First-Pass Anger Camera Radiocardiography:Biventricular Ejection Fraction, Flow,and Volume MeasurementsMartin L. Nusynowitz, Anthony R. Benedetto, Richard A. Walsh,and Mark R. Starling*

Division of Nuclear Medicine, Department of Radiology, University of Texas Medical Branch,Galveston, and Divisions of Nuclear Medicine and Cardiology, University of Texas HealthScience Center, San Antonio, Texas

First-pass (FP) right and left ventricular ejection fraction results were compared with

equilibrium radiocardiographic (ER) measurements, and FP left ventricular ejection fraction(LVEF) values were compared with biplane contrast angiographie (CA) measurements in 13patients with and seven patients without régurgitantvalvular disease. Régurgitantfractionswere calculated from differences between the FP right and left ventricular stroke volumes.Ejection fractions determined by FP were precise (mean CV = 9.6% RVEF, 13.4% LVEF).Mean LVEF by FP and ER were essentially identical, and both were lower than by CA.LVEF(FP)correlated with LVEF by ER and CA (r = 0.88, p < 0.001). Mean RVEF by both FPand ER were also correlated (r = 0.82, p < 0.001 ). There was correlation between FP(corrected) and CA left ventricular stroke (r = 0.77), end-diastolic (r = 0.88), and end-systolic(r = 0.91) volumes, but underestimates were noted when uncorrected flows were used (r =0.52-0.71). The FP régurgitantfraction measurements separated the patients with régurgitantvalvular disease from those without and agreed well with CA grading of régurgitation.

J NucÃ-Med 28:950-959,1987

/eft ventricular ejection fraction measurements byfirst-pass (FP) radiocardiographic techniques correlatewell with contrast angiography (CA) (1,2), but becauseof the complexity of the shape of the right ventricle,geometric methods of determining right ventricularejection fraction (RVEF) are difficult (3,4). A numberof radiocardiographic RVEF methods have been described and appear to be satisfactory, as judged byinternal consistency of repeated measurements and correlation of FP with equilibrium radiocardiographic(ER) methods (5,6). Because of problems arising fromoverlying cardiac chambers, first-pass RVEF methodshave utilized right anterior oblique (RAO) or anteriorprojections in order to spatially separate atrial activityfrom that of the right ventricle; these FP techniques

Received Dec. 20, 1985; revision accepted Dec. 23, 1986.For reprints contact: Martin L. Nusynowitz, MD, Div. of

Nuclear Medicine, Dept. of Radiology. Univ. of Texas MedicalBranch, Galveston, TX 77550.

" Present address: University of Michigan Medical Center, De

partment of Internal Medicine, Div. of Cardiology, VA MedicalCenter, 2215 Fuller Rd., Ann Arbor, MI 48105.

also allow for temporal separation of the left and rightventricles. Equilibrium radiocardiographic methods ( 7)performed in the left anterior oblique (LAO) projectionhave used special collimation (30°slant hole) in an

attempt to separate the right atrium from the rightventricle.

First-pass RAO projection techniques, from a theoretical viewpoint at least, allow determination of RVEF;however, none of the techniques, including equilibriummethods, has been subjected to thorough validation byindependent means because of the difficulty of defininga widely acceptable "gold standard" for comparison.

Furthermore, the RAO projection FP technique doesnot permit simultaneous calculation of the cardiac output and stroke volume from either ventricle becausethe spatial overlap of the chambers in the equilibriumphase precludes the determination of equilibrium concentrations for either ventricle, which are necessary forthe Stewart-Hamilton calculational method.

First-pass radionuclide methods for the determination of cardiac output (and, hence, stroke volume) havebeen used successfully with probe and multicrystal detectors (8,9). These studies have shown the validity of

950 Nusynowitz, Benedetto, Walsh et al The Journal of Nuclear Medicine

FP measurements as compared with CA results, butwidespread application has not ensued. Furthermore,the use of these methods with Anger cameras has beenextremely limited.

Because we were interested in the determination ofejection fraction, flow, and volume information fromboth ventricles, we performed ejection fraction, flow,and volume measurements in a modified LAO projection, which allowed evaluation of both ventricles simultaneously. The purposes of our investigation wereto ascertain if valid measurements of left and rightventricular ejection fractions, flows, and volumes couldbe made using noninvasive FP radiocardiographicmethods and Anger scintillation cameras and to determine if the right and left ventricular flow data could beused to estimate the degree of régurgitantvalvular disease.

