Skip Navigation



European Journal of Echocardiography Advance Access published online on September 1, 2008
European Journal of Echocardiography, doi:10.1093/ejechocard/jen209
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
10/4/485    most recent
jen209v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Veronesi, F.
Right arrow Articles by Mor-Avi, V.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Veronesi, F.
Right arrow Articles by Mor-Avi, V.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

Published on behalf of the European Society of Cardiology. All rights reserved. © The Author 2008. For permissions please email: journals.permissions@oxfordjournals.org

Semi-automated analysis of dynamic changes in myocardial contrast from real-time three-dimensional echocardiographic images as a basis for volumetric quantification of myocardial perfusion

Federico Veronesi1,2,3, Enrico G. Caiani2, Eran Toledo3, Cristiana Corsi1, Keith A. Collins3, Georgeanne Lammertin3, Claudio Lamberti1, Roberto M. Lang3 and Victor Mor-Avi3,*

1 Department of Electronics, Computer Science and Systems, Università di Bologna, Bologna, Italy;
2 Department of Biomedical Engineering, Politecnico di Milano, Milan, Italy
3 Department of Medicine, Section of Cardiology, University of Chicago Medical Center, Chicago, Illinois, USA

Received 7 April 2008; accepted after revision 14 July 2008.

* Corresponding author. Tel: +1 773 702 1842; fax: +1 773 702 1034. E-mail address: vmoravi{at}medicine.bsd.uchicago.edu


   Abstract
Top
 Abstract
Introduction
 Methods
 Results
 Discussion
 Funding
 References
 
Aims: Despite the potential of real-time three-dimensional (3D) echocardiography (RT3DE) to assess myocardial perfusion, there is no quantification method available for perfusion analysis from RT3DE images. Such method would require 3D regions of interest (ROI) to be defined and adjusted frame-by-frame to compensate for cardiac translation and deformation. Our aims were to develop and test a technique for automated identification of 3D myocardial ROI suitable for translation-free quantification of myocardial videointensity over time, MVI(t), from contrast-enhanced RT3DE images.

Methods and results: Twelve transthoracic RT3DE (Philips) data sets obtained in pigs during transition from no contrast to steady-state enhancement (Definity) were analysed using custom software. Analysis included: (i) semi-automated detection of left ventricular endo- and epicardial surfaces using level-set techniques in one frame to define a 3D myocardial ROI, (ii) rigid 3D registration to reduce translation and rotation, (iii) elastic 3D registration to compensate for deformation, and (iv) quantification of MVI(t) in the 3D ROI from the registered and non-registered data sets to assess the effectiveness of registration. For each MVI(t) curve we computed % variability during steady-state enhancement (100 x SD/mean) and goodness of fit (r2) to the indicator dilution equation MVI(t) = A[1–exp(–βt)]. Analysis of myocardial contrast throughout contrast inflow was feasible in all data sets. Three-dimensional registration improved MVI(t) curves in terms of both % variability (2.8 ± 1.8 to 1.5 ± 0.9%; P < 0.05) and goodness of fit (r2 from 0.79 ± 0.2 to 0.90 ± 0.1; P < 0.05).

Conclusion: This is the first study to describe a new technique for semi-automated volumetric quantification of myocardial contrast from RT3DE images that includes registration and thus provides the basis for 3D measurement of myocardial perfusion.

Keywords: Real-time 3D echocardiography; Myocardial perfusion; Image registration


   Introduction
Top
 Abstract
 Introduction
Methods
 Results
 Discussion
 Funding
 References
 
Although the use of contrast echocardiography to assess myocardial perfusion has been investigated for over a decade, the reliance on partial information contained in specific cross-sectional planes posed by the two-dimensional (2D) nature of ultrasound imaging has been one of the stumbling blocks on the way of this methodology to wider acceptance. Despite the potential of contrast-enhanced real-time three-dimensional (3D) echocardiography (RT3DE) to assess myocardial perfusion more accurately,1,2 there are no tools for quantitative 3D analysis of perfusion, and only initial data are available on volumetric assessment of perfusion.3–8 One of the major difficulties in designing such volumetric analysis tools is that they require the ability to define 3D regions of interest (ROI), in which myocardial videointensity would be measured over time [MVI(t)] during dynamic changes in myocardial contrast. Moreover, even after 3D ROIs are defined, they need to be adjusted frame-by-frame to compensate for cardiac translation and deformation, adding another layer of complexity to the problem. Such adjustments are crucial, since ROIs fixed in space are highly unlikely to maintain a fixed anatomical position, even when images are acquired at the same phase of the cardiac cycle using ECG (electrocardiogram) triggering. Thus, spatially fixed ROIs inevitably result in noisy MVI(t) curves, which limit the reproducibility of quantitative perfusion analysis.

