validation of spect equilibrium radionuclide angiographic right ventricular parameters by cardiac...

ORIGINAL ARTICLES

Validation of SPECT equilibrium radionuclideangiographic right ventricular parameters by cardiacmagnetic resonance imaging

Kenneth Nichols, PhD,a Rola Saouaf, MD,b Ala’eldin A. Ababneh, MD,a Robyn J.Barst, MD,a Marlon S. Rosenbaum, MD,a Mark W. Groch, PhD,c Abu H. Shoyeb,MD,a and Steven R. Bergmann, MD, PhDa

Background. Recent advances in the treatment of primary pulmonary hypertension (PPH),and in surgery to correct tetralogy of Fallot (TOF), have rekindled interest in evaluating rightventricular (RV) volume and ejection fraction (EF). The purpose of this investigation was todetermine the accuracy of RV functional parameters assessed by single photon emissioncomputed tomography (SPECT) equilibrium radionuclide angiography (ERNA).

Methods and Results. Twenty-eight patients with PPH (n � 15) or TOF (n � 13) (aged 28� 14 years; 57% male) were analyzed by means of SPECT ERNA algorithms that automaticallyidentified mid-RV tomographic planes, generated regions isolating the right ventricle from otherstructures, and presented RV-segmented regions as a cinematic display. RV EF and volumeswere computed and compared with values obtained by magnetic resonance imaging (MRI).Mean values were not different between SPECT ERNA and MRI for RV EF, end-diastolicvolume, and end-systolic volume (42% � 11% vs 41% � 10%, 135 � 67 mL vs 139 � 91 mL,and 87 � 54 mL vs 85 � 61 mL, respectively; P � not significant for all comparisons).Significant linear correlation (P < .0001) was found between SPECT ERNA and MRI for RVEF, end-diastolic volume, and end-systolic volume (r � 0.85, r � 0.94, and r � 0.93, respectively). Nostatistically significant trends or biases for RV EF were found. Intraobserver and interobservercomparisons demonstrated good reproducibility. As expected, RV volume was significantlyhigher and RV EF was significantly lower for patients with PPH and TOF than were values forindividuals at low likelihood for coronary artery disease or other cardiac disease.

Conclusions. SPECT ERNA provides accurate, reproducible assessment of RV volumes andEF and should prove useful in evaluating the magnitude of RV dysfunction in patients and inproviding an objective means with which to assess the results of therapeutic interventions. (JNucl Cardiol 2002;9:153-60.)

Key Words: Right ventricle • tomography • radioisotopes • cardiac volume • magneticresonance imaging • tetralogy of Fallot • primary pulmonary hypertension

See related editorial, p 226

Advances in the treatment of primary pulmonaryhypertension (PPH),1 and in surgery to correct tetralogyof Fallot (TOF),2 have rekindled interest in accuratenoninvasive measurement of right ventricular (RV) vol-umes and ejection fraction (EF). In patients with PPH orTOF, the right ventricle, right atrium (RA), and RVoutflow tract may be markedly abnormal, and accurateassessment of RV function is needed to evaluate diseaseprogression and to assess the efficacy of therapeuticinterventions. However, it can be difficult to be certainthat an injected tracer is well mixed with blood during

From the Division of Cardiology,a Department of Medicine, andDepartment of Radiology,b Columbia University College of Physi-cians and Surgeons, New York, NY, and Division of NuclearMedicine,c Department of Radiology, Northwestern UniversityHospital, Chicago, Ill.

Presented in part at the 50th Annual Scientific Meeting of the AmericanCollege of Cardiology, Orlando, Fla, March 17-21, 2001.

Supported by a grant from Siemens Medical Systems, Inc.Received for publication June 20, 2001; final revision accepted Aug 14,

2001.Reprint requests: Kenneth Nichols, PhD, Division of Cardiology,

Columbia University, PH9-993B, 622 W 168th St, New York, NY10032; [email protected].

