Author + information
- Received July 21, 2009
- Revision received September 16, 2009
- Accepted September 23, 2009
- Published online February 1, 2010.
- Manish Bansal, MD,
- Leanne Jeffriess, BS,
- Rodel Leano, BS,
- Julie Mundy, MBBS and
- Thomas H. Marwick, MBBS, PhD⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Thomas H. Marwick, University of Queensland Department of Medicine, Princess Alexandra Hospital, Ipswich Road, Brisbane, Qld 4102, Australia
Objectives Comparison of myocardial tissue-velocity imaging (TVI) and speckle-tracking echocardiography (STE) for prediction of viability at dobutamine echocardiography (DbE).
Background Use of TVI-based strain imaging during DbE may facilitate the prediction of myocardial viability but has technical limitations. STE overcomes these but requires evaluation for prediction of viability.
Methods We studied 55 patients with ischemic heart disease and left ventricular systolic dysfunction (left ventricular ejection fraction <0.45) who were undergoing DbE for assessment of myocardial viability and who subsequently underwent myocardial revascularization. TVI was used to measure longitudinal end-systolic strain (longS) and peak systolic strain rate (SR) at rest and at low-dose dobutamine (LDD). Longitudinal, radial, and circumferential strain and strain rate were measured with STE. Segmental functional recovery was defined by improved wall-motion score on side-by-side comparison of echocardiographic images before and 9 months after revascularization and areas under the receiver operator characteristic curves were used to compare methods.
Results Of the 375 segments with abnormal resting function, 154 (41%) showed functional recovery. Only circumferential resting and low-dose STE strain and low-dose longitudinal strain and SR predicted functional recovery independent of wall-motion analysis. Among different strain parameters, only TVI-based longitudinal end-systolic strain and peak systolic SR at LDD had incremental value over wall-motion analysis (areas under the receiver operator characteristic curves of 0.79, 0.79, and 0.74, respectively). STE measurements of strain and SR identified viability only in the anterior circulation, whereas TVI strain and SR accurately identified viability in both anterior and posterior circulations.
Conclusions Combination of TVI or STE methods with DbE can predict viability, with TVI strain and SR at LDD being the most accurate. TVI measures can predict viability in both anterior and posterior circulations, but STE measurements predict viability only in the anterior circulation.
The detection of myocardial viability in patients with ischemic left ventricular (LV) dysfunction has potential therapeutic and prognostic implications (1). Dobutamine echocardiography (DbE) has comparable accuracy to other approaches for the assessment of myocardial viability (2), but involves subjective assessment of wall-thickening responses to dobutamine stimulation and is highly operator-dependent (3).
Measurement of myocardial deformation with tissue-velocity imaging (TVI) during DbE has comparable accuracy to expert wall-motion analysis for prediction of functional recovery (4). However, the wider application of Doppler strain in clinical practice has been constrained by its susceptibility to signal noise and dependence on the angle of insonation (5). Angle dependency is not a problem with 2-dimensional speckle-tracking echocardiography (STE), but this technique is dependent on image quality and operates at limited frame rate (6,7). STE-based myocardial strain has been validated against other techniques, used to measure the amplitude and timing of function at rest, and applied (with more difficulty) during stress (8–10). However, its utility in the DbE assessment of myocardial viability has not been studied. In the present study, we sought to determine the relative accuracy of TVI- and STE-based measurements of myocardial strain for the detection of myocardial viability.
Fifty-five patients with ischemic heart disease and LV systolic dysfunction (ejection fraction <0.45) who were undergoing DbE for assessment of myocardial viability and who subsequently underwent myocardial revascularization were included in the study. This study involved analysis of TVI images recorded by Hanekom et al. (4) and new analysis and comparison of strain and strain rate (SR) measurements with STE. A follow-up echocardiogram was performed 9 months after revascularization to assess recovery of regional myocardial function.
Dobutamine stress echocardiography
A standard dobutamine-atropine stress protocol was performed with low-dose images at 5 and 10 μg/kg/min. Blood pressure and 12-lead electrocardiography were recorded at baseline and at the end of every stage.
