- Email: [email protected]
Automated border detection on contrast enhanced echocardiographic images
International Journal of Cardiology 103 (2005) 164 – 167 www.elsevier.com/locate/ijcard
Automated border detection on contrast enhanced echocardiographic images James N. Kirkpatrick, Roberto M. Lang, Savitri E. Fedson, Allen S. Anderson, James Bednarz, Kirk T. SpencerT University of Chicago Hospitals and Clinics, 5841 S. Maryland Ave. MC 5084, Chicago, IL 60637, United States Received 31 March 2004; received in revised form 26 July 2004; accepted 7 August 2004 Available online 19 February 2005
Abstract Background: Accurate determination of left ventricular ejection fraction (LV EF) is of paramount importance in the evaluation of patients with cardiovascular disease. Quantitative techniques for the automated calculation of EF exist however, the robustness of these techniques is dependent on adequate endocardial border definition and therefore are difficult to use in patients with limited images. We sought to combine the endocardial border enhancing effects of contrast echocardiography with an automated border detection technique to provide quantitative and accurate determination of LV EF. Methods: Thirty-nine consecutive patients referred to nuclear cardiology for EF determination underwent radionuclide angiography followed by echocardiographic imaging using prototype software that allowed automated border detection during contrast infusion. Results: Adequate LV cavity opacification with contrast was possible in 38/39 patients. The mean radionuclide EF was 50F16% (range 19– 73). There was no statistically significant difference between the mean nuclear EF and averaged echocardiographically determined EF (51F18%). The mean bias was 0.6 with limits of agreement that were +15 and 14. Conclusion: This study demonstrated that prototype software successfully tracked the contrast enhanced endocardial border allowing accurate calculation of LV EF. D 2005 Elsevier Ireland Ltd. All rights reserved. Keywords: Ejection fraction; Border detection; Transthoracic echocardiography; Contrast agents
1. Introduction Accurate determination of left ventricular (LV) function is of paramount importance in the evaluation of patients with cardiovascular disease. Transthoracic echocardiography is the most commonly used modality for the assessment of ejection fraction (EF), but the applicability of echocardiography for the evaluation of EF is often limited by image quality and the use of qualitative rather than quantitative techniques. The advent of tissue harmonic and contrast
T Corresponding author. E-mail address: [email protected] (K.T. Spencer). 0167-5273/$ - see front matter D 2005 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ijcard.2004.08.056
echocardiographic imaging have allowed definition of the endocardial border in nearly every patient presenting to the echocardiography laboratory [1–3]. However, use of harmonic contrast enhanced images still requires manual tracing of systolic and diastolic endocardial borders to determine EF. Quantitative techniques for the automated detection of the endocardium exist and have proven useful for the evaluation of LV systolic and diastolic performance [4,5]. However, the robustness of these techniques is dependent on adequate endocardial border definition, similar to hand tracing, and therefore cannot be used in patients with limited endocardial delineation. We sought to combine the endocardial border enhancing effects of contrast echocardiography with an automated border detection technique
J.N. Kirkpatrick et al. / International Journal of Cardiology 103 (2005) 164–167
165
generated (Fig. 1). LV end-diastolic (EDV) and end-systolic (ESV) volumes were computed as the maximum and minimal points on the LV volume waveform. EF was calculated independently for both apical views as (EDV ESV)/EDV. The mean values for EF for the two techniques were compared using a paired t-test. A Bland–Altman analysis was also performed with the radionuclide EF as the standard for comparison. Statistical significance was considered Pb0.05. 40 30
Fig. 1. Contrast enhanced apical four-chamber view of the LV with prototype border detection activated.
20
to provide quantitative and accurate determination of LV EF.