Copyright © 2002 by the American Society of Nuclear Cardiology.1071-3581/2002/$35.00 � 0 43/1/119464doi:10.1067/mnc.2002.119464

153

the RV phase of bolus transit for gated first-pass RV EFassessment,3,4 and conventional planar equilibrium ra-dionuclide angiography (ERNA) RV EF values can beunreliable in patients in whom the RA or RV outflowtract (or both) is substantially enlarged.5,6 Neither first-pass nor planar ERNA provides 3-dimensional wallmotion or direct measurement of RV volume.

Cardiac magnetic resonance imaging (MRI) is ex-cellent for overall RV evaluation.7,8 Spin-echo acquisi-tions enable discrimination of epicardial and endocardialboundaries for calculation of RV volume despite theeccentric shape of the RV chamber.9–11 However, car-diac MRI may not always be readily available, and atpresent analyses are time-consuming.

SPECT ERNA should provide unambiguous separa-tion of the cardiac chambers12–14 because, like MRI, it isa fully 3-dimensional means of acquiring data synchro-nously with the heartbeat. Although SPECT ERNA algo-rithms are commercially available,15 few reports haveaddressed the right ventricle.15,16 SPECT ERNA producesreproducible results through automation, but even the mostrecent reports have focused on left ventricular (LV) meas-urements,17 for which there are several well-establishedalternatives for reliably computing LV parameters, whereasRV assessment is generally more problematic. Thereforethe objective of this study was to determine the accuracyof SPECT ERNA measurements of the right ventricle,with MRI used as the standard for comparison.

METHODSPatients

Twenty-eight consecutive patients (aged 28 � 14 years[range, 7–56 years]; 57% male) who required clinically indicatedtechnetium 99m—labeled red blood cell studies for evaluation forPPH (n � 15) or TOF (n � 13) were studied prospectivelybetween July 1, 1999, and November 10, 2000. All patients hadevidence of pulmonary insufficiency. All patients underwentcardiac MRI for calculation of RV volume and RV EF. Allcorrelative SPECT ERNA and MRI studies were performed within 1month of each other (mean interval, 10 � 10 days). No patient hadany significant cardiac event between studies, and none had changesin medical or surgical therapy. No patients were excluded fromthis study on the basis of inadequate SPECT or MRI image quality.

A second group composed of 15 consecutive femalepatients (aged 53 � 16 years) with breast cancer referred forLV EF quantitation prior to chemotherapy was also studied.These patients had a low likelihood of coronary artery diseaseand no history of cardiac disease. SPECT ERNA data for thesepatients were acquired to form an initial estimate of the normallimits to be expected from the application of the SPECT ERNAmethods, as described below.

Planar ERNA

For all patients, immediately prior to SPECT ERNA dataacquisition, conventional planar ERNA was performed in the

left anterior oblique projection that optimized septal separationof the right ventricle from the left ventricle. For adults, injectedTc-99m—pertechnetate activity was 925 MBq (25 mCi) fol-lowing injection of 5 mg of pyrophosphate. For patients under18 years of age whose body weight was less than 70 kg, theseinjections were scaled downward linearly for body weight. A20% energy window centered on 140 keV was used for dataacquisition, with low-energy, general-purpose collimation. Datawere acquired as 64 � 64 matrices, gated for 24 frames per R-Rinterval for 10 minutes. Processing was performed with commer-cially available software (SYMA; Elscint, Hackensack, NJ), typ-ical of processing used for ERNA data.18 For RV EF assessment,observers drew initial RV outlines, primarily guided by thevisual impression of the RV shape as seen at end diastole, aidedby Fourier amplitude and phase maps.5 For all time intervals ofthe complete ERNA study, algorithms generated regions auto-matically on the basis of the limiting end-diastolic (ED) region,which observers reviewed and modified as necessary.