We have recently described a technique for translation-free perfusion analysis from 2D contrast-enhanced images based on frame-by-frame automated detection of left ventricular (LV) endocardial boundary, which eliminated the need for manual tracing and frame-by-frame tracking of myocardial ROIs.9 We hypothesized that a similar approach could be used for frame-by-frame semi-automated definition of 3D ROIs, which in combination with 3D image registration would allow translation-free volumetric analysis of myocardial contrast over time.

Our goal was to develop such volumetric analysis technique and test its applicability to RT3DE images. We used the 3D level-set techniques for detection of the endo- and epicardial boundaries and optical flow analysis for frame-by-3D image registration of the detected myocardium. This technique was tested on transthoracic contrast-enhanced RT3DE data sets obtained in pigs by measuring MVI(t) during transition from no contrast to steady-state contrast enhancement. This transient contrast inflow (TCI) manoeuvre was used as an alternative to high-energy ultrasound pulses that are not applicable with RT3DE imaging.8,10,11 This approach allowed us to study the effects of image registration on the stability of myocardial contrast during steady-state enhancement and on the goodness of fit of the transition phase to the indicator dilution equation commonly used to model myocardial contrast replenishment. These two parameters are essential for the ability to obtain low noise MVI(t) curves, a prerequisite for reliable volumetric analysis of myocardial perfusion from RT3DE images.


   Methods
Top
 Abstract
 Introduction
 Methods
Results
 Discussion
 Funding
 References
 
Animal preparation
Experiments were performed in five male farm pigs (20–28 kg). Animals were pre-treated with telazol (2.2 mg/kg, im) and atropine sulphate (0.05 mg/kg, im). Indomethacine (100 mg, per OGT) was given to suppress allergic reactions. Following intubation, pigs were mechanically ventilated (Drager) and anaesthetized with isoflurane (0.5–2.5% mixed with oxygen). ECG, blood pressure and expiratory gases were monitored (Datex, Cardiocap). At the end of experiment, a lethal dose of pentobarbital sodium (120 mg/kg, iv) was given.

Ultrasound imaging
RT3DE imaging was performed using a SONOS 7500 system (Philips) equipped with a matrix-array transducer (X4) in the harmonic mode (1.6 MHz transmit frequency). The spatial aperture was set to be the widest available (58° x 29°) for triggered acquisition of series of end-systolic pyramidal data sets. Imaging was performed from the left parasternal approach, as a way to reduce attenuation in a densely opacified right ventricular cavity. Volumetric short-axis data sets containing the mid-portion of the LV (Figure 1A) with the right ventricular cavity mostly excluded from the scan volume were obtained from this transducer position. Contrast enhancement was achieved by intravenous infusion of Definity, the US equivalent of Luminity (Bristol-Myers Squibb, 1.3 mL in 25 mL saline at 150–260 mL/h). Infusion rate was determined by maximizing myocardial contrast without visible attenuation. To minimize bubble destruction, the minimal mechanical index necessary to visualize myocardial contrast was used (0.4–0.8). Figure 2 shows a schematic representation of the TCI manoeuvre, which included: (i) optimization of contrast infusion rate and imaging settings during steady-state enhancement, (ii) infusion interruption to allow contrast clearance, which was expedited by continuous 2D imaging at high mechanical index; (iii) resumption of contrast infusion, resulting in contrast inflow. Image acquisition started ~5 s prior to the resumption of infusion to capture the entire transition from non-contrast to reinstated steady-state enhancement. Images were saved digitally for off-line analysis.


Figure 1
View larger version (83K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 1 Schematic representation of volumetric parasternal short-axis imaging of the heart (A). Actual real-time three-dimensional echocardiography data set that contains the mid-portion of the left ventricle, obtained using transthoracic imaging in a pig (B) shown with en-face short-axis views of a mid-papillary three-dimensional slice obtained before the onset of myocardial contrast inflow (C), during (D) and after (E) reinstatement of steady-state contrast enhancement.