Echocardiography was performed using a standard commercial echocardiography system (Vivid Seven, General Electric Medical Systems, Milwaukee, Wisconsin) with an M3S probe. Parasternal long-axis and mid-ventricular short-axis views, as well as 3 standard apical views (4-, 2-, and 3-chamber) were acquired at rest and at each stage during dobutamine stress. Gray-scale images were obtained at a frame rate of 60 to 100 frames/s using harmonic (1.7/3.4 MHz) B-mode imaging. Separate harmonic color tissue-velocity images were also acquired with a color frame rate of 100 to 185 frames/s depending on the sector width. For each view, 3 consecutive cardiac cycles were acquired during a breath hold.
Regional wall-motion analysis was performed by 2 independent observers, blinded to the patients' clinical data, using both side-by-side digital displays and review of videotape. In each of 16 LV segments (11), wall motion was scored as 1, normokinetic; 1.5, mildly hypokinetic; 2, moderate to severely hypokinetic; 3, akinetic; or 4, dyskinetic. A dysfunctional segment was considered to be viable if there was at least a 1-grade improvement at low-dose dobutamine (LDD) with or without worsening at peak dose.
Measurement of deformation parameters
Myocardial strain and SR with STE and TVI were measured at rest and at LDD using the same 16-segment model (EchoPAC-PC 6.0, General Electric Medical Systems, Milwaukee, Wisconsin). Gray-scale harmonic images were used for STE strain measurements and color tissue-velocity images were used for measurement of TVI strain. Only longitudinal strain (longS) and longitudinal (longSR) were measured with TVI, whereas STE measured longitudinal (apical views), radial, and circumferential (short-axis view) deformation. Radial strain (radS) and circumferential strain (circS) and radial strain rate (radSR) and circumferential strain rate (circSR) were measured only in the mid-ventricular segments because only 1 short-axis view was available. Only those segments that had baseline resting wall-motion abnormalities were included in the measurement.
Measurement of myocardial strain with STE (Fig. 1)
STE strain and SR were measured in 1 cardiac cycle per view. After manual tracing of the endocardial border in end systole, a region of interest was manually adjusted to include the entire myocardial thickness. The software then tracked speckles frame-by-frame throughout the entire cardiac cycle. The automated software then generated traces depicting regional strain and SR, from which end-systolic strain and peak systolic SR were recorded.
Measurement of myocardial strain with TVI
A region of interest (12 × 6 mm) was placed in the middle of each myocardial segment and tracked manually through the cardiac cycle. In each segment, traces depicting regional longS and longSR were generated over the whole cardiac cycle, and end-systolic longS and peak systolic longSR were recorded. Segments that were poorly visualized or had insonation angle >30° were excluded from the measurements.
Revascularization and follow-up
All patients underwent percutaneous or surgical revascularization, selected according to the judgment by the treating physician. Follow-up echocardiography was performed 9 months after revascularization. Segments with resting dysfunction that were adequately revascularized were evaluated for regional recovery by 2 independent observers, blinded to patients' clinical and DbE data. Segments were deemed to have recovered if the follow-up echocardiogram revealed improvement on side-by-side comparison of resting wall motion. None of the patients had a cardiac event during follow-up.
STE and TVI strain and SR were performed by 2 independent observers. To assess interobserver variability, measurements of each strain technique were repeated by an independent observer on the same echocardiographic images in 5 patients (40 myocardial segments) for longS and longSR and in 10 patients (30 myocardial segments) for radial and circumferential strain and SR.
The statistical analysis was performed using standard software (SPSS 14.0, SPSS Inc, Chicago, Illinois). Comparisons between segments showing functional recovery at follow-up and those that did not recover were performed with independent samples t test for continuous variables and chi-square test for categorical variables. Receiver-operator characteristics (ROC) curves were created to assess ability of different strain parameters to predict functional recovery; the optimal cut point was based on the Youden index (12). Comparisons between ROC curves for different strain and SR parameters were made according to the method suggested by DeLong et al. (13). Because concerns have been raised about accuracy of STE strain measurements in the regions supplied by right coronary and left circumflex arteries (posterior circulation) related to image quality (9), we reanalyzed our data by different myocardial regions. Finally, multivariable logistic regression analysis was performed using functional recovery as the dependent variable to determine the independent predictive value of wall-motion analysis and various strain and SR parameters. In order to avoid problems of collinearity, this evaluation was set up in a series of separate logistic analyses for each of the deformation parameters, including wall-motion analysis and a patient-based variable (to correct for repeated measures within each patient). All values are expressed as mean ± SD or percentages. A p value of ≤0.05 was considered statistically significant.