SPECT ERNA

Immediately after planar ERNA, all patients underwentSPECT ERNA. A dual-detector gamma camera (CardiaL,Elscint) was used to collect images at 64 projections over a180° circular arc. Low-energy, general-purpose collimatorswere used to acquire 64 � 64 tomograms with a pixel size of6.4 mm for 20 seconds per projection. Tomograms wereacquired with patients at rest, at 8 frames per R-R interval, withuse of a 100% R-wave window. All SPECT ERNA patient datawere screened algorithmically for arrhythmic artifacts,19 on thebasis of which no data were excluded. Eight frames per R-Rinterval were used, as opposed to a higher number, to guaranteeroutine collection of sufficient tomographic counts.

Pre-reconstruction spatial smoothing was performed bymeans of Butterworth filtering (cutoff, 0.55 cycles/cm; power,5.0) for gated tomograms, followed by quantitative rampfiltering in the transaxial plane applied during backprojection toform transaxial slices.20 Images were reoriented into short-axis(SA) sections through use of manual choices of anterior,inferior, septal, and lateral limits and approximate LV symme-try axes. Reconstruction-limiting angles and planes were cho-sen to ensure that the entire right ventricle and left ventricle andportions of the RA and pulmonary artery were included inreoriented images.

SPECT ERNA Algorithms

By searching for maximum count areas in volumetricregions likely occupied by the right ventricle, algorithmsidentified mid-RV planes, indicated as boxes framing estimatedmid-plane locations projected onto simultaneous cines of SA,vertical long-axis (VLA), and horizontal long-axis (HLA)projections (Figure 1). Background counts per 3-dimensionalimage pixel (voxel) were tabulated in the volume beneath theright ventricle at the edge of SA imaging matrices.

An ED VLA region was generated that included the rightventricle within count thresholds set at 35% of the differencebetween background counts and maximum ED VLA RV

154 Nichols et al Journal of Nuclear CardiologyValidation of SPECT ERNA right ventricular parameters March/April 2002

counts, while avoiding the RA and pulmonary artery based onFourier phase images. Algorithms then generated an end-systolic (ES) VLA RV region, based on the ED region andphase images. Algorithms generated the third (and final) RVED SA region at the 35% difference threshold between back-ground counts and maximum ED SA RV counts, while avoid-ing LV and pulmonary artery areas (Figure 2).

Moving tricuspid valve planes were interpolated from RVED and ES valve planes for all R-R intervals. These planeswere used to limit, in the RA direction, the number of SA slicesincluded in subsequent volume calculations. Usually, thisresulted in an appropriately greater number of basal ED planesbeing included in defining the right ventricle than basal ESplanes. ED and ES VLA regions also were used to defineautomatically the moving pulmonary valve plane, such thatmaximum heights of SA outlines were limited to conform topulmonary valve plane limits.

All SPECT ERNA SA regions for all tomographic sec-tions and all R-R intervals were generated automatically ascorresponding to those contiguous regions containing countsgreater than 35% of the difference between maximum volu-metric RV ED counts over background counts, a thresholdvalue that has been used frequently by investigators in derivingmyocardial surfaces from myocardial perfusion or blood poolimages.21 The set of all regions superimposed on all tomogramsat all levels was shown to observers in cinematic format (Figure3). If observers perceived that alterations were needed to betterdefine RV regions, they reran the algorithms and manuallyintervened to change mid-RV planes and RV outlines toachieve optimal agreement of calculated with perceived actualRV boundaries. Observers recorded the percent of time it wasnecessary to make changes of any sort during any stage of dataprocessing. The primary criterion for identifying voxels be-longing to the right ventricle was to find those for which countsdecreased as the left ventricle was observed to contract.

RV ED and ES volumes were computed geometrically as3-dimensional volumes defined at 35% of maximum cardiaccounts, and EFs were computed from changes in countssummed over all included RV voxels. These steps wereaccomplished through use of computer programming software(IDL; Research Systems Inc, Boulder, Colo) implemented on aWindows-NT 500-MHz personal computer. For interobservervariability assessment, calculations were performed independ-ently by a second observer. For intraobserver reproducibilityassessment, calculations were performed by 1 observer on 2occasions separated by more than 1 month, without reference tothe values obtained on the first occasion. Both observers wereblinded to the other’s results, as well as to a third observer’sresults of MRI calculations, as described below.

