-Carlsen PF. Marving J. Rasmussen S. Haunso S, Pedersen JF. Accuracy of absolute left ventricular volumes and cardiac output determined by radionuclide cardiography. Int 1 Cordial 1984;6:505-521. 24. Levine RA. Gibson TC, Aretz T, Gillam LD, Guyer DE, King ME, Weyman AE. Echocardiographic measurement of right ventricular volume. Circulation 1984;69:497-504. 25. Als AV, Paulin S, Aroesty JM. Biplane angiographic volumetry using the right anterior oblique and half-axial left anterior oblique technique. Rodiology 1978;126:551-514. 26. Roger WJ, Smith LR, Beam PR, Elliott LP, Rackey C, Russell RO. Quantitative axial oblique contrast left ventriculography: validation of the method by demonstrating improved visualization of regional wall motion and mitral valve function with accurate volume determinations. Am Heart [ 1982; 103:185-194.

27. Urschel CW, Bemis CE, Borkenhagen ]SD, Sonnenblick EH. The influence of right ventricular filling pressure on left ventricular pressure and dimension. Circulation 1971;43/44(suppJ):125-145. 26. Levenson LW, Liedtke A], Hunter AS, Gentzer RD. Gault JH. Left ventricular chamber distortion in mitral stenosis (abstr). Circulation 1973;48:117. 29. Hunter HS, Gentzler RD, Gault ]H, Vincent WR. Distortion of left ventricular chamber anatomy by right ventricular volume overload (abstr). Am r Cardiol 1974;33:145. 30. Fraker TD, Wise NK. Harrel FE, Behar VS. Comparison of biplane and single plane left ventricular volumes in atrial septal defect. Cathet Cardiovasc Diagn 1980;6:39-48. 31. Kelley MJ. Higgins CB, Kirkpatrick SE. Axial left ventriculography in discrete subaortic stenosis. Radiology 1980;135:77-80.