Hemodynamic Evaluation of the Heart With a Nonfluoroscopic Electromechanical Mapping Technique
Background Clinical cardiac volumetric measurement techniques are essential for assessing cardiac performance but produce significant inaccuracies in extrapolation of the volume of a three-dimensional (3D) object from two-dimensional images and lack the ability to associate cardiac electrical and mechanical activities. In this study, we tested the accuracy of cardiac volumetric measurements using a new catheter-based system.
Methods and Results The system uses magnetic technology to accurately locate a special catheter at a frequency of 125 Hz and is currently used in the field of electrophysiology, in which activation maps are superimposed on the 3D geometry of the cardiac chamber. The mapping procedure is based on sequentially acquiring the location of the tip and local electrogram while in contact with the endocardium. The 3D geometry of the chamber is reconstructed in real time, and its volume could be calculated at every time step (8 ms). The volumetric measurements of the system were found to be highly accurate for simple phantoms (mean±SEM deviation, 2.3±1.1%), left ventricular casts (9.6±1.3%), and a dynamic test jig. In addition, left ventricular volumes of 12 swine were measured. Intraobserver and interobserver variabilities were found to be minimal (ejection fraction, 6.5±1.9% and 7.1±2.0%; stroke volume, 4.5±1.0% and 11.3±2.4%). Comparison with the thermodilution method for measuring stroke volume showed an average deviation of 8.1±2.2%. Typical pressure-volume loops were also obtained.
Conclusions The new mapping image provides, for the first time, simultaneous information regarding cardiac mechanics, hemodynamics, and electrical properties. Furthermore, all this information is achieved without the use of fluoroscopy, contrast medium, or complicated image processing.
Volumetric measurements of the chambers of the heart during the different stages of the cardiac cycle are essential for the calculation of specific parameters such as EF and SV. These parameters play a major role in the assessment of patients with a variety of cardiac disorders in the determination of the diagnosis and prognosis and guidance of the management of these patients.1,2 Consequently, a variety of interventional and noninterventional techniques have been developed to measure these volumes, such as echocardiography,3–5 angiocardiography,6,7 isotope-scanning techniques,8,9 computed tomography,10 and MRI.11
Although widely used, currently used clinical methods, such as conventional echocardiography and angiocardiography, are limited by the appreciable inaccuracies that result from extrapolation of the volume of a 3D object from 2D measurements, geometric assumption regarding the shape of the ventricle, position and orientation of the imaging planes, and problematic edge detection. These limitations can result in significant inaccuracies and interobserver variabilities.12,13
Moreover, at present, none of the methods discussed above have the ability to associate between the electrical activity of the heart and its regional and global (volumetric changes) mechanics. With the growing interest in evaluation of the electromechanical coupling of the heart and with new data available regarding the importance of the hemodynamic effect of single-site or multisite pacing, a single method that could combine information on both the cardiac mechanics and electrical functions could provide valuable information for both basic and clinical cardiology.
A new method is presented for in vivo measurement of cardiac volumes using a computerized nonfluoroscopic, catheter-based mapping system that enables the generation of 3D electromechanical maps. The new method is currently used in the field of electrophysiology, in which activation maps are superimposed on the anatomy of the chamber.14,15 The method is based on the use of a new locatable catheter connected to an endocardial mapping and navigation system. The system uses magnetic technology to accurately determine the location and orientation of the catheter and simultaneously records the intracardiac local electrogram from its tip. The 3D geometry of the chamber, from which the volume can be calculated, is reconstructed in real time from the set of endocardial sites that were sampled. The electrophysiological information is color coded and superimposed on the electromechanical map.
The overall objective of the present study was to test the possible role of this new mapping method in evaluating cardiac mechanics. Several stages were undertaken in the process of testing the accuracy of the volumetric measurements of this technique. First, simple objects with known volumes were tested. We then studied the accuracy of the volumetric measurements of LV casts and a custom-built dynamic test jig. Finally, the accuracy and reproducibility of in vivo LV volume measurements were studied.
The navigation and mapping system is composed of a miniature passive magnetic field sensor, an external ultralow magnetic field emitter (location pad), and a processing unit (NOGA, Biosense).
The locatable catheter (NAVI-STAR, Cordis-Webster) is similar to a regular electrophysiological 8F deflectable-tip catheter; just proximal to the tip electrode lies the location sensor, totally embedded within the catheter.