Figure 1. The ED cinematic frame of tomographic slices forthe same patient in SA (top), VLA (middle), and HLA (bottom)orientations. The identity of the tomographic section at themiddle of the right ventricle is determined automatically andshown to an observer as a highlighted image, along with allother tomographic sections. Only the 3 frames closest tomid-RV locations are depicted here.

Figure 2. Automatically generated RV regions are shownsuperimposed on the VLA ED (left), VLA ES (middle), and SAED (right) mid-RV tomographic sections. The right ventricle(RV), RA, and left ventricle (LV) also are labeled.

Figure 3. The ED frames of all tomographic slices for the samepatient are shown with the final RV regions used to computedvolume and EF, with SA sections in the top 3 rows, VLAsections in the middle 2 rows, and HLA sections in the bottom2 rows.

Journal of Nuclear Cardiology Nichols et al 155Volume 9, Number 2;153-60 Validation of SPECT ERNA right ventricular parameters

MRI

Cardiac gated gradient-echo cine MRI evaluation wasperformed by means of a 1.5-T scanner (either LX or SignaHorizon; GE Medical Systems, Inc, Milwaukee, Wis) with abody phased-array surface coil. The cine images were acquiredthrough use of a “spoiled gradient recall” non—breath-holdtechnique (n � 8) (TR min, TE 13, FA 30, matrix 256 � 128,FOV 30-41, NEX 1, S1 thickness 8, gap 0) or a breath-holdtechnique (n � 20) (TR min, TE min full, FA 15, matrix 256� 128, FOV 30-41, NEX 1, S1 thickness 8, gap 0). For thebreath-hold technique, the breath was held in end expirationafter the patient took a deep breath. Several levels of 8-mm-thick myocardial slices were acquired for 16 to 20 timeintervals per cardiac cycle. The cine sequences were gatedretrospectively and covered the complete R-R cycle. Foracquisition of 2-dimensional MRI spin-echo and cine imagesalong axes specific to an individual patient’s heart, an initialscan was performed to isolate the heart. This was followed bya further acquisition along the approximate axes of the LVapex-to-base direction. Vertical 2-chamber, horizontal 4-cham-ber, and SA views were obtained along the axes specific to anindividual patient’s heart. The SA images encompassed theentire ventricle (apex to atrioventricular valves). Total imagingtime, including patient set-up time, was less than 1 hour.

Data were analyzed with the aid of semi-automated algo-rithms (Cardiac Analysis software, GE Medical Systems, Inc) forthe delineation of the left and right heart chambers through all SAlevels of the heart at end diastole and end systole. Automaticallygenerated outlines were reviewed and amended as necessary. MRIvalues were derived independently by Simpson’s rule from semi-automated SA regions that were modified manually to conform toendocardial borders and to take into account trabeculations,primarily in the apical third of the right ventricle.

Statistical Analysis

Numerical results are reported as mean values �1 SD.Differences among EF results are reported in absolute EF units,not as percentages of EFs. For comparisons between 2 meth-ods, paired t tests were used to compute whether 2-tailedprobabilities indicated statistical significance, at the level of P� .05. The unpaired t test was used to compare mean valuesbetween different populations. Linear regression analysis wasused to compare calculations of volumes and EFs betweenmodalities, for which computed probability values of noassociation were considered statistically significant if � .05,and in conjunction with Bland-Altman analyses of differencesversus means of paired values, to search for trends andsystematic errors. Statistical significance of pairs of differentregression results was assessed by the Fisher z test.

RESULTS

Comparisons With MRI

Although planar ERNA RV EFs correlated signifi-cantly with MRI (r � 0.76; P � .0001) (Figure 4), RV

EF was slightly (but statistically significantly) lower forplanar ERNA than for MRI (35% � 11% vs 41% �10%, P � .04) in patients with PPH or TOF. PlanarERNA RV EF also was significantly lower than forSPECT ERNA values (35% � 11% vs 42% � 11%, P �.02), with significant correlation between them (r � 0.70;P � .0001) (Figure 4). SPECT ERNA RV EF correlatedsignificantly more strongly with MRI than did planarERNA (r � 0.85 vs r � 0.75; P � .01).