0
10
-10 -20
2. Materials and methods Thirty-nine consecutive patients referred to nuclear cardiology for EF determination were enrolled. First pass radionuclide angiocardiography was performed using 25–30 mCi of Tc-99m tetrofosmin. Images were obtained using a SIM-400 optically divided single crystal camera in the anterior projection at 25 (F4) frames per cardiac cycle. Data were analyzed using the frame method for LV ejection fraction by commercially available software. The dualregion of interest (ROI) LV EF was determined as the enddiastolic ROI counts minus the end-systolic ROI counts, divided by the background subtracted end-diastolic ROI counts. The subjects then underwent echocardiographic imaging within 1 h of their radionuclide test. Imaging was performed in harmonic mode with an S-4 transducer on a Sonos 5500 (Philips Medical, Andover, MA). Contrast enhancement was achieved by using an intravenous infusion of 0.2–0.3 ml/min of Definity (Dupont Pharmaceuticals, Wilmington, DE). The contrast infusion was adjusted to provide adequate apical filling without cavity attenuation. Imaging was performed using a low mechanical index (0.3–0.4) to minimize bubble destruction, systolic fading and apical swirling. Echocardiographic imaging was performed sequentially from the apical four-chamber (AP4) and two-chamber (AP2) views, taking care not to foreshorten the LV cavity. Prototype acoustic quantification (AQ) software (Philips Medical, Andover, MA), modified for use during contrast enhancement, was activated and optimized by adjusting the overall gain, time gain compensation and lateral gain controls to ensure adequate endocardial tracking. The LV volume signals were continuously acquired during quiet respiration and a single, averaged waveform automatically
-30 -40 10
20
30
40
50
60
70
80
60
70
80
60
70
80
Nuclear EF (%) 40 30 20 10 0 -10 -20 -30 -40 10
20
30
40
50
Nuclear EF (%) 40 30 20 10 0 -10 -20 -30 -40 10
20
30
40
50
Nuclear EF (%)
Fig. 2. Bland–Altman graphs comparing LV EF determined from contrast enhanced automated border detection algorithm in the AP4 and AP2 views as well the average of AP4 and AP2.
166
J.N. Kirkpatrick et al. / International Journal of Cardiology 103 (2005) 164–167
3. Results Adequate LV cavity opacification with contrast was possible in 38/39 patients. After optimizing imaging settings, adequate endocardial border tracking was visually confirmed in all subjects in whom cavity opacification was sufficient. Mean patient age was 55F14 years old (31 males/18 females). The mean radionuclide EF was 50F16% (range 19–73). There was no statistically significant difference between the mean nuclear EF and the AP4, AP2 or averaged echocardiographically determined EF (49F17, 51F18 and 51F18, respectively). Bland–Altman graphs are shown in Fig. 2. The AP4 AQ EF slightly underestimated (bias 0.8) and AP2 AQ EF slightly overestimated (bias 1.3) the true EF. The mean bias was lowest at 0.6 for the averaged AQ EF with limits of agreement that were +15 and 14. There were several individual EF differences that were somewhat large (N15%). These disparities seemed to occur when the nuclear EF was in the higher range (N50%), whereas the agreement at lower EFs was generally better. There was no significant correlation between EF difference and subject body surface area.
threshold as tissue and those below this threshold as LV cavity. After contrast enhancement there is a significant increase in the backscatter of the LV cavity. Using the traditional algorithm, the LV cavity would be incorrectly identified as tissue thereby preventing endocardial border identification. The modified algorithm used in this study breversesQ the thresholding, identifying the blood pool as pixels that exceed a predefined backscatter threshold. With adequate cavity opacification, the endocardial-blood pool border tracking is facilitated. Although widely used and without significant side effect, a limitation of echocardiographic contrast agents is their cost. However, these costs (under $100) are substantially less than that of radionuclide angiography, MRI or CT estimation of ejection fraction. In addition this methodology requires expertise in transthoracic imaging to avoid foreshortening or misaligning the LV image plane, which leads to significant underestimation of LV EF. This study demonstrated that this prototype software successfully tracked the contrast enhanced endocardial border. When this border was used to calculate LV EF there was good agreement with the gold standard radionuclide angiography.
4. Discussion Two-dimensional echocardiography is the most common method for assessing LV EF, as it is noninvasive, accurate and reproducible. Quantification of EF is typically performed by hand-tracing endocardial borders. Recent advances in ultrasound technology, such as tissue harmonic imaging and echocardiographic contrast agents, have greatly increased the number of patients in whom endocardial borders are sufficiently defined to permit hand tracing for quantification of LV EF [6]. Use of tissue harmonic and contrast enhanced echocardiographic imaging has also been shown to improve the accuracy of ejection fraction determination in sub-optimal images [7–9]. Automated identification of endocardial borders avoids tedious manual tracing of still frame echocardiographic images. These automated techniques are accurate when compared to other modalities for the determination of LV EF, and utility of these methods for the evaluation of LV systolic and diastolic function has been established [4,5,10–14]. However, automated border detection methods are also dependent on adequate endocardial definition to work properly. Tissue harmonic imaging has previously been shown to improve automated endocardial border tracking [1,15]. Despite the use of harmonic imaging, some patients still have inadequate endocardial definition and require cavity opacification with echocardiographic contrast. This study demonstrates that automated border detection can be used during echocardiographic contrast injection. The automated border detection technique used in previous studies identifies pixels above a certain backscatter
Acknowledgements Disclosures: Supported in part by a Grant from Dupont Pharmaceuticals.