The locator pad is located beneath the operating table and generates ultralow magnetic fields (5×10−6 to 5×10−5 T) that contain the information necessary to resolve the location and orientation of the sensor in 6 df. The accuracy of the location capabilities of the system has been tested before and was shown to be 0.16 and 0.73 mm for both in vitro and in vivo studies, respectively.15
Two locatable catheters were introduced into the heart. The first catheter was inserted into the coronary sinus or RV apex and served as a reference catheter. A mapping catheter was then inserted into the mapped chamber. The location of the tip of the mapping catheter while inside the heart was recorded relative to the fixed intracardiac reference catheter. This relative location determination enabled correction for motion of the animal as well as of the heart within the animal (ie, respiration). Through movement of the catheter inside the heart, the system continuously analyzed its end-diastolic location and presented it to the user, thus enabling navigation without the use of fluoroscopy (Fig 1⇓).
The mapping procedure is based on dragging the mapping catheter randomly over the endocardium and sequentially acquiring the location of the tip while in contact with the endocardium. The set of points collected comprises an irregularly sampled data set of location points that are members of the endocardial surface. Chamber geometry is then reconstructed, in real time, using the set of sampled location points. The endocardial surface is presented as a set of polygons (triangles) whose vertices are the sampled points. The LAT at each site is determined as the time difference between a selected fiducial point on the body-surface ECG and the steepest negative intrinsic deflection (minimum dV/dt) in the unipolar intracardiac electrogram (filtered at 0.5 to 400 Hz) recorded from the tip of the mapping catheter. The activation map is color coded and superimposed on the 3D chamber geometry. The center of mass of the reconstructed chamber is automatically calculated from the set of the surface points. The volume of the chamber can be calculated from the sum of the volumes of all tetrahedrons constructed when connecting the center of mass to all triangles forming the reconstructed surface.
At each site, the spatial coordinates of the tip of the catheter were recorded at a frequency of 125 Hz. Hence, the movement of each endocardial site throughout the cardiac cycle could be tracked and analyzed. This enabled the creation of a dynamically beating image of the 3D map from which the volume at each time step (8 ms) could be calculated.
Phantom Volumetric Measurements
To test the accuracy of the reconstruction algorithm, two objects with known volumes were used: an ellipsoid with a volume of 83 mL and a tennis ball with a volume of 158 mL. Each object was mapped using up to 70 sampled points, and the volume of the reconstructed object was compared with the actual volume. The procedure was carried out five times for each object.
Volumetric Measurements of LV Casts
At this stage, we tested the accuracy of the volumetric measurements of six casts of the swine LV. The casts were made using polyester from six postmortem pig hearts. Each cast was mapped using the locatable catheter, which sequentially acquired points from the cast surface. The procedure was carried out twice for each cast. The volume calculated by the system was then compared with the actual volume of the cast.
Volumetric Measurements of a Specially Built Test Jig
A cylinder-shaped, dynamic test jig was used to study the in vitro accuracy of the volumetric measurements of the system. The maximal and minimal volumes and the calculated EF were measured by the system and compared with the actual volumes of the test jig (maximal volume, 547.3 mL; minimal volume, 377 mL; SV, 169 mL; EF, 31%).
Studies were performed on healthy male pigs weighing 30 to 40 kg. The experimental protocol was approved by the Animal Use and Care Committee of the Technion Faculty of Medicine. All animals were anesthetized with pentobarbital (30 mg/kg), intubated, and placed on a Harvard large-animal mechanical respirator. Vascular access was obtained using vascular cutdown of the jugular veins, carotid artery, and femoral vein and artery as needed.
In some of the experiments, a 5F solid-state pressure catheter (Millar Instruments) was positioned in the LV for simultaneous pressure recording. A 6F Swan-Ganz thermodilution catheter was advanced from the jugular vein to be positioned in the pulmonary artery.
The protocol used in our studies consisted of taking the first three points using fluoroscopy (two at the base near the aortic and mitral valves and one at the apex). EDV and ESV were calculated as the maximal and minimal volumes throughout the cardiac cycle. SV was calculated as the difference between the EDV and the ESV, and EF was calculated as SV divided by EDV.