Mean SPECT ERNA and MRI values were notstatistically significantly different for RV EF, ED vol-ume, and ES volume (42% � 11% vs 41% � 10%, 135� 67 mL vs 139 � 91 mL, and 87 � 54 mL vs 85 � 61mL, respectively) in patients with PPH or TOF. Signif-icant linear correlation (P � .0001) was found betweenSPECT ERNA and MRI for both RV ED and ESvolumes (r � 0.94 and r � 0.93, respectively) (Figure 5).Bland-Altman analyses of differences versus meanSPECT ERNA and MRI values demonstrated no statis-tically significant trends or biases for RV EF. Significantunderestimation of RV ED volume by SPECT ERNAcompared with MRI was found, when all data wereanalyzed including the highest MRI value of 496 mL (r� 0.61; P � .001) (Figure 6). Results for RV ES volumewere similar to those for RV ED volume (y � 18 mL �0.82x, standard error of the estimate [SEE] � 20 mL for

Figure 4. Linear regression curves of planar ERNA versusMRI values of RV EF (A) and planar ERNA versus SPECTERNA values of RV EF (B). Solid lines are linear regressioncurves, and dashed lines are the lines of identity.


SPECT ERNA vs MRI linear regression; y � 14 mL ��0.13x, P � not significant, SEE � 21 mL for Bland-Altman analyses of RV ES volume values [compare withFigures 5 and 6]).

To determine whether the unusually high value ofMRI RV ED volume of 496 mL was unduly influencingresults, RV volume and EF data were reanalyzed exclud-ing that one subject’s information. When this was done,there were no statistically significant differences betweencomparing SPECT ERNA with MRI values for anyparameter.

Comparisons With Patients Without CardiacDisease

For the 15 patients without cardiac disease evaluatedbefore beginning chemotherapy, SPECT ERNA RV EDvolume was 59 � 15 mL and RV EF was 54% � 9%.Compared with these values, SPECT ERNA RV EDvolume of 135 � 66 mL was significantly higher (P �.0001) and RV EF of 41% � 10% lower (P � .01) in the28 patients with PPH or TOF. These results are consis-tent with the expectation that the PPH and TOF popula-tion should exhibit abnormally high RV volumes anddepressed EF.

Data Processing Reproducibility

Interobserver comparisons showed no statisticallysignificant differences between repeated measurementsof any parameters and significant linear correlation (P �.0001) between repeated measurements of RV EF, EDvolume, and ES volume (r � 0.92, r � 0.91, and r �0.95, respectively) (Figure 7). Similarly, intraobserverreproducibility studies showed no statistically significantdifferences between repeated measurements of any pa-rameters and significant linear correlation (P � .0001)between repeated measurements of RV EF, ED volume,and ES volume (r � 0.90, r � 0.98, and r � 0.97,respectively) (Figure 8). Bland-Altman analyses con-firmed that there were no statistically significant interob-server or intraobserver biases nor any trends detected.This degree of data processing reproducibility was facil-itated through automation, as, overall, observers found itnecessary to alter automatically determined choices ofmid-RV planes and/or RV regions 25% of the time. Themost common cause of automation failure was excessivebackground counts beneath the right ventricle, whichnecessitated manual alterations to the inferior wall of RVregions. Data processing time was found to range from10 seconds, when no manual changes were required, to 2minutes, when changes were required for all choices.

Figure 5. Linear regression curves of SPECT ERNA versusMRI values of RV EF (A) and RV ED volume (RVEDV) (B).Solid lines are linear regression curves, and dashed lines are thelines of identity.