Clinical characteristics of the study population
Table 1 summarizes the clinical characteristics, baseline LV ejection fraction, and ongoing medical therapy in the study population.
Wall-motion scoring and functional recovery
Only segments that had resting wall-motion abnormality that were adequately revascularized were included in the analysis. Of these 375 segments, 154 (41%) showed functional recovery at the end of follow-up. At baseline DbE, 160 segments (43%) were designated as viable on the basis of the dobutamine response. Among the segments showing functional recovery, 110 were accurately predicted by wall-motion analysis, yielding a sensitivity of 71%. In contrast, wall-motion scoring correctly identified 171 of the 221 segments that did not recover after revascularization (specificity 77%, p < 0.001).
Feasibility and reproducibility of strain and SR measurements
We were able to perform longS and longSR measurements in 359 segments (96%) using TVI and in 344 segments (92%, p = 0.03) using STE. The radS, radSR, circS, and circSR measurements using STE could be performed in 97% of the available 149 segments. The interobserver variability for different strain and SR measurements with the 2 techniques is summarized in Table 2.
STE measurements and functional recovery
Table 3 summarizes various strain and SR measurements with STE in the segments with and without functional recovery. Segments showing functional recovery showed significantly higher strain in all 3 directions at rest but only longS and circS at LDD. Of SR parameters, only longSR at rest and circSR at rest and LDD showed significant difference in the 2 groups.
On ROC analysis, longS and longSR at rest and LDD, radS, and radSR at rest and circS and circSR at rest and at LDD were significant predictors of functional recovery (Figs. 2 to 5).⇓⇓⇓ Increments in any of these variables in response to dobutamine stimulation were not found to be predictors of functional recovery. Comparison of the ROC curves showed no significant difference with respect to their predictive accuracy for functional recovery.
TVI measurements and functional recovery
Both longS and longSR at rest and at LDD were significantly higher in segments that recovered after revascularization as compared with those that did not recover (Table 3). All strain and SR parameters predicted functional recovery (Figs. 2 and 3), but longS and longSR at LDD were stronger predictors of functional recovery than the corresponding parameters at rest (all p values <0.001). The increment in longS (area under the receiver-operator characteristic curve [AUC] 0.73 [95% confidence interval: 0.68 to 0.79], p < 0.001; optimal cutoff –3.0 for sensitivity and specificity 0.70) and ΔlongSR (AUC 0.76 [95% confidence interval: 0.71 to 0.81], p < 0.001; optimal cutoff –0.24 for sensitivity 0.77 and specificity 0.68) with dobutamine stimulation were also good predictors of recovery.
Comparison among wall-motion analysis and STE- and TVI-based strain measurements
TVI longS and longSR at LDD (AUC 0.79 and 0.79, respectively) were the only predictors of functional recovery close to an AUC of 0.8 (the usual threshold for a “good” test). There was no significant difference in the diagnostic accuracy between these parameters and wall-motion analysis during DbE (AUC 0.74).
Because there was a significant linear relationship between corresponding strain and SR measurements, separate multivariable analyses were performed to determine whether any of the deformation measurements had incremental value over wall-motion analysis for prediction of functional recovery. Only circumferential resting and low-dose strain and TVI-based longS and longSR at LDD emerged as independent predictors of functional recovery (Table 4).
Regional differences in the predictive accuracy of strain and SR parameters
To address concerns about the influence of image quality on the accuracy of STE strain measurements, we reanalyzed our data separately for different myocardial regions. For this purpose, basal and mid-segments of posterior, lateral, and inferior walls and basal interventricular septum were designated to the posterior circulation. On ROC curve analysis, most of the STE strain and SR parameters were found to predict functional recovery in the anterior circulation. In contrast, both longS and longSR at LDD with TVI were highly significant predictors of functional recovery in both anterior and posterior circulations (Table 5).
The salient findings of the present study are as follows: 1) longitudinal strain and SR measured at rest and at LDD using either TVI or STE are predictors of functional recovery after revascularization; 2) STE measurements of circumferential strain and SR at rest and LDD and radial strain and SR at rest also predict functional recovery; 3) only TVI based longitudinal strain and SR at LDD have incremental value over wall-motion analysis in prediction of function recovery; and 4) STE measurements identify viability in the anterior circulation, whereas TVI strain and SR accurately identify viability in both