 


Figure 2
View larger version (17K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 2 Schematic representation of the transient contrast inflow sequence. The lower portion of the diagram shows the time line of imaging steps that comprise the sequence. The upper portion shows the timing of infusion interruption and resumption (arrows) and their effects on the myocardial contrast with the level of enhancement shown schematically in the shaded band above the time axis.

 
Image analysis
Images were analysed using custom software implemented in Matlab software (MathWorks, Natick, MA, USA). Initially, semi-automated endocardial surface detection based on the level-set approach,12,13 as described previously,14 was performed on a fully contrast-enhanced reference frame selected from the steady-state phase. This technique uses an implicit representation of curves as a partial differential equation to track boundaries, without geometrical assumptions or a priori shape knowledge.12 Four LV short-axis planes were selected from the 3D data set, and a small number of endocardial boundary points (8–12) were manually initialized in each plane. Papillary muscles were included within the LV cavity, when visible. The selected points were connected to define a set of polygons. For each polygon, a signed distance function was calculated,15 and a rough surface corresponding to the endocardium was computed using linear interpolation of the signed distance functions. This surface was then used as the initial condition for the level-set partial differential equation, which guided surface evolution within the volumetric data set towards the endocardium under the constraints of two forces: (i) an interface tension force that depends on the curvature of the evolving surface and has a regulating effect, and (ii) a force that attracts this surface towards the image boundaries. When the two forces balance each other, the evolution reaches a steady state and the resultant surface was used to represent the endocardium. Thereafter, this procedure was repeated to detect the LV epicardial surface,16 and the shell confined between the two surfaces was used as the myocardial ROI.

Three-dimensional registration procedure included two consecutive steps. First, rigid transformation of the heart was performed by shifting and rotating each frame throughout the image sequence for optimal match between the position and spatial orientation of the LV in the current and the reference frames. Matching was achieved by maximizing weighted cross-correlation of the pyramidal data sets representing these two frames, while excluding the LV cavity from analysis in order that the changes in contrast intensity not to affect the registration. The goal of this step was to compensate for translation and rotation of the heart relative to the reference frame that predominantly occurs because of respiratory motion. After the data set was shifted and rotated, it was subjected to an elastic transformation. First, using an algorithm based on optical flow techniques,17 displacement field representing point-wise motion between the two images was computed. Then using this field, the current frame was warped by forcing the endo- and epicardial LV boundaries into their position in the reference frame.

Once the myocardial ROI was defined and 3D registration completed, mean MVI was calculated by averaging pixel intensity in each consecutive frame throughout the sequence to generate MVI(t) curves. This procedure was also performed with the original non-registered image sequence to assess the effects of registration on MVI(t) curves. For each curve, we assumed that myocardial contrast inflow followed the indicator dilution equation: MVI(t) = A[1–e–βt] + C, where C is the initial MVI before contrast inflow, A the maximum contrast-induced increase in MVI, and β the characteristic constant related to tissue blood flow. For both the registered and non-registered MVI(t) curves, non-linear fitting to the indicator dilution equation was performed and the goodness of fit was estimated in terms of r2. In addition, the post-replenishment steady-state portion of the curve was used to estimate signal variability (in % of its mean value) as 100 x SD/mean, with and without registration.

Statistical analysis
To assess the effectiveness of 3D registration, goodness of fit and % variability were averaged for the registered, and separately for the non-registered data sets. The differences were tested for significance using paired t-test with Bonferroni corrections for repeated measures to take into account that more than one data set were obtained in the same animals.


   Results
Top
 Abstract
 Introduction
 Methods
 Results
Discussion
 Funding
 References
 
Twelve pyramidal data sets were analysed. Figure 1B shows an example of a data set that contains the mid-portion of the LV, and en-face short-axis views of a mid-papillary 3D slice obtained at different phases of the TCI sequence are shown in Figure 1C–E. Semi-automated detection of LV endo- and epicardial surfaces (Figure 3A and B) was feasible in all data sets and resulted in identification of the myocardial ROI (Figure 3C) in every one of the 12 data sets. Analysis of one data set required ~10 min, including initialization of the endo- and epicardial boundaries in multiple planes, calculation of the 3D endo- and epicardial surfaces and surface adjustments when necessary, and calculation of MVI(t) throughout the TCI image sequence.