Twelve animals were studied. The LV was mapped twice during sinus rhythm by the same observer, who was blinded to the volumetric results to determine intraobserver variability. After acquisition of the first map, the user immediately proceeded to acquire the second map. Points were rejected from the according to the following criteria: (1) premature beat or a beat after a premature beat, (2) cycle length that deviated >15% of the average cycle length, (3) end-diastolic location stability (measured as the difference in end-diastolic location between two consecutive beats) that was >2 mm, and (4) LAT stability (defined as the difference between the LATs of two consecutive beats) that was >2 ms.
In all pigs, thermodilution cardiac output measurements were acquired with a cardiac output computer in the standard way. Four measurements of cardiac output were obtained and averaged immediately. SV was calculated by dividing the average thermodilution cardiac output by the heart rate and later compared with the calculated SV measured by the new method.
In eight of the pigs studied, the LV was also mapped by a second observer, immediately after completion of the second map, and interobserver variability was determined.
To ensure hemodynamic stability among the three maps that were obtained, both heart rate and SV measurements were monitored throughout the experiment. No animals were eliminated from experimentation because of failure to meet the criteria of the experiments.
Results are reported as mean±SEM. The volumetric measurements of the casts by the mapping technique were compared with their actual volume by simple linear regression analysis. In the animal studies, intraobserver and interobserver variabilities for determination of EDV, ESV, SV, and EF were calculated as the difference between the two measurements expressed as a percentage of the average value. Simple linear regression and r values were used to study the correlation between SV calculated by the new method and SV derived from thermodilution measurements. SEE values were calculated in the usual manner.
Initially, the accuracy of the reconstruction algorithm was tested with known phantoms. As can be seen from Table 1⇓, the calculated volume of the reconstructed objects was very close to the actual volume, with the average deviation being 3.2±0.4% and 2.3±1.1% for the tennis ball and ellipsoid, respectively. Fig 2⇓ illustrates a typical graph of the volume of the reconstructed object as a function of the number of points sampled. Both curves reached a plateau value after 40 to 60 points.
Volumetric Measurements of LV Casts
At this stage, the volumetric measurements of the reconstructed images of six LV casts were compared with their actual volumes. The volumetric measurements highly correlated with the actual volume of the cast (r=.94, y=0.89x+2.4), with an average deviation of 9.6±1.3% (Table 2⇓).
Volumetric Measurements of a Specially Built Test Jig
A custom-built dynamic test jig was used to test the in vitro accuracy of the new mapping method in determining dynamic volume changes of a moving object. As can be seen from Table 3⇓, the volume measurements of the system (maximal and minimal volumes) and the calculated hemodynamic parameters (SV and EF) were highly accurate, with an average deviation of 1.4±0.3%, 0.7±0.3%, 6.0±0.4%, and 5.2±0.9% for maximal volume, minimal volume, SV, and EF, respectively.
From our initial experience, we found that after acquisition of the first three points under fluoroscopic guidance, the remainder of the mapping procedure could be achieved without the use of fluoroscopy. We also noted that the mapping procedure could be achieved relatively quickly, during an average of 15 to 20 minutes. No complications were noted.
Typical end-diastolic and end-systolic electromechanical maps of the swine LV are presented in Fig 3⇓. All maps showed a similar activation pattern during sinus rhythms, with the earliest site of activation (the red area in the map) located in the superior septum. The activation then spread to the remainder of the ventricle, with the posterolateral areas and atrioventricular ring being activated last (blue and purple areas). Total activation time of this ventricle was 55 ms. The measured volumes of the LV shown in this figure were EDV, 61.0 mL; ESV, 33.2 mL; SV, 27.9 mL; and EF, 46%. Fig 4⇓ represents the relation between the calculated parameters and the number of points sampled during a typical procedure. Note that the EDV initially increased significantly and then reached a plateau at ≈50 points. In contrast, the EF reached its final value after ≈10 points.
As described above, the volume of the ventricle could be calculated at time steps of 8 ms. This enabled continuous determination of the volume throughout the cardiac cycle (during both the systolic and diastolic periods). A typical plot of the LV volume during the cardiac cycle is presented in Fig 5⇓ (top left). Fig 5⇓ (top right) plots the intraventricular pressure recorded simultaneously. Using the information presented in both plots, we could now determine the relationship between pressure and volume throughout the cardiac cycle. Thus, typical PV curves, as presented in Fig 5⇓ (bottom), could be obtained.