Figure 6. Bland-Altman curves of differences versus meanvalues between SPECT ERNA and MRI values of RV EF (A)and RV ED volume (RVEDV) (B). Solid lines are linearregression curves, and dashed lines are the lines of identity.


DISCUSSION

Significant progress in surgical correction of TOF2

and advances in therapeutic interventions for PPH1 haverenewed interest in measuring RV volumes and EF. Inaddition, an increasingly large segment of the populationis affected by systemic hypertension,22 which is alsoassociated with right heart disease,23 and centers per-forming lung reduction surgery for emphysema havereported variable results, which might be improved byknowledge of presurgical RV EF.24 A widely availableautomated imaging technique for assessing RV EF andvolume could be important in diagnosis and evaluationof therapeutic interventions in patients with RV disor-ders.

We found excellent correlation between SPECTERNA and MRI measurements for RV volume and EF.In contrast, RV EF was underestimated by planar ERNA,as a result of the contamination of RV counts by RAcounts, which significantly reduces computed RV EF, asRA activity increases as RV activity decreases (Figure2).

SPECT ERNA has the potential advantage overMRI of deriving its volumetric information from de-tected gamma rays rather than from geometrical consid-

erations so that RV trabeculations should not interferewith received counts. That correlation of RV EF betweenSPECT ERNA and MRI was not higher than r � 0.85may have been influenced by the necessity for the MRIobserver to adjust RV outlines in the apical third of theright ventricle for trabeculations. Whereas excellentagreement was found between SPECT ERNA and MRIfor RV volumes, particularly for volumes less than 250mL, SPECT ERNA tended to underestimate the highestMRI volume of 496 mL, possibly because of radiationattenuation through the reduction of both counts andcontrast between blood pool and noncardiac radioactivityas depth increased. Adjustments may be needed for thehighest volumes to obtain optimal accuracy, which mightbe accomplished through maximum likelihood recon-struction algorithms along with attenuation scans, similarto techniques applied successfully to myocardial perfu-sion gated SPECT data.25 In the meantime, regressionresults such as those obtained in this investigation can beused to predict what MRI values would have beenobtained given the SPECT ERNA measurements, in afashion similar to that performed for predicting cardiacMRI and angiographic results from myocardial perfusiongated SPECT measurements as provided in some com-mercially available software packages.26

Figure 7. Interobserver variability is demonstrated by linearregression curves of SPECT ERNA measurements graphed for2 different observers of RV EF (A) and RV ED volume(RVEDV) (B). Solid lines are linear regression curves, anddashed lines are the lines of identity.

Figure 8. Intraobserver reproducibility is demonstrated bylinear regression curves of SPECT ERNA observation 2 versusobservation 1 values for the same observer of RV EF (A) andRV ED volume (RVEDV) (B). Solid lines are linear regressioncurves, and dashed lines are the lines of identity.


Because of their demonstrated accuracy and repro-ducibility, planar ERNA and myocardial perfusion gatedSPECT can be used for evaluation of LV EF, which is aprime prognostic indicator,27 and LV volumes also helpto refine prognostic assessment.28,29 RV EF and volumemeasurements should serve a similar role for stagingright heart function and assessing cardiac remodeling be-fore and after surgery or other interventional strategies.

Study Limitations

The finding of volume underestimation may havebeen skewed by one unusually high value, and more datafor the largest right ventricles are needed to clarify thisissue. Data for larger groups of patients without cardiacdisease also will be needed for a realistic formation ofnormal limits, because all of those analyzed here werefemale patients; it will be necessary to acquire data formale patients as well for gender-specific SPECT ERNARV normal limits.30

It should be noted that patients with congenital heartdisease usually need high-quality anatomic images, aswell as the functional information, and that MRI showsconsiderably more anatomic detail than does SPECTERNA. Also, MRI can quantitate valvular regurgitation,unlike SPECT ERNA. Therefore MRI would be the firstchoice for evaluating patients with congenital heartdisease, but its ability to detect and quantify RV dys-function rapidly and easily may well increase the use ofSPECT ERNA.