References [1] Spencer KT, Bednarz J, Rafter PG, Korcarz C, Lang RM. Use of harmonic imaging without echocardiographic contrast to improve two-dimensional image quality. Am J Cardiol 1998;82:794 – 9. [2] Cohen JL, Cheirif J, Segar DS, Gillam LD, Gottdiener JS, Hausnerova E, et al. Improved left ventricular endocardial border delineation and opacification with OPTISON (FS069), a new echocardiographic contrast agent—results of a phase III multicenter trial. J Am Coll Cardiol 1998;32:746 – 52. [3] Grayburn PA, Weiss JL, Hack TC, Klodas E, Raichlen JS, Vannan MA, et al. Phase III multicenter trial comparing the efficacy of 2% dodecafluoropentane emulsion (EchoGen) and sonicated 5% human albumin (Albunex) as ultrasound contrast agents in patients with suboptimal echocardiograms. J Am Coll Cardiol 1998;32:230 – 6. [4] Spencer KT, Mor-Avi V, Weinert L, Lang RM. Evaluation of left ventricular diastolic performance using automated border detection. Echocardiogr—A J Cardiovasc Ultrasound and Allied Tech 1999;16: 51 – 62. [5] Perez JE. Online quantification of left ventricular function: correlation with various imaging modalities. Echocardiogr—A J Cardiovasc Ultrasound and Allied Tech 1999;16:43 – 9. [6] Spencer KT, Bednarz J, Mor-Avi V, Weinert L, Tan J, Godoy I, et al. The role of echocardiographic harmonic imaging and contrast enhancement for improvement of endocardial border delineation. J Am Soc Echocardiogr 2000;13:131 – 8. [7] Van Camp G, Franken R, Schoors D, Hagers Y, Koole M, Demoor D, et al. Impact of second harmonic imaging on the determination of the global and regional left ventricular function by 2D echocardiography:
J.N. Kirkpatrick et al. / International Journal of Cardiology 103 (2005) 164–167
[8]
[9]
[10]
[11]
a comparison with MIBI gated SPECT. Eur J Echocardiog 2000;1:122 – 9. Nahar T, Croft L, Shapiro R, Fruchtman S, Diamond J, Henzlova M, et al. Comparison of four echocardiographic techniques for measuring left ventricular ejection fraction. Am J Cardiol 2000;86:1358 – 62. Hundley WG, Kizilbash AM, Afridi I, Franco F, Peshock RM, Grayburn PA. Administration of an intravenous perfluorocarbon contrast agent improves echocardiographic determination of left ventricular volumes and ejection fraction: comparison with cine magnetic resonance imaging. J Am Coll Cardiol 1998;32:1426 – 32. Chandra S, Bahl VK, Reddy CB, Bhargava B, Malhotra A, Wasir HS. Comparison of echocardiographic acoustic quantification system and radionuclide ventriculography for estimating left ventricular ejection fraction: validation in patients without regional wall motion abnormalities. Am Heart J 1997;133:359 – 63. Hardy BG, Rice MJ, McDonald RW. Comparison of echocardiographic acoustic quantification with off-line manual tracing for determining left ventricular volume and ejection fraction in pediatric patients. Am J Cardiol 1998;82:1428–4.
167
[12] Lucariello RJ, Sun Y, Doganay G, Chiaramida SA. Sensitivity and specificity of left ventricular ejection fraction by echocardiographic automated border detection: comparison with radionuclide ventriculography. Clin Cardiol 1997;20:943 – 8. [13] Gorcsan III J, Lazar JM, Schulman DS, Follansbee WP. Comparison of left ventricular function by echocardiographic automated border detection and by radionuclide ejection fraction. Am J Cardiol 1993;72:810 – 5. [14] Lindower PD, Rath L, Preslar J, Burns TL, Rezai K, Vandenberg BF. Quantification of left ventricular function with an automated border detection system and comparison with radionuclide ventriculography. Am J Cardiol 1994;73:195 – 9. [15] Tsujita-Kuroda K, Zhang GC, Sumita Y, Hirooka K, Hanatani A, Nakatani S, et al. Validity and reproducibility of echocardiographic measurement of left ventricular ejection fraction by acoustic quantification with tissue harmonic imaging technique. J Am Soc Echocardiogr 2000;13:300 – 5.