To determine intraobserver and interobserver variabilities, each heart was mapped three times. In this group, mean EDV and ESV were 51.5±3.2 and 29.6±1.5 mL, respectively, with mean calculated SV and EF of 21.8±2.3 mL and 42.0±2.0%, respectively. Intraobserver variability (Table 4⇓) was found to be relatively small (EDV, 8.0±1.5%; ESV, 12.2±3.1%; SV, 4.5±1.0%; EF, 6.5±1.9%). Similarly, interobserver variability (Table 5⇓) was relatively small (EDV, 5.9±1.7%; ESV, 7.5±1.1%; SV, 11.3±2.4%; EF, 7.1±2.0%).
Comparison with Thermodilution Measurements
SV was calculated with the system as EDV−ESV and later correlated with SV measurements using the thermodilution technique. The average error between the two methods was found to be 8.1±2.2%. Fig 6⇓ depicts the linear regression plot. SV measurements highly correlated with those determined by thermodilution, with r value of .93, slope of 0.95, y intercept of −0.12 mL, and SEE of 1.7 mL.
We presented the volumetric accuracy data for a new catheter-based method for studying cardiac mechanical, hemodynamic, and electrophysiological functions. The new method samples the location of a special catheter throughout the cardiac cycle at a plurality of endocardial sites and reconstructs a dynamic 3D electromechanical map of the chamber. For the volumetric measurements of the new method to be correct, three conditions must be met: (1) the location of each sampled site must be highly accurate, (2) the number of sites sampled should be adequate to represent the entire chamber, and (3) the reconstruction algorithm should be accurate and reproducible.
To test the accuracy and repeatability of the volumetric measurements of the new method, several steps were undertaken. First, the location accuracy of the new method has been shown in both in vivo and in vitro studies to be <1 mm.15 To further evaluate the precision of the reconstruction algorithm, the accuracy of volumetric measurements of simple phantoms and LV casts was tested and found to be highly accurate (2.3±1.1% and 3.2±0.4% deviation for the phantom study and 9.6±1.3% for the LV cast study). To test the volumetric accuracy of moving objects, we used a cylinder-shaped dynamic test jig. We again found that the volumetric measurements were accurate, with the error ranging from 0.7% to 5%.
The final stage of the study demonstrated that in closed-chest anesthetized pigs, measurement of the LV volumes and hemodynamic parameters could be achieved with excellent reproducibility and a high degree of precision. Both intraobserver and interobserver variabilities were found to be very low. In addition, we observed an excellent correlation with SV measured with thermodilution, which was previously established as an accurate method for determining SV.16 Interestingly, we noted that EF reached stable values after a relatively small number of sites were sampled, whereas absolute volume determination (EDV, ESV) required sampling of more points. This finding is not surprising because both EDV and ESV increase simultaneously when more points are sampled, therefore resulting in a relatively constant determination of EF.
Comparison With Other Imaging Modalities
Quantitative determination of LV volumes has long been a cornerstone of diagnostic cardiology. However, common clinical modalities, such as angiographic methods, echocardiography, and radionuclide angiography, may produce significant inaccuracies that result from various assumptions regarding the geometry of the chamber.
Angiocardiographic methods for determining LV may produce significant inaccuracies in determining absolute volumes due to the extrapolation of the volume from 2D images, using geometric assumptions about the ellipsoid nature of the ventricle. Image distortion, varying spatial orientation of the ventricle, and lack of sharpness of ventricular borders may limit the accuracy of these volumetric measurements even further. Several studies have shown that ventriculography tends to overestimate LV volumes.6,12 This may be due to inclusion of areas contained within the trabeculae carneae cordis, to the fact that from any viewing angle the angiographic silhouette creates a maximal area, and to hemodynamic effects of the contrast medium. These limitations may result in lack of precision. For instance, three groups of investigators examined biplane angiographic estimates of SV and compared them with cardiac output measurements based on either the Fick method or the indicator dye dilution method; these different studies reported variation of up to 50%,17 30%,18 and 37%.19
Echocardiography has the advantages of being noninvasive and of not using hazardous radiation. Consequently, various 2D echocardiographic approaches for measuring LV volumes have been described.3,4 Nevertheless, all of these methods make some assumptions regarding the cavity shape and geometric location among the different tomographic imaging planes. Thus, the operator has no means of localizing the image plane in 3D space and must rely solely on image content to position images for volume calculation.