MATERIALS AND METHODS

The study group consisted of 20 patients in whom cardiaccatheterization was performed for evaluation of either coronary artery disease, valvular heart disease, or congestive heartfailure. The radiocardiographic studies were performed within48 hr of the catheterization, usually on the day preceding thecatheterization procedure.

Right heart catheterization was performed using a balloon-tipped flow-directed catheter, and brachial artery catheterization was performed using an NAMIC* catheter. Cardiac out

put determinations were obtained using both the Pick principle and indocyanine green dye indicator dilution method. Alldeterminations were performed in triplicate.

Left heart catheterization was performed using standardbrachial or femoral techniques. Simultaneous biplane leftventricular cineangiograms were obtained in the 30°RAO and60°LAO/20" caudal projections following injection of 40-60

ml Renografm-76 at 10-15 ml/sec and 500 psi at a framingrate of 60/sec. Left ventricular end-diastolic and end-systolicvolumes were calculated using a modified biplane formula(10). Left ventricular ejection fraction was calculated in thestandard fashion.

The blood volume (Vb) was determined by using 5 /¿Ciiodine-125 (I25I)radioiodinated human serum albumin,* be

cause the rate of egress of this material from the circulation ismuch lower than that of technetium-99m ("Te) albumin

albumin preparations (//). Corrections for overestimation ofthe red cell volume were employed using techniques previously described (72).

Following the determination of Vb, the FP study was undertaken. A rapid bolus injection (without flushing) of 8-10mCi of ["Tcjalbumin* in a volume of

defined on (he spatially smoothed and background-subtractedimages. The FP results were compared with the ER images.Left ventricular FP results were compared with results of CA.

For determination of flow by FP radiocardiography, theStewart-Hamilton principle was employed. The cardiac output, F, is given by:

mlmin A,

(1)

where Cf (cpm/ml) is the equilibrium concentration of thetracer in the blood, Vh(ml) is the blood volume, and A, (min-cpm/ml) is the area under the extrapolated FP curve. Strokevolume (SV) was obtained from F and heart rate (HR):

SV = F/HR. (2)

Using the FP ejection fraction, end-diastolic volume (EDV)and end-systolic volume (ESV) were calculated by the equations:

EDV = SV/EFESV = EDV - SV.

(3)(4)

The stroke volumes, as obtained from the right ventricularregion of interest [RVSV,RV|]and from the lung region ofinterest [RVSV(LU)],were used without correction. The strokevolume obtained from the LV was evaluated both in anuncorrected manner and with the following correction forbolus smearing and régurgitation.

Sequential FP determination of the RV, lung, and uncorrected LV flows (stroke volumes) in 13 patients with left-sidedrégurgitantvalvular disease (R+) and seven patients withoutrégurgitation(R-) showed progressive diminution in both

groups (Table 1). Because there is no significant differencebetween the ratios of the R+ and R—group on the side wherethere was no régurgitation,it can be assumed that both bolussmearing and régurgitationdistort the radiocardiograms similarly. That is to say, disruption of the tracer bolus due to therégurgitantprocess has the same qualitative effect as smearinghas on the tracer bolus as it proceeds downstream, as evidenced by declining flows, even in the nonregurgitant group(R-).

Accordingly, identical corrections may be applied to bothbolus smearing and régurgitationin order to calculate the totalcorrected LV flow. The uncorrected left ventricular flow results from the algebraic summation of the forward left ventricular flow (F) and the back flow (régurgitantflow and/orbolus smearing effect, B; Fig. l ):

uncorrected LV flow = F - B. (5)

Solving for B.

B = F - (uncorrected LV flow). (6)

TABLE 1Flow Ratios of Right Ventricle to Lung and Lung to LeftVentricle in Patients With (R+) and Without (R-) Left-

Sided RégurgitantValvular Disease

RV/lung flowLung/LV (uncorrected)

flowR-1

.29 ±0.081.22 + 0.09R+

p1.24

±0.07 N.S.1.88 ±0.20

ROI previously described for RV, lungs, and LV. The heartrate was determined by electrocardiographic monitoring during the FP study and was used in the determination of strokevolume.

Unless stated otherwise, all values are given as x ±1 s.e.Standard parametric statistics were employed and productmoment correlation coefficients were calculated (14).