Automated border detection on contrast enhanced echocardiographic images James N. Kirkpatrick, Roberto M. Lang, Savitri E. Fedson, Allen S. Anderson, James Bednarz, Kirk T. SpencerT University of Chicago Hospitals and Clinics, 5841 S. Maryland Ave. MC 5084, Chicago, IL 60637, United States Received 31 March 2004; received in revised form 26 July 2004; accepted 7 August 2004 Available online 19 February 2005
Abstract Background: Accurate determination of left ventricular ejection fraction (LV EF) is of paramount importance in the evaluation of patients with cardiovascular disease. Quantitative techniques for the automated calculation of EF exist however, the robustness of these techniques is dependent on adequate endocardial border definition and therefore are difficult to use in patients with limited images. We sought to combine the endocardial border enhancing effects of contrast echocardiography with an automated border detection technique to provide quantitative and accurate determination of LV EF. Methods: Thirty-nine consecutive patients referred to nuclear cardiology for EF determination underwent radionuclide angiography followed by echocardiographic imaging using prototype software that allowed automated border detection during contrast infusion. Results: Adequate LV cavity opacification with contrast was possible in 38/39 patients. The mean radionuclide EF was 50F16% (range 19– 73). There was no statistically significant difference between the mean nuclear EF and averaged echocardiographically determined EF (51F18%). The mean bias was 0.6 with limits of agreement that were +15 and 14. Conclusion: This study demonstrated that prototype software successfully tracked the contrast enhanced endocardial border allowing accurate calculation of LV EF. D 2005 Elsevier Ireland Ltd. All rights reserved. Keywords: Ejection fraction; Border detection; Transthoracic echocardiography; Contrast agents
1. Introduction Accurate determination of left ventricular (LV) function is of paramount importance in the evaluation of patients with cardiovascular disease. Transthoracic echocardiography is the most commonly used modality for the assessment of ejection fraction (EF), but the applicability of echocardiography for the evaluation of EF is often limited by image quality and the use of qualitative rather than quantitative techniques. The advent of tissue harmonic and contrast
T Corresponding author. E-mail address: [email protected] (K.T. Spencer). 0167-5273/$ - see front matter D 2005 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.ijcard.2004.08.056
echocardiographic imaging have allowed definition of the endocardial border in nearly every patient presenting to the echocardiography laboratory [1–3]. However, use of harmonic contrast enhanced images still requires manual tracing of systolic and diastolic endocardial borders to determine EF. Quantitative techniques for the automated detection of the endocardium exist and have proven useful for the evaluation of LV systolic and diastolic performance [4,5]. However, the robustness of these techniques is dependent on adequate endocardial border definition, similar to hand tracing, and therefore cannot be used in patients with limited endocardial delineation. We sought to combine the endocardial border enhancing effects of contrast echocardiography with an automated border detection technique
J.N. Kirkpatrick et al. / International Journal of Cardiology 103 (2005) 164–167
165
generated (Fig. 1). LV end-diastolic (EDV) and end-systolic (ESV) volumes were computed as the maximum and minimal points on the LV volume waveform. EF was calculated independently for both apical views as (EDV ESV)/EDV. The mean values for EF for the two techniques were compared using a paired t-test. A Bland–Altman analysis was also performed with the radionuclide EF as the standard for comparison. Statistical significance was considered Pb0.05. 40 30
Fig. 1. Contrast enhanced apical four-chamber view of the LV with prototype border detection activated.
20
to provide quantitative and accurate determination of LV EF.