Conclusion

RV volumes and EFs can be calculated accuratelyby SPECT ERNA. This technique is a feasible approachfor obtaining RV parameters routinely in patients withsignificant right heart enlargement as well as right heartdysfunction. SPECT ERNA should prove to be useful indiagnosis, as well as in following disease progressionand evaluating the efficacy of therapeutic interventionsin patients with RV abnormalities.

Acknowledgment and Disclosure

We gratefully acknowledge Ketan Bhatia, BS, JaniceGibbs, BS, Mohammed Khan, BS, Gladys P. Kusterer, MS, andCarlo Laraia, CNMT, for their assistance in collecting clinicalpatient data and Julia Rheem, MBBS, for assistance withanalysis of image data.

The work described in this article was funded by a grantfrom Siemens Medical Systems, Inc. Negotiations are underway for the algorithms described here to be marketed bySyntermed, Inc, at which time some of the authors (K.N. andM.W.G.) will receive royalties, as will Columbia University.

References

1. Barst RJ, Rubin LJ, Long WA, McGoon MD, Rich S, BadeschDB, et al. A comparison of continuous intravenous epoprostenol(prostacyclin) with conventional therapy in primary pulmonaryhypertension. New Engl J Med 1996;334:296-301.

2. Schamberger MS, Hurwitz RA. Course of right and leftventricular function in patients with pulmonary insufficiencyafter repair of tetralogy of Fallot. Pediatr Cardiol 2000;21:244-8.

3. Berger HJ, Matthay RA, Loke J, Marshall RC, Gottschalk A,Zaret BL. Assessment of cardiac performance with quantitativeradionuclide angiocardiography: right ventricular ejectionfraction with reference to findings in chronic obstructivepulmonary disease. Am J Cardiol 1978;41:897-905.

4. Maltz DL, Treves S. Quantitative radionuclideangiocardiography. determination of Qp:Qs in children.Circulation 1973;47:1049-56.

5. Maddahi J, Berman DS, Matsuoka DT, Waxman AD, StankusKE, Forrester JS, et al. A new technique for assessing rightventricular ejection fraction using rapid multiple-gatedequilibrium cardiac blood pool scintigraphy: description,validation and findings in chronic coronary artery disease.Circulation 1979;60:581-9.

6. Schulman DS. Assessment of the right ventricle withradionuclide techniques. J Nucl Cardiol 1996;3:253-64.

7. Markewicz W, Sechtem U, Higgins CB. Evaluation of the rightventricle by magnetic resonance imaging. Am Heart J 1987;113:8-15.

8. Bouchard A, Higgins CB, Byrd BF III, Amparo EG, Osaki L,Axelrod R. Magnetic resonance imaging in pulmonary arterialhypertension. Am J Cardiol 1985;56:938-42.

9. Johnson DE, Vacek J, Gollub SB, Wilson DB, Dunn M.Comparison of gated cardiac magnetic resonance imaging andtwo-dimensional echocardiography for the evaluation of rightventricular thrombi: a case report with autopsy correlation.Cathet Cardiovasc Diagn 1988;14:266-8.

10. Sechtem U, Pflugfelder PW, Gould RG, Cassidy MM, HigginsCB. Measurement of right and left ventricular volumes inhealthy individuals with cine MR imaging. Radiology 1987;163:697-702.

11. Boxt LM, Katz J, Kolb T, Czegledy FP, Barst RJ. Directquantitation of right and left ventricular volumes with nuclearmagnetic resonance imaging in patients with primary pulmonaryhypertension. J Am Coll Cardiol 1992;19:1508-15.

12. Faber TL, Stokely EM, Templeton GH, Akers MS, Parkey RW,Corbett JR. Quantification of three-dimensional left ventricularsegmental wall motion and volumes from gated tomographicradionuclide ventriculograms. J Nucl Med 1989;30:638-49.

13. Bartlett ML, Srinivasan G, Barker WC, Kitsiou AN, Dilsizian V,Bacharach SL. Left ventricular ejection fraction. comparison ofresults from planar and SPECT gated blood-pool studies. J NuclMed 1996;37:1795-9.