RESULTS

The patients final diagnoses and results of the ejectionfraction and coefficient of variation determinations arepresented in Table 2. Coefficient of variation, a standard statistical measure of precision, reflects the degreeof reproducibility of ejection fraction from beat to beat.

The mean coefficient of variation for the first-pass

left ventricular ejection fraction [LVEF(FP)] was 13.4 ±1.4%. The mean LVEF(n»,was 42.9 ±3.6%, whereas,the mean LVEF(KR) was 43.9 ±4.0%; there was nosignificant difference between methods (paired t-test).

The mean contrast angiographie left ventricular ejection fraction [LVEF(rA>] was 49.0 ±4.2%; this wassignificantly higher than the results of FP or ER (p <0.01). The mean LVEF(R»)in the subgroup of patientswith régurgitantvalvular disease (R+) was 36 ±4%,whereas, in those without régurgitation(R-) the mean

LVEF(FP)was 55 ±4% (p < 0.01).The mean coefficient of variation for the first-pass

right ventricular ejection fraction [RVEF(Fp)] was 9.6 ±

80

80

ÕÕ 40

UJ

20

20 40 80 80

LVEFrFP1(%)FIGURE 2Correlation of left ventricular ejection fractions from equilibrium and first-pass radiocardiography. Y = 0.99X -I-1.4;r = 0.89; p < 0.001.

1.0%. The mean RVEF(FP) was 41.4 ±3.4% and themean RVEF,ER) was 42.8 ±3.3%, there was no significant difference between the paired results. However,when these patients were divided into those withoutrégurgitantvalvular disease (R-) and those with régur

gitant valvular disease (R+), the mean RVEF,FP) for R+

TABLE 2Final Diagnosis, Degree of Régurgitation,Ejection Fractions and their Coefficients of Variation

Patientno.1234567891011121314151617181920Age/sex56F60F58F50F67F62F58

M31M45F61

M38M53M49F25

M60F17

M34M82

M67M66FDiagnosis'NLCAD,

MRNLNLCAD,

MRAS,MS. AR.MRCADCM,

MRCAD,MRCAD,MRCADARMRRepaired

Tetralogy,MRMRInfundibular

StenosisMSAR,

MRMR,CADCM,

MRDegree

of RVEF^p,régurgitation(%)0+10044410414141042434141004343416135606349704824195541372538245433383122X

=41.4s.e.= 3.4CV111081451012710777193686621149.61.0LVEF.fp,693469614470472634324244292135524537541342.93.6CV151816174—10141291218201594101383113.41.4RVEF™62335552455663143352335418432557394039—42.83.3LV(%r603362623570581334304146342231625660591043.94.0LVEF^,743880644657631438464850302640755957591549.04.2

'NL, normal, no disease found; CAD, coronary artery disease; MR, mitral régurgitation; AR, aortic régurgitation; MS, mitral stenosis; AS, aortic

stenosis; CM, cardiomyopathy.

Volume 28 •Number 6 •June 1987 953

80

60

ïï 40

LU

OC 20

600

100

20 40 60 80

FIGURE 3Correlation of right ventricular ejection fractions from equilibrium and first-pass radiocardiography. Y = 0.78X + 9.9;r = 0.82; p< 0.001.

was 36 ±4%, significantly lower than that for R-, 51

±4% (p < 0.05).The validity of the FP method employed for deter

mination of LVEF is confirmed by the correlations withthe ER method (r = 0.89, p < 0.001) and with the CAmethod (r = 0.88, p < 0.001). The regression of

LVEF(ER>on LVEF(FP) is essentially the line of identity(Fig. 2). Similarly, the correlation of RVEF(ER) andRVEF(FPI was significant (r = 0.82, p < 0.001; Fig. 3).

100 200 300 400 500 600

LVEDV|FpC|(ml)FIGURE 5Correlation of left ventricular end diastolic volumes bycontrast angiocardiography and the corrected first-passmethod. Y = 0.84X + 3; r = 0.88; p + 0.001.

The corrected FP left ventricular stroke volumes andthe CA stroke volumes correlated (Fig. 4), with nosignificant difference between paired data [LVSV(FPC)=96 + 5 ml, LVSV(CA) = 92 ±7 ml]. The uncorrected

values, however, correlated poorly; the LVSV(FPU)was52 ±6 ml, significantly lower than for CA (p < 0.001).