In addition, the in vivo beating heart displays a complicated series of motions in all dimensions in the thoracic cavity during the cardiac cycle. The heart shows variable twist motions, contractions along its long axis, and changes in its relation with the thoracic wall. Because all the techniques described use 2D imaging planes, they face common problems in selecting the appropriate areas or diameters for measurements due to the through-plane motion of the measured area.
The method described in the present study, despite obvious limitations, may bring a unique value to the 3D evaluation of cardiac mechanics because of two conceptual advantages. First, the 3D geometry of the chamber is reconstructed, and thus the volume can be calculated without extrapolation based on geometric assumptions from 2D measurements. Second, the dynamic of the chamber is based on the actual motion of each endocardial point sampled and does not rely on the imaging plane.
The reproducibility of the volumetric measurements of the new method as quantified by intraobserver variability (SV, 4.5%; EF, 6.5%) and interobserver variability (SV, 11.3%; EF, 7.1%) was found to be relatively small and may be comparable to or even better than other techniques. For example, Baur et al20 compared intraobserver and interobserver variabilities for echocardiography and MRI. Intraobserver variability was relatively low for both techniques (in most instances, <5%), whereas interobserver variability was considerably larger (≈10%). In addition, biological variability, defined as the difference in volumetric measurements between beats, was the largest source of error in echocardiography and ranged up to 19% for SV. This type of variability is decreased in MRI, radionuclide angiography, and the present technique because the results are averaged over several beats. Gopal et al21 assessed cardiac function by 3D echocardiography and compared it with conventional echocardiography. Interobserver variability was found to be 10.2%, 26.1%, and 33.3% for 3D echocardiography, quantitative 2D echocardiography using an apical biplane summation of disk algorithm, and visual estimation, respectively. In a similar study, these authors found an interobserver variability of cardiac volumes that ranged from 5% to 8% for 3D echocardiography and 6% to 9% for MRI in normal subjects.22 Similar results were found in other studies comparing quantitative 3D echocardiography with radionuclide angiography23 and MRI24 and observer variabilities found for cine–computed tomography10 and radionuclide angiography.8 The results of the phantom and cast studies may also be comparable to other methods.9,25,26 For example, Krebs et al25 found a correlation (r=.97) similar to that of our study for measurement of LV cast volume using a multiplanar transesophageal echo probe, which was significantly better than that for biplane angiography (r=.88).
Possible Clinical Advantages and Uses
The new method described in this study may offer a number of possible clinical advantages. The ability to accurately determine global hemodynamic parameters, without the use of fluoroscopy, contrast medium, and image processing, may overcome some of the shortcomings of angiocardiographic volumetric measurements. This applies especially to contrast material injection, which can be hazardous and lead to various complications such as renal failure, heart failure due to the injection of hyperosmotic solution, allergic reaction, and so on.27 Moreover, as discussed above, the ability to present accurate 3D information regarding cardiac mechanics may overcome some of the limitations of cardiac hemodynamic evaluation during ventriculography.
The ability to accurately determine LV volume throughout the cardiac cycle simultaneously with intracavitary pressure recordings enables determination of the PV relation (PV loop). In the past 20 years, PV relation analysis has evolved as a dominant approach to estimate cardiac systolic and diastolic abnormalities and is better able to separate chamber function into primary systolic and diastolic functions.28,29 Numerous studies have expanded the usefulness of this approach by defining relations among ventricular PV data, myocardial energetics, and different hemodynamic and pharmacological conditions.30,31
The present study has shown that PV loops can be obtained with the use of the new technique. Nevertheless, these should be regarded as average PV loops because they lack the ability to represent beat-to-beat changes in PV relations due to the sequential nature of the mapping procedure.
The ability to supply valuable information regarding the 3D geometry and mechanics of the chamber, combine PV information, and associate mechanical and electrical data may provide a powerful clinical and research tool. Hence, substantial information may be added with respect to systolic and diastolic function of the LV in both health and disease. Of special interest may be the evaluation of LV diastolic function because it may contribute significantly to LV dysfunction in a large number of patients. Furthermore, by altering ventricular filling times through the introduction of a premature beat or a pause, after fixed pacing periods, Starling’s filling effect on LV systolic and diastolic performance might be examined.
Another important application of the new method would be to study the relationship between abnormal electrical activation sequences and the resulting mechanical events. This would allow better understanding of the role of electromechanical coupling in LV performance and the influence of various conductive disorders on cardiac mechanics and may bring a unique value to the evaluation of the effects of pacing.