0
10
-10 -20
2. Materials and methods Thirty-nine consecutive patients referred to nuclear cardiology for EF determination were enrolled. First pass radionuclide angiocardiography was performed using 25–30 mCi of Tc-99m tetrofosmin. Images were obtained using a SIM-400 optically divided single crystal camera in the anterior projection at 25 (F4) frames per cardiac cycle. Data were analyzed using the frame method for LV ejection fraction by commercially available software. The dualregion of interest (ROI) LV EF was determined as the enddiastolic ROI counts minus the end-systolic ROI counts, divided by the background subtracted end-diastolic ROI counts. The subjects then underwent echocardiographic imaging within 1 h of their radionuclide test. Imaging was performed in harmonic mode with an S-4 transducer on a Sonos 5500 (Philips Medical, Andover, MA). Contrast enhancement was achieved by using an intravenous infusion of 0.2–0.3 ml/min of Definity (Dupont Pharmaceuticals, Wilmington, DE). The contrast infusion was adjusted to provide adequate apical filling without cavity attenuation. Imaging was performed using a low mechanical index (0.3–0.4) to minimize bubble destruction, systolic fading and apical swirling. Echocardiographic imaging was performed sequentially from the apical four-chamber (AP4) and two-chamber (AP2) views, taking care not to foreshorten the LV cavity. Prototype acoustic quantification (AQ) software (Philips Medical, Andover, MA), modified for use during contrast enhancement, was activated and optimized by adjusting the overall gain, time gain compensation and lateral gain controls to ensure adequate endocardial tracking. The LV volume signals were continuously acquired during quiet respiration and a single, averaged waveform automatically
-30 -40 10
20
30
40
50
60
70
80
60
70
80
60
70
80
Nuclear EF (%) 40 30 20 10 0 -10 -20 -30 -40 10
20
30
40
50
Nuclear EF (%) 40 30 20 10 0 -10 -20 -30 -40 10
20
30
40
50
Nuclear EF (%)
Fig. 2. Bland–Altman graphs comparing LV EF determined from contrast enhanced automated border detection algorithm in the AP4 and AP2 views as well the average of AP4 and AP2.
166
J.N. Kirkpatrick et al. / International Journal of Cardiology 103 (2005) 164–167
3. Results Adequate LV cavity opacification with contrast was possible in 38/39 patients. After optimizing imaging settings, adequate endocardial border tracking was visually confirmed in all subjects in whom cavity opacification was sufficient. Mean patient age was 55F14 years old (31 males/18 females). The mean radionuclide EF was 50F16% (range 19–73). There was no statistically significant difference between the mean nuclear EF and the AP4, AP2 or averaged echocardiographically determined EF (49F17, 51F18 and 51F18, respectively). Bland–Altman graphs are shown in Fig. 2. The AP4 AQ EF slightly underestimated (bias 0.8) and AP2 AQ EF slightly overestimated (bias 1.3) the true EF. The mean bias was lowest at 0.6 for the averaged AQ EF with limits of agreement that were +15 and 14. There were several individual EF differences that were somewhat large (N15%). These disparities seemed to occur when the nuclear EF was in the higher range (N50%), whereas the agreement at lower EFs was generally better. There was no significant correlation between EF difference and subject body surface area.
threshold as tissue and those below this threshold as LV cavity. After contrast enhancement there is a significant increase in the backscatter of the LV cavity. Using the traditional algorithm, the LV cavity would be incorrectly identified as tissue thereby preventing endocardial border identification. The modified algorithm used in this study breversesQ the thresholding, identifying the blood pool as pixels that exceed a predefined backscatter threshold. With adequate cavity opacification, the endocardial-blood pool border tracking is facilitated. Although widely used and without significant side effect, a limitation of echocardiographic contrast agents is their cost. However, these costs (under $100) are substantially less than that of radionuclide angiography, MRI or CT estimation of ejection fraction. In addition this methodology requires expertise in transthoracic imaging to avoid foreshortening or misaligning the LV image plane, which leads to significant underestimation of LV EF. This study demonstrated that this prototype software successfully tracked the contrast enhanced endocardial border. When this border was used to calculate LV EF there was good agreement with the gold standard radionuclide angiography.
4. Discussion Two-dimensional echocardiography is the most common method for assessing LV EF, as it is noninvasive, accurate and reproducible. Quantification of EF is typically performed by hand-tracing endocardial borders. Recent advances in ultrasound technology, such as tissue harmonic imaging and echocardiographic contrast agents, have greatly increased the number of patients in whom endocardial borders are sufficiently defined to permit hand tracing for quantification of LV EF [6]. Use of tissue harmonic and contrast enhanced echocardiographic imaging has also been shown to improve the accuracy of ejection fraction determination in sub-optimal images [7–9]. Automated identification of endocardial borders avoids tedious manual tracing of still frame echocardiographic images. These automated techniques are accurate when compared to other modalities for the determination of LV EF, and utility of these methods for the evaluation of LV systolic and diastolic function has been established [4,5,10–14]. However, automated border detection methods are also dependent on adequate endocardial definition to work properly. Tissue harmonic imaging has previously been shown to improve automated endocardial border tracking [1,15]. Despite the use of harmonic imaging, some patients still have inadequate endocardial definition and require cavity opacification with echocardiographic contrast. This study demonstrates that automated border detection can be used during echocardiographic contrast injection. The automated border detection technique used in previous studies identifies pixels above a certain backscatter
Acknowledgements Disclosures: Supported in part by a Grant from Dupont Pharmaceuticals.