Correlation was noted between the corrected FP andCA left ventricular end-diastolic volumes (Fig. 5) with

a significant difference (p < 0.01) between the paired

Ü

>CO

200

160

1201

80

40

00 40 80 120 160 200

LVSV[FPC](ml)

<O

COUJ

500

400

300

200 1

100

0 ~\ 1 1 1 1 r

100 200 300 400 500

LVESV[FPCl(ml)FIGURE 4 FIGURE 6Correlation of left ventricular stroke volumes by contrast Correlation of left ventricular end systolic volumes byangiocardiography and the corrected first-pass method. Y contrast angiocardiography and the corrected first-pass= 1.05X - 8; r = 0.77; p < 0.001. method. Y = 0.86X - 12; r = 0.91; p < 0.001.

954 Nusynowitz, Benedetto, Walsh et al The Journal of Nuclear Medicine

TABLE 3Comparison of Right Ventricular Flows as Obtained from First-Pass Right Ventricular and Lung Regions of Interest

with Results of Green Dye and Fick Measurements

Linear regressionequationRight

Ventricular ROIRVSV(GO)= 0.48[RVSV(RV)] +RVSV(FK;K)= 0.64[RVSV(RV)] -

Lung ROI:RVSV(QO)= 0.82[RVSV(LU)] +RVSV(F1CK)=1.19[RVSV(LU)]-18.1

H2.91.3

-27.6s.e.e.

(ml)12.3

20.011.3

14.5Con-elation

coefficient(r)0.81

0.760.85

0.88P

o

150

100

50

50 100 150

RVSV[LU](ml)FIGURE 9Correlation of right ventricular stroke volumes performedby the Pick and first-pass radiocardiographic methods. Y= 1.19X - 28; r = 0.88; p < 0.001.

From the RVEF and the RVSV,LU>,the right ventricular end-systolic and end-diastolic volumes were calculated for the 13 patients with régurgitantleft-sidedvalvular disease (R+) and for the seven patients withoutvalvular régurgitation(R-). Table 4 presents these data.

The mean right ventricular stroke volumes of the R+group were not significantly different from those in theR- group. The group with régurgitantvalvular diseasehad higher right ventricular end-diastolic and end-systolic volumes, but they did not differ significantly.

Finally. Table 5 compares the R-l- with R- groups.Figure 10 compares the grading of the extent of régurgitation at the time of cardiac catherization and thecalculation of régurgitantfraction from the FP method.Values of RF >18% were strongly indicative of régurgitation.

TABLE 4Comparison of Right Ventricular Parameters of 13

Patients with Left-Sided RégurgitantValvular Disease(R+) and in Seven Patients Without (R-)

RVEFR+

R-RVSVR+R-RVEDVR+

R-RVESVR+

R-Mean36%

5168ml

85202ml

173134ml

88s.e.4

46716

2116

17t2.411.821.101.86P

20

10

-10

Degree of régurgitationFIGURE 10Scattergram of left ventricular régurgitantfractions usingcorrection methods described in the text versus the semi-quantitative degree of régurgitationestimated visually during contrast angiocardiography.

956 Nusynowitz, Benedetto, Walsh et al The Journal of Nuclear Medicine

technique resulted in overestimation of the ejectionfraction.

Steele et al. (76) introduced the FP radiocardi-ographic technique for the determination of RVEFusing an RAO projection. The method appears satisfactory, because the right ventricle can be separated fromthe surrounding chambers both temporally and spatially; it has had wide application (13,15,17-19). Indeed, the first-pass RVEF in the RAO projection currently is the accepted method of determining RVEF,although there is a paucity of validation studies withother methods. The precision of this technique has beenwell documented; accuracy of the method is only inferential (13,15). Other FP methods using krypton-Simand xenon-133 have been reported recently (20-22).

A desirable goal in the performance of the noninva-sive radiocardiographic study is the opportunity to notonly obtain biventricular ejection fractions but also tomeasure cardiac output and stroke volume for eachventricle. This would enable calculation of end-systolicand end-diastolic volumes for each ventricle as well aspermit (theoretically, at least) the quantitation of shuntsand valvular régurgitation.The disadvantage of employing the RAO projection is its inability to quantitatestroke volumes of each chamber, because the right andleft ventricles superimpose, precluding the measurement of equilibrium concentrations of tracer in theventricles.