Figure 3
View larger version (62K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 3 Example of the detected left ventricular endocardial (cyan) and epicardial (blue) surfaces shown in two different orientations (A and B) superimposed on the original real-time three-dimensional echocardiography data set. The automatically defined myocardial region of interest is shown in (C).

 
The 3D registration required additional time, which ranged between 15 and 25 min depending on the number of heart beats included in the TCI image sequence (up to 50 frames). Figure 4 shows an example of cross-sectional short-axis views extracted from non-registered and corresponding registered pyramidal data sets obtained at different phases of the TCI. Although the effects of the elastic transformation can be visualized as altered image features, this figure demonstrates the effects of 3D registration since at every phase of the sequence the registered LV fits well into the endo- and epicardial boundaries obtained from the reference frame. Figure 5 shows the resultant MVI(t) curves obtained from the same data set with and without 3D registration, which demonstrate the marked decrease in the level of noise during the phase of post-replenishment steady-state contrast enhancement noted with the registration. The decrease in MVI(t) signal variability (2.8 ± 1.8 to 1.5 ± 0.9%; P < 0.05) and the increase in the goodness of fit to the indicator dilution inflow model (r2 from 0.79 ± 0.2 to 0.90 ± 0.1; P < 0.05) in the 12 data sets are summarized in Figure 6. Importantly, improvement in both parameters was statistically significant, thus proving the effectiveness of the 3D registration algorithm.


Figure 4
View larger version (59K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 4 Example of cross-sectional short-axis views extracted from non-registered (top) and corresponding registered (bottom) pyramidal data sets obtained at different phases of the transient contrast inflow at which left ventricular (LV) cavity appeared shifted relative to the reference frame. The red and blue contours represent the cross-section of the LV endocardial and epicardial surfaces, respectively. Note the shift between the endo- and epicardial contours and the LV myocardium on the non-registered frames (top), which is eliminated by three-dimensional registration (bottom), despite the dynamics of contrast enhancement.

 


Figure 5
View larger version (18K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 5 Example of myocardial videointensity over time, MVI(t), curves measured during transient contrast enhancement in one animal.

 


Figure 6
View larger version (22K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 6 Summary of myocardial videointensity over time [MVI(t)] signal variability and goodness of fit (r2) of the MVI(t) curves to the indicator dilution inflow model in the 12 data sets (*P < 0.05; t-test with Bonferroni corrections for repeated measures of data sets obtained in the same animals).

 

   Discussion
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
Funding
 References
 
Despite the abundance of previous studies demonstrating the potential use of contrast echocardiography to quantify myocardial perfusion, this approach has not become part of the daily clinical routine in the context of diagnosis and evaluation of ischaemic heart disease. It has been recognized that the 2D nature of conventional contrast-enhanced echocardiographic imaging has been one of the limiting factors in its ability to provide accurate information on the extent and severity of perfusion abnormalities. This is because planar images cannot accurately reflect the complexity of individual coronary anatomy and account for the wide variability in its relationship with myocardial perfusion territories. Hence the immense appeal for 3D imaging in the context of myocardial perfusion, which received a major boost from the development of RT3DE technology. This latter technology eliminated the need for multiplane acquisitions3,4 with repeated contrast manoeuvres necessary for perfusion quantification by analysis of beat-by-beat changes in myocardial contrast, such as contrast inflow or washout following bolus injections,18–23 or alternatively its replenishment after destructive high-energy ultrasound pulses delivered during contrast infusion.24–28

Still, the quantification of tissue blood flow in a certain area of the myocardium, either as absolute values in mL/min/g or as quantitative indices calculated from contrast intensity time curves, requires the definition of a myocardial ROI where dynamic changes in contrast intensity can then be analysed. Traditionally, quantitative 2D analysis of myocardial perfusion has been based on manual tracing of ROIs in a single imaging plane and frame-by-frame realignment of these ROIs to compensate for cardiac translation. Although this methodology is subjective and time-consuming, it is relatively straightforward and easy to implement in software. In contrast, drawing ROIs in 3D space is significantly more complex, both conceptually and from the point of view of software implementation. Needless to say, realignment of 3D ROIs throughout the image sequence to compensate for cardiac translation and deformation is anything but straightforward.