The use of pacing to alter the hemodynamic state in patients with cardiomyopathies is an emerging technique in cardiovascular medicine.32–34 Dual-chamber pacing changes the temporal relation between the atrial and ventricular contraction and the sequence of ventricular activation, and both factors probably contribute substantially to the hemodynamic benefit achieved with pacing. Such benefit has been shown to occur in patients with hypertrophic obstructive cardiomyopathy,32,33 in which RV apical pacing may reduce the dynamic pressure gradient over the LV outflow tract, probably by altering the activation sequence of the septum. Other preliminary studies have also shown significant improvement in functional class with pacing in patients with dilated cardiomyopathy.34 Recent work may suggest that the position of the pacing electrode and the resulting activation sequence also have important hemodynamic consequences, even in structurally normal hearts.35
Limitations of the New Method
In contrast to conventional volumetric measurement techniques, which sample the endocardial surface simultaneously and in a regular sequence, the acquisition of the spatial location of the endocardial sites that forms the basis for the 3D reconstruction of the dynamic geometry of the chamber by the new method is sequential and irregular. Hence, a major prerequisite of the new mapping method is that the rhythm be monomorphic and stable throughout the mapping procedure. Also, the fact that the data used in a single reconstruction are recorded over several heartbeats limits the temporal resolution of the method.
In addition, because the geometry of the reconstructed image is highly dependent on the sampled sites, failure to collect data points from the entire ventricle will result in underestimation of the true volume. Inability to collect data points from specific areas in the ventricles may result from user variabilities or the mechanical relations between the catheter and the geometry of the chamber.
By introducing filtering criteria for the acceptance of sampled points and sampling the first points with fluoroscopy, we were able to achieve accurate and reproducible measurements with different users. However, with regard to the number of points that must be acquired, speed of acquisition, and volumetric measurement accuracy, the results of the present study should be restricted to the animal model used and still must be verified clinically. For example, in coronary artery disease patients with dilated hearts or aneurysms, additional sampled points may be required to accurately determine LV volumes.
The present study has presented a new method for the 3D assessment of LV mechanical properties. The study has shown through several stages that the new method is capable of highly accurate and reproducible determination of LV volumes in both in vitro and in vivo studies.
The ability to provide accurate information regarding the 3D mechanics of the LV without the use of fluoroscopy, contrast medium, complicated image processing, or extrapolation of the volume from 2D images may provide an important clinical advantage. Furthermore, the ability to accurately combine detailed information regarding the mechanical and electrical functions of the heart in a spatially oriented fashion may provide a useful tool for both basic and clinical cardiology.
Selected Abbreviations and Acronyms
|LAT||=||local activation time|
|LV||=||left ventricular, ventricle|
|MRI||=||magnetic resonance imaging|
|RV||=||right ventricular, ventricle|
This work was supported by a grant from Biosense. We thank Ruth Singer for editing the manuscript.
Dr Ben-Haim is one of the founders of Biosense, Inc.
- Received April 28, 1997.
- Revision received July 24, 1997.
- Accepted August 2, 1997.
- Copyright © 1997 by American Heart Association
McGhie AI, Willerson JT, Corbett JR. Radionuclide assessment of ventricular function and risk stratification after myocardial infarction. Circulation. 1991;84(suppl I):I-167–I-176.
Cintron G, Johnson G, Francis G, Cobb F, Cohn JN, for the V-HeFT VA Cooperative Studies Group. Prognostic significance of serial changes in left ventricular ejection fraction in patients with congestive heart failure. Circulation. 1993;87(suppl VI):VI-17–VI-23.
Feigenbaum H. Echocardiography, 4th ed. Philadelphia, Pa: Lea & Febiger; 1986:153–155.
Gueret P, Meerbaum S, Wyatt HL, Uchiyama T, Lang TW, Corday E. Two-dimensional echocardiographic quantitation of left ventricular volumes and ejection fraction. Circulation. 1980;62:1308–1318.
Tortoledo FA, Quinones MA, Fernandez GC, Waggoner AD, Winters WL. Quantification of left ventricular volumes by two-dimensional echocardiography: a simplified and accurate approach. Circulation. 1982;67:579–584.
Nichols K, De Puey EG, Gooneratne N, Saledsky H, Friedman M, Cochoff S. First-pass ventricular ejection fraction using a single-crystal nuclear camera. J Nucl Med. 1994;35:1292–1300.