This report demonstrates that the precision of determining RVEF from the LAO projection is as good asthat of the left ventricle as judged by similar meancoefficients of variation. The accuracy of this method,indeed, with all other RVEF methods, is difficult tojudge. However, the good correlation of RVEF(FP)withRVEF(ER)and the insignificant mean difference betweenpaired determinations is strong supportive evidence forthe accuracy of the RVEF(FP),demonstrating internalconsistency between methods; the RVEF(ER)has shownhigh correlation with RVEF(Fp>from the RAO projection (18,19).

Further supporting evidence for the validity of theRVEF(Fp)can be seen by comparing the results in patients with (R+) and without (R-) régurgitantvalvulardisease. Those with régurgitationwould be expected tohave lower RVEF values because of higher pulmonaryartery pressures (19,23); supportive evidence for thisfinding was also observed in the present study.

First-pass radionuclide methods for quantitative determination of ventricular flow, volumes, and régurgitant fraction have had limited application, despite thefact that the theoretical basis of such studies has beenknown for years. This is attributable to a variety ofreasons, predominantly arising from limitations ofequipment (for example, deadtime losses and, thus,count rate limitations of early Anger cameras). Hence,FP methodology was limited to multicrystal cameras

not widely available and with spatial resolution inferiorto Anger cameras. Modern Anger cameras and computer systems are now widely available and possess highcount rate capability, thus, permitting the performanceof FP studies with only minimal deadtime losses whenusing significant amounts of radiopharmaceutical.Nevertheless, the technique of performing such studiesis very important, and meticulous attention to detail isnecessary.

1. A tight bolus injection is essential. This is readilyperformed with the patient in a supine position using ajugular vein injection; flushing of the syringe is unnecessary, and the procedure is less painful than antecubitalinjections. Antecubital injections are adequate, but onlyif care is taken to use the basilic vein and the dose israpidly flushed. With either method, valsalva maneuvers by the patient must be avoided.

2. Radioiodinated serum albumin should be used forblood volume determination because the disappearanceof [Tc]albumin from the circulation is rapid (77);whether or not such disappearance affects the bloodvolume (Vb) and equilibrium concentration (Cf) portions of the Stewart-Hamilton equation equally in opposite directions, so that the effects of disappearance of[Tc]albumin from the blood cancel, is unknown andremains to be tested. If it does turn out that the lossfrom the circulation results in an expanded Vb valuebalanced by a decrease in Cf, then the technique maybe simplified by elimination of the radioiodinatedserum albumin determination of Vb. With the presenttechnique, [Tc]albumin losses from the circulationmust be corrected. Furthermore, corrections must bemade for overestimation of the Vbdue to the differencein total body and peripheral venous hematocrit (72).

3. In estimating the area under the FP curve we haveused a monoexponential extrapolation of the initialdownslope. Alternative methods, such as gamma variate techniques, may also be used, but the adequacy ofother methods was not tested in this study.

4. Automation of calculation reduces error and allows answers to be obtained quickly; we have developedprograms for this purpose.

As can be seen from the data in Tables 1 and 2 thereis a step-down in stroke volume as one proceeds fromRV to lung to LV (uncorrected data). Possible reasonsfor this step-down include deadtime losses, which decrease as the bolus dissipates, inadequate mixing, orbolus smearing, which has the effect of increasing thearea under the FP curve, thereby, resulting in a calculated lower flow as one proceeds downstream. Whencorrelations of RVSV from analysis of the right ventricular region of interest or the lung region of interest weremade with green dye and Pick stroke volumes, betterresults were obtained with the lung curve (Table 3).This is probably due to both better mixing of theradiopharmaceutical with the blood as it traverses the

Volume 28 •Number 6 •June 1987 957

lung and elimination of right atrial activity (24,25). Ofcourse, the calculation of RVEF, of necessity, mustcome from the right ventricular region of interest curve.Accordingly, we have adopted the pulmonary flow asrepresentative of RVSV and calculated RV volumesand régurgitantfractions from these data (Table 2).

The degree of correlation between the RVSV asderived from the first pass lung curve [RVSV(LU)]andwith the Pick and green dye methods is surprisinglyhigh, considering the studies were performed on different days, in different environments, with different degrees of invasiveness, and under different conditions ofsedation. The correlation between Pick and green dyestroke volumes performed almost simultaneously at thetime of cardiac catheterization was similar (r = 0.87).Thus, the validity of the RVSV from both the rightventricular region of interest and the lung region ofinterest is established, but the latter is preferred.