To overcome these limitations, we initially developed a technique for automated identification of myocardial ROI for fast, translation-free analysis of myocardial contrast enhancement from 2D images.9 Our approach was based on automated detection of the endocardial boundaries, which are relatively easy to detect in the presence of contrast in the LV cavity, followed by outward expansion into the myocardium and segmentation (Figure 4). We tested this approach on contrast-enhanced 2D images obtained in pigs undergoing coronary occlusions, and found that it allowed automated translation-free quantification of regional myocardial perfusion, without the need for ROI tracing.

The present study was designed to test the feasibility of the 3D equivalent of this approach, which is essential for the advancement of volumetric quantification of myocardial perfusion. There are two major methodological differences in the approach we used in the 3D implementation. First, instead of detecting the endocardium and expanding it outwards as we did with 2D images, we used a similar algorithm to detect the epicardium,16 which allowed us defining the myocardial ROI more accurately irrespective of variations in myocardial thickness. Secondly, while the 2D technique inherently minimized motion artefacts by independently detecting the endocardial boundary on each consecutive frame and manually correcting it as necessary, implementing this approach in 3D space for both endo- and epicardial surfaces turned out to be impractical and time-consuming, since it required manual corrections of both surfaces in multiple frames. This difficulty led us to seek a more efficient alternative solution in the form of automated frame-by-frame image registration designed to fit the LV in every frame into its reference position. This approach was implemented in two consecutive steps: the rigid registration that addressed the problem of translation and rotation, followed by the elastic registration that modified the shape of the ventricle to resemble the reference frame when necessary. Once the myocardium was forced into its reference position, the 3D myocardial ROI detected in the reference frame was essentially fixed, allowing the generation of translation-free MVI(t) curves throughout the image sequence.

To test the effectiveness of this approach as a potential solution for motion artefacts in 3D space, we compared MVI(t) curves obtained from the same RT3DE data sets with and without registration and calculated for each curve two parameters that are key for robust volumetric quantification of myocardial perfusion: (i) goodness of fit of the MVI(t) curve to the indicator dilution contrast replenishment model, which describes the exponential behaviour during the contrast inflow phase; and (ii) the noise level in the steady-state post-replenishment portion of the MVI(t) curve, that evaluates the stability of data during the ‘quiescent’ period during which no changes in contrast are expected. These two parameters reflect the quality of information on blood flow dynamics and are crucial for reliable analysis.11 Importantly, both parameters were significantly improved by registration, proving the effectiveness of this approach and supporting its applicability to series of ECG-triggered contrast-enhanced RT3DE images obtained during the TCI manoeuvre, despite the frame-by-frame changes in both contrast characteristics and LV shape.

Although the small number of contrast intensity curves used in this study could be potentially viewed as a limitation, our results were statistically significant despite the limited size of the data sample. Accordingly, we could not justify sacrificing additional animals to further increase the sample. In addition, we could be criticized for not including in our study design an independent reference technique for myocardial perfusion, such as microspheres. However, this study was designed to test the new algorithm for 3D image registration throughout the cardiac cycle, rather than to quantify myocardial perfusion. Such quantification presents a significant challenge of its own and remains to be validated against an appropriate reference technique in the future.

One limitation of our analysis technique is the subjective choice of a reference frame, which affects the registration process and may thus affect the measured MVI(t) curves. In our experience, selecting a reference frame during the post-replenishment steady-state enhancement period provided more stable curves than those obtained when using a non-enhanced reference frame at the beginning of the image sequence. This is likely due to the fact that endocardial blood-tissue interface is better visualized with contrast enhancement.

One important limitation of the current RT3DE technology is the limited spatial aperture-angle, which does not allow imaging the entire ventricle in real-time. This limitation mostly stems from insufficient computational resources and does not represent a fundamental constraint of this technology. Therefore, while anticipating a technological solution to this issue, we focused on testing the feasibility of our approach for 3D image registration in a partial volume of the heart that can be scanned using the current RT3DE equipment.

In summary, this is the first study to test the feasibility of a new semi-automated technique that allows true 3D quantification of myocardial contrast including effective registration of contrast-enhanced RT3DE images, and thus may facilitate volumetric quantification of myocardial perfusion. This methodology or its further refined derivatives may prove a valuable addition to the non-invasive diagnostic cardiac imaging arsenal and become part of the routine cardiology practice.


   Funding
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 Funding
References
 
This study was funded by the Echo Investigator's Award from the American Society of Echocardiography (V. Mor-Avi).