References [1] Spencer KT, Bednarz J, Rafter PG, Korcarz C, Lang RM. Use of harmonic imaging without echocardiographic contrast to improve two-dimensional image quality. Am J Cardiol 1998;82:794 – 9. [2] Cohen JL, Cheirif J, Segar DS, Gillam LD, Gottdiener JS, Hausnerova E, et al. Improved left ventricular endocardial border delineation and opacification with OPTISON (FS069), a new echocardiographic contrast agent—results of a phase III multicenter trial. J Am Coll Cardiol 1998;32:746 – 52. [3] Grayburn PA, Weiss JL, Hack TC, Klodas E, Raichlen JS, Vannan MA, et al. Phase III multicenter trial comparing the efficacy of 2% dodecafluoropentane emulsion (EchoGen) and sonicated 5% human albumin (Albunex) as ultrasound contrast agents in patients with suboptimal echocardiograms. J Am Coll Cardiol 1998;32:230 – 6. [4] Spencer KT, Mor-Avi V, Weinert L, Lang RM. Evaluation of left ventricular diastolic performance using automated border detection. Echocardiogr—A J Cardiovasc Ultrasound and Allied Tech 1999;16: 51 – 62. [5] Perez JE. Online quantification of left ventricular function: correlation with various imaging modalities. Echocardiogr—A J Cardiovasc Ultrasound and Allied Tech 1999;16:43 – 9. [6] Spencer KT, Bednarz J, Mor-Avi V, Weinert L, Tan J, Godoy I, et al. The role of echocardiographic harmonic imaging and contrast enhancement for improvement of endocardial border delineation. J Am Soc Echocardiogr 2000;13:131 – 8. [7] Van Camp G, Franken R, Schoors D, Hagers Y, Koole M, Demoor D, et al. Impact of second harmonic imaging on the determination of the global and regional left ventricular function by 2D echocardiography:
J.N. Kirkpatrick et al. / International Journal of Cardiology 103 (2005) 164–167
[8]
[9]
[10]
[11]
a comparison with MIBI gated SPECT. Eur J Echocardiog 2000;1:122 – 9. Nahar T, Croft L, Shapiro R, Fruchtman S, Diamond J, Henzlova M, et al. Comparison of four echocardiographic techniques for measuring left ventricular ejection fraction. Am J Cardiol 2000;86:1358 – 62. Hundley WG, Kizilbash AM, Afridi I, Franco F, Peshock RM, Grayburn PA. Administration of an intravenous perfluorocarbon contrast agent improves echocardiographic determination of left ventricular volumes and ejection fraction: comparison with cine magnetic resonance imaging. J Am Coll Cardiol 1998;32:1426 – 32. Chandra S, Bahl VK, Reddy CB, Bhargava B, Malhotra A, Wasir HS. Comparison of echocardiographic acoustic quantification system and radionuclide ventriculography for estimating left ventricular ejection fraction: validation in patients without regional wall motion abnormalities. Am Heart J 1997;133:359 – 63. Hardy BG, Rice MJ, McDonald RW. Comparison of echocardiographic acoustic quantification with off-line manual tracing for determining left ventricular volume and ejection fraction in pediatric patients. Am J Cardiol 1998;82:1428–4.
167
[12] Lucariello RJ, Sun Y, Doganay G, Chiaramida SA. Sensitivity and specificity of left ventricular ejection fraction by echocardiographic automated border detection: comparison with radionuclide ventriculography. Clin Cardiol 1997;20:943 – 8. [13] Gorcsan III J, Lazar JM, Schulman DS, Follansbee WP. Comparison of left ventricular function by echocardiographic automated border detection and by radionuclide ejection fraction. Am J Cardiol 1993;72:810 – 5. [14] Lindower PD, Rath L, Preslar J, Burns TL, Rezai K, Vandenberg BF. Quantification of left ventricular function with an automated border detection system and comparison with radionuclide ventriculography. Am J Cardiol 1994;73:195 – 9. [15] Tsujita-Kuroda K, Zhang GC, Sumita Y, Hirooka K, Hanatani A, Nakatani S, et al. Validity and reproducibility of echocardiographic measurement of left ventricular ejection fraction by acoustic quantification with tissue harmonic imaging technique. J Am Soc Echocardiogr 2000;13:300 – 5.