14. Groch MW, Marshall RC, Erwin WD, Schippers DJ, BarnettCA, Leidholt EM Jr. Quantitative gated blood pool SPECT forthe assessment of coronary artery disease at rest. J Nucl Cardiol1998;5:567-73.

15. Van Kriekinge SD, Berman DS, Germano G. Automaticquantification of left ventricular ejection fraction from gatedblood pool SPECT. J Nucl Cardiol 1999;6:498-506.

16. Chin BB, Bloomgarden DC, Xia W, Kim HJ, Fayadz ZA,Ferrari VA, et al. Right and left ventricular volume and ejectionfraction by tomographic gated blood-pool scintigraphy. J NuclMed 1997;38:942-8.


17. Vanhove C, Franken PR, Defrise M, Momen A, Everaert H,Bossuyt A. Automatic determination of left ventricular ejectionfraction from gated blood-pool tomography. J Nucl Med 2001;42:401-7.

18. Christian PE, Nortmaan CA, Taylor A. Comparison of fullyautomated and manual ejection fraction calculations: validationand pitfalls. J Nucl Med 1985;26:775-82.

19. Nichols K, Yao SS, Kamran M, Faber TL, Cooke CD, DePueyEG. Clinical impact of arrhythmias on gated SPECT cardiacmyocardial perfusion and function assessment. J Nucl Cardiol2001;8:19-30.

20. Crawford CR. CT filtration aliasing artifacts. IEEE Trans MedImaging 1991;10:99-102.

21. Nichols K, DePuey EG, Rozanski A. Automation of gatedtomographic left ventricular ejection fraction. J Nucl Cardiol1996;3:475-82.

22. Joint National Committee on Detection, Evaluation, andTreatment of High Blood Pressure. The fifth report of the JointNational Committee on Detection, Evaluation, and Treatment ofHigh Blood Pressure (JNC V). Arch Intern Med 1993;153:154-83.

23. Chronic cor pulmonale. Report of an expert committee. WorldHealth Organ Tech Rep Ser 1961;213:1-51.

24. Geddes D, Davies M, Koyama H, Hansell D, Pastorino U,Pepper J, et al. Effect of lung-volume-reduction surgery in

patients with severe emphysema. New Engl J Med 2000;343:239-45.

25. Links JM, Becker LC, Rigo P, Taillefer R, Hanelin L, Anstett F,et al. Combined corrections for attenuation, depth-dependentblur, and motion in cardiac SPECT: a multicenter trial. J NuclCardiol 2000;7:414-25.

26. Faber TL, Cooke DC, Folks RD, Vansant JP, Nichols KJ, DePueyEG, et al. Left ventricular function from gated SPECT perfusionimages: an integrated method. J Nucl Med 1999;40:650-9.

27. Iskandrian AS, Hakki AH, Goel IP, Mundth ED, Kane-MarschSA, Schenk CL. The use of rest and exercise radionuclideventriculography in risk stratification in patients with suspectedcoronary artery disease. Am Heart J 1985;110:864-72.

28. Fuster V, Gersh BJ, Giuliani ER, Tajik AJ, Brandenburg RO,Frye RL. The natural history of idiopathic dilatedcardiomyopathy. Am J Cardiol 1981;47:525-31.

29. Germano G, Kiat H, Kavanaugh PB, Mariel M, Mazzanti M, SuHT, et al. Automatic quantification of ejection fraction fromgated myocardial perfusion SPECT. J Nucl Med 1995;36:2138-47.

30. Rozanski A, Nichols K, Yao SS, Malhotra S, Cohen R, DePueyEG. Development and application of normal limits for leftventricular ejection fraction and volume measurements fromTc-99m sestamibi myocardial perfusion gated SPECT. J NuclMed 2000;41:1445-50.


validation of spect equilibrium radionuclide angiographic right ventricular parameters by cardiac...

Documents