Conflict of interest: none declared.


   References
Top
 Abstract
 Introduction
 Methods
 Results
 Discussion
 Funding
 References
 

  1. Lang RM, Mor-Avi V, Sugeng L, Nieman PS, Sahn DJ. Three-dimensional echocardiography: the benefits of the additional dimension. J Am Coll Cardiol (2006) 48:2053–69.[Abstract/Free Full Text]
  2. Mor-Avi V, Lang RM. Three-dimensional echocardiographic evaluation of myocardial perfusion. Cardiol Clin (2007) 25:273–82.[CrossRef][Web of Science][Medline]
  3. Bae RY, Belohlavek M, Tanabe K, Greenleaf JF, Seward JB. Rapid three-dimensional myocardial contrast echocardiography: volumetric quantitation of nonperfused myocardium after intravenous contrast administration. Echocardiography (1999) 16:357–65.[CrossRef][Web of Science][Medline]
  4. Yao J, De CS, Delabays A, Masani N, Udelson JE, Pandian NG. Bulls-eye display and quantitation of myocardial perfusion defects using three-dimensional contrast echocardiography. Echocardiography (2001) 18:581–8.[CrossRef][Web of Science][Medline]
  5. Camarano G, Jones M, Freidlin RZ, Panza JA. Quantitative assessment of left ventricular perfusion defects using real-time three-dimensional myocardial contrast echocardiography. J Am Soc Echocardiogr (2002) 15:206–13.[CrossRef][Web of Science][Medline]
  6. Chen LX, Wang XF, Nanda NC, Miller AP, Xie MX, Zhuang L, et al. Real-time three-dimensional myocardial contrast echocardiography in assessment of myocardial perfusion defects. Chin Med J (Engl) (2004) 117:337–41.[Medline]
  7. Pemberton J, Li X, Hickey E, Karamlou T, Sandquist CA, Ungerleider R, et al. Live real-time three-dimensional echocardiography for the visualization of myocardial perfusion–a pilot study in open-chest pigs. J Am Soc Echocardiogr (2005) 18:956–8.[CrossRef][Web of Science][Medline]
  8. Toledo E, Lang RM, Collins KA, Lammertin G, Williams U, Weinert L, et al. Imaging and quantification of myocardial perfusion using real-time three-dimensional echocardiography. J Am Coll Cardiol (2006) 47:146–54.[Abstract/Free Full Text]
  9. Caiani EG, Lang RM, Caslini S, Collins KA, Korcarz CE, Mor-Avi V. Quantification of regional myocardial perfusion using semiautomated translation-free analysis of contrast-enhanced power modulation images. J Am Soc Echocardiogr (2003) 16:116–23.[CrossRef][Web of Science][Medline]
  10. Toledo E, Lang RM, Collins KA, Lammertin G, Weinert L, Mor-Avi V. Quantitative echocardiographic evaluation of myocardial perfusion using interrupted contrast infusion technique: in vivo validation studies and feasibility in human beings. J Am Soc Echocardiogr (2005) 18:1304–11.[CrossRef][Web of Science][Medline]
  11. Toledo E, Collins KA, Williams U, Lammertin G, Bolotin G, Raman J, et al. Interrupted infusion of echocardiographic contrast as a basis for accurate measurement of myocardial perfusion: ex vivo validation and analysis procedures. J Am Soc Echocardiogr (2005) 18:1312–20.[CrossRef][Web of Science][Medline]
  12. Malladi R, Sethian JA, Vemuri BC. Shape modeling with front propagation - a level set approach. IEEE Trans on Pattern Anal Machine Intelligence (1995) 17:158–75.[CrossRef]
  13. Sethian JA. Level Set Methods and Fast Marching Methods (1999) Cambridge: Cambridge University Press.
  14. Corsi C, Lang RM, Veronesi F, Weinert L, Caiani EG, MacEneaney P, et al. Volumetric quantification of global and regional left ventricular function from real-time three-dimensional echocardiographic images. Circulation (2005) 112:1161–70.[Abstract/Free Full Text]
  15. Corsi C, Saracino G, Sarti A, Lamberti C. Left ventricular volume estimation for real-time three-dimensional echocardiography. IEEE Trans Med Imaging (2002) 21:1202–8.[CrossRef][Web of Science][Medline]
  16. Caiani EG, Corsi C, Sugeng L, MacEneaney P, Weinert L, Mor-Avi V, et al. Improved quantification of left ventricular mass based on endocardial and epicardial surface detection with real time three dimensional echocardiography. Heart (2006) 92:213–9.[Abstract/Free Full Text]