Links JM, Becker LC, Shindledecker JG, Guzman P, Burow RD, Nickoloff EL, Alderson PO, Wagner HN. Measurement of absolute left ventricular volume from gated blood pool studies. Circulation. 1982;65:82–91.
Reiter SJ, Rumberger JA, Feiring AJ, Stanford W, Marcus ML. Precision of measurements of right and left ventricular volume by cine computed tomography. Circulation. 1986;74:890–900.
Gepstein L, Hayam G, and Ben-Haim SA. A novel method for nonfluoroscopic catheter-based electroanatomical mapping of the heart: in vitro and in vivo accuracy results. Circulation. 1997;95:1611–1622.
Dodge HT, Hay RE, Sandler H. An angiocardiographic method for directly determining left ventricular stroke volume in man. Circ Res. 1962;11:739–745.
Kennedy JW, Baxley WA, Figley MM, Dodge HT, Blackmon JR. Quantitative angiocardiography, I: the normal left ventricle in man. Circulation. 1966;34:272–278.
Baur LHB, Schipperheyn JJ, van der Velde EA, van der Wall EE, Reiber JHC, van der Geest RJ, van Dijkman PRM, Gerritsen JG, van Eck-Smit BLF, Voogd PJ, Bruschke AVG. Reproducibility of left ventricular size, shape and mass with echocardiography, magnetic resonance imaging and radionuclide angiography in patients with anterior wall infarction: a plea for core laboratories. Int J Cardiac Imaging. 1996;12:233–240.
Gopal AS, Shen Z, Sapin PM, Keller AM, Schnellbaecher MJ, Leibowitz DW, Akinboboye OO, Rodney RA, Blood DK, King DL. Assessment of cardiac function by three-dimensional echocardiography compared with conventional noninvasive methods. Circulation. 1995;92:842–853.
Gopal AS, Keller AM, Rigling R, King DL Jr, King DL. Left ventricular volume and endocardial surface area by three--dimensional echocardiography: comparison with two-dimensional echocardiography and nuclear magnetic resonance imaging in normal subjects. J Am Coll Cardiol. 1993;22:258–270.
Nosir YFM, Fioretti PM, Vletter WB, Boersma E, Salustri A, Postma JT, Reijs AEM, Ten Cate FJ, Roelandt RTC. Accurate measurement of left ventricular ejection fraction by three-dimensional echocardiography: a comparison with radionuclide angiography. Circulation. 1996;94:460–466.
Apfel HD, Shen Z, Gopal AS, Vangi V, Solowiejczuk D, Altmann K, Barst RJ, Boxt LM, Allan LD, King DL. Quantitative three dimensional echocardiography in patients with pulmonary hypertension and compressed left ventricles: comparison with cross sectional echocardiography and magnetic resonance imaging. Heart. 1996;76:350–354.
Krebs W, Klues HG, Steinert S, Sivarajan M, Job FP, Flachskampf FA, Franke A, Reineke T, Hanrath P. Left ventricular volume calculations using a multiplanar transoesophageal echoprobe: in vitro validation and comparison with biplane angiography. Eur Heart J. 1996;17:1279–1288.
Sagawa K, Maughan L, Suga H, Sunagawa K. Cardiac Contraction and the Pressure-Volume Relationship. New York, NY: Oxford University Press; 1988.
Kass DA, Maughan WL. From Emax to pressure-volume relations: a broader view. Circulation. 1988;77:1203–1212.
Suga H. Ventricular energetics. Physiol Rev. 1990;70:247–277.
Sunagawa K, Maughan WL, Burkhoff D, Sagawa K. Left ventricular interaction with arterial load studied in isolated canine ventricle. Am J Physiol. 1983;215:H773–H780.
Fananapazir L, Epstein ND, Curiel RV, Panza JA, Tripodi D, McAreavey D. Long-term results of duel-chamber (DDD) pacing in obstructive hypertrophic cardiomyopathy: evidence for progressive symptomatic and hemodynamic improvement and reduction of left ventricular hypertrophy. Circulation. 1994;90:2731–2742.
Rosenqvist M, Bergfeldt L, Haga Y, Rydén J, Öwall A, Rydén L. High right septal pacing in complete heart block is possible and improves left ventricular performance. Circulation. 1993;88(suppl I):I-19. Abstract.