At the time we performed this study, there was not agood angiographie method available to test the validityof the derived right ventricular end-systolic and end-diastolic volumes. However, if the ejection fraction isvalid and the RVSV is valid, then the derived volumesshould also be valid. The values for RVEDV andRVESV obtained are certainly reasonable and believable, lending credence to their validity; furthermore,similar techniques applied to the left ventricle result inLVEDV and LVESV that correlate well with contrastangiographie data.

The flow data derived directly from the left ventricular region of interest (uncorrected) resulted in unrealistic volumes for left ventricular stroke, end-diastolic,and end-systolic volumes. Furthermore, the uncorrected LV data correlated poorly with the CA results.In contrast, the results of the left ventricular flow andvolumes, when corrected, correlate highly with the CAdata. It should be remembered, however, that the CAtechnique is subject to a number of limitations: (a) theiodinated contrast material employed, unlike radiola-beled tracers, produces inotropic and chronotropic effects that alter the parameters being studied; (b) alimited number of images are obtained and a "representative beat" is selected to determine end-systolic and

end-diastolic volumes; (c) the placement of the cardiacedge for determination of the end-systolic and end-diastolic volumes is subject to observer interpretation;(d) with cardiomegaly, the assumption that the heartremains an ellipsoid of revolution becomes less realistic.Notwithstanding these limitations, the CA method hasremained the standard method against which newertechniques are compared; it is gratifying, therefore, thatthe FP corrected left ventricular results described hereincorrelate so highly. Further support of the validity ofthe correction of LVSV is derived from the good agreement with the catheterization methods of estimatingrégurgitantfraction.

In conclusion, FP determination of the LVEF in theLAO projection gives precise and accurate measurements. The precision of the RVEF is similar. Althoughthere is no good way to corroborate the accuracy of theRVEF measurement due to the lack of a widely accepted "gold standard," the values obtained seem to be

in keeping with the expectations of this patient group.Furthermore, the significant correlation of the RVEF(FP)and the RVEF,ER)lends substance to the accuracy ofthis method. The corrected left ventricular stroke volume (LVSV(FPC))correlates highly with LVSV(CA)andthe results obtained are similar. Similar high correlations are obtained for LVEDV(FPC)and LVESV(FPc)andCA techniques. Right ventricular stroke volume[RVSV(LU)]results are similar to and correlate highlywith both green dye and Pick methods. Right ventricular end-diastolic and end-systolic volumes appear tobe reasonable. Finally, the régurgitantfractions derivedfrom the methods employed distinguish the R+ andR- groups of patients and agree well with results of

catheterization. Therefore, the LAO projection technique, as described in this investigation, is a validmethod for obtaining simultaneous right and left ventricular ejection fractions, flows, volumes, and régurgitant fractions.

NOTES*North American Instrument Corp., Hudson Falls, NY.*Mallinckrodt, Inc., St. Louis, MO.*Medi-Physics, Inc., Richmond, CA.§Medical Data Systems, Ann Arbor, MI.

ACKNOWLEDGMENTS

This work was supported by New Investigator ResearchAward R23 HL-27508, NIH (M.R.S.).

REFERENCES

1. Marshall RC, BergerHJ, Costin JC, et al. Assessmentof cardiac performance with quantitative radionuclideangiography: sequential left ventricular ejection fraction, normalized left ventricular ejection rate, andregional wall motion. Circulation 1977;56:820-829.

2. Folland ED, Hamilton GW, Larson SM, et al. Theradionuclide ejection fraction: a comparison of threeradionuclide techniques with contrast angiography. JNucÃ-Med 1977; 18:1159-1166.

3. Horn V, Mullins CB, Safter SI, et al. A comparison ofmathematical models for estimating right ventricularvolumes in animals and man. Clin Cardio! 1979;2:341-347.

4. Dell'Italia LJ, Starling MR, Walsh RA, et al. Validation of attenuation-corrected equilibrium radionuclideangiographie determinations of right ventricular volume: comparison with cast-validated biplane cine-ventriculography. Circulation 1985;72:317-326.