  17. Veronesi F, Corsi C, Caiani EG, Sarti A, Lamberti C. Tracking of left ventricular long axis from real-time three-dimensional echocardiography using optical flow techniques. IEEE Trans Inf Technol Biomed (2006) 10:174–81.[CrossRef][Web of Science][Medline]
  18. Vandenberg BF, Kieso R, Fox-Eastham K, Chilian W, Kerber RE. Quantitation of myocardial perfusion by contrast echocardiography: analysis of contrast gray level appearance variables and intracyclic variability. J Am Coll Cardiol (1989) 13:200–6.[Abstract]
  19. Kaul S, Kelly P, Oliner JD, Glasheen WP, Keller MW, Watson DD. Assessment of regional myocardial blood flow with myocardial contrast two-dimensional echocardiography. J Am Coll Cardiol (1989) 13:468–82.[Abstract]
  20. Keller MW, Spotnitz WD, Matthew TL, Glasheen WP, Watson DD, Kaul S. Intraoperative assessment of regional myocardial perfusion using quantitative myocardial contrast echocardiography: an experimental evaluation. J Am Coll Cardiol (1990) 16:1267–79.[Abstract]
  21. Mor-Avi V, David D, Akselrod S, Bitton Y, Choshniak I. Myocardial regional blood flow: quantitative measurement by computer analysis of contrast enhanced echocardiographic images. Ultrasound Med Biol (1993) 19:619–33.[CrossRef][Web of Science][Medline]
  22. Skyba DM, Jayaweera AR, Goodman NC, Ismail S, Camarano G, Kaul S. Quantification of myocardial perfusion with myocardial contrast echocardiography during left atrial injection of contrast. Implications for venous injection. Circulation (1994) 90:1513–21.[Abstract/Free Full Text]
  23. Leistad E, Ohmori K, Peterson TA, Christensen G, DeMaria AN. Quantitative assessment of myocardial perfusion during graded coronary artery stenoses by intravenous myocardial contrast echocardiography. J Am Coll Cardiol (2001) 37:624–31.[Abstract/Free Full Text]
  24. Pelberg RA, Wei K, Kamiyama N, Sklenar J, Bin J, Kaul S. Potential advantage of flash echocardiography for digital subtraction of B-mode images acquired during myocardial contrast echocardiography. J Am Soc Echocardiogr (1999) 12:85–93.[CrossRef][Web of Science][Medline]
  25. Mor-Avi V, Caiani EG, Collins KA, Korcarz CE, Bednarz JE, Lang RM. Combined assessment of myocardial perfusion and regional left ventricular function by analysis of contrast-enhanced power modulation images. Circulation (2001) 104:352–7.[Abstract/Free Full Text]
  26. Wei K. Assessment of myocardial blood flow and volume using myocardial contrast echocardiography. Echocardiography (2002) 19:409–16.[CrossRef][Web of Science][Medline]
  27. Dijkmans PA, Knaapen P, Sieswerda GT, Aiazian E, Visser CA, Lammertsma AA, et al. Quantification of myocardial perfusion using intravenous myocardial contrast echocardiography in healthy volunteers: comparison with positron emission tomography. J Am Soc Echocardiogr (2006) 19:285–93.[CrossRef][Web of Science][Medline]
  28. Jeetley P, Hickman M, Kamp O, Lang RM, Thomas JD, Vannan MA, et al. Myocardial contrast echocardiography for the detection of coronary artery stenosis: a prospective multicentre study in comparison with single-photon emission computed tomography. J Am Coll Cardiol (2006) 47:141–5.[Abstract/Free Full Text]

Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 Eur J EchocardiogrHome page
L. P. Badano
Contrast enhanced real-time three-dimensional echocardiography for quantification of myocardial perfusion: a step forward
Eur J Echocardiogr, June 1, 2009; 10(4): 465 - 466.
[Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
10/4/485    most recent
jen209v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Veronesi, F.
Right arrow Articles by Mor-Avi, V.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Veronesi, F.
Right arrow Articles by Mor-Avi, V.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?