5. Maddahi J, Berman DS, Matsuoka DT, et al. A new

958 Nusynowitz, Benedetto, Walsh et al The Journal of Nuclear Medicine

technique for assessing right ventricular ejection fraction using rapid multiple-gated equilibrium cardiacblood pool scintigraphy. Circulation 1979; 60:581-589.6. Starling MR, Dell'Italia LJ, Walsh RA, et al. Accurate

estimates of absolute left ventricular volumes fromequilibrium radionuclide angiographie count data using a simple geometric attenuation correction. JACC1984; 3:789-798.

7. Holman BL, Wynne J, Zielonka JS, et al. A simplifiedtechnique for measuring right ventricular ejection fraction using the equilibrium radionuclide angiocardiogram and the slant-hole collimator. Radiology 1981;138:429-435.

8. Berger HJ, Matthay RA, Pytlik LM, et al. First-passradionuclide assessment of right and left ventricularperformance in patients with cardiac and pulmonarydisease. Semin NucÃ-Med 1979; 9:275-295.

9. Steele P, Kirch D, Vogel R. Evaluation of centralcirculatory dynamics with the radionuclide angiocardiogram. In: Strauss HW, Pitt B, eds. Cardiovascularnuclear medicine. St Louis: CV Mosby Co., 1979:105-115.

10. Wynne J, Green LH, Mann T, et al. Estimation of leftventricular volumes in man from biplane cineangio-grams filmed in oblique projections. Am J Cardici1978; 41:726-732.

11. Nusynowitz ML, Straw JD, Benedetto AR, et al. Bloodclearance rates of technetium-99m albumin preparations [Concise communication]. J NucÃ-Med 1978;19:1142-1145.

12. Nusynowitz ML, Blumhardt R. Estimation of true redcell volume from RISA red cell volume. Am J Roent-genol 1974; 120:549-552.

13. Tobinick E, Scheiben HR, Henning H, et al. Rightventricular ejection fraction in patients with acuteanterior and inferior myocardial infarction assessedby radionuclide angiography. Circulation 1978;57:1078-1084.

14. SnedecorGW, Cochran WG. Statistical methods. 7thed. Ames, IO: Iowa State University Press, 1980.

15. Berger HJ, Matthay RA, Loke J, et al. Assessment ofcardiac performance with quantitative radionuclide

angiocardiography: right ventricular ejection fractionwith reference to findings in chronic obstructive pulmonary disease. Am J Cardio! 1978; 41:897-905.

16. Steele P, Kirch D, LeFree M, et al. Measurement ofright and left ventricular ejection fractions by radionuclide angiocardiography in coronary artery disease.Chest 1976; 70:51-56.

17. Brent BN, Berger HJ, Matthay RA, et al. Physiologiccorrelates of right ventricular ejection fraction inchronic obstructive pulmonary disease: a combinedradionuclide and hemodynamic study. Am J Cardiol1982; 50:255-261.

18. Starling MR, Dell'Italia LJ, Nusynowitz ML, et al.Estimates of left-ventricular volumes by equilibriumradionuclide angiography: importance of attenuationcorrection. J NucÃ-Med 1984; 25:14-20.

19. Reduto LA, Berger HJ, Cohen LS, et al. Sequentialradionuclide assessment of left and right ventricularperformance after acute transmural myocardial infarction. Ann Intern Med 1978; 89:441-447.

20. Ham HR, Franken PR, Georges B, et al. Evaluationof the accuracy of steady-state krypton-81 m methodfor calculating right ventricular ejection fraction. JNucÃ-Med 1986; 27:593-601.

21. Caplin JL, Flatman WD, Dymond DS. Gated rightventricular studies using krypton-Sim: comparisonwith first-pass studies using gold-195m. J NucÃ-Med1986; 27:602-608.

22. Martin W, Tweddel AC, McGhie I, Hutton I. Gatedxenon scans for right ventricular function. J NucÃ-Med1986; 27:609-615.

23. Korr KS, Gandsman EJ, Winkler ML, et al. Hemodynamic correlates of right ventricular ejection fraction measured with gated radionuclide angiography.Am J Cardiol 1982; 49:71-77.

24. Swan HJC, Beck W. Ventricular nonmixing as asource of error in the estimation of ventricular volumeby the indicator-dilution technic. Circ Res 1960;8:989-998.

25. Maseri A, Enson Y. Mixing in the right ventricle andpulmonary artery in man: evaluation of ventricularvolume measurements from indicator washout curves.J Clin Invest 1968; 47:848-859.

Volume 28 •Number 6 •June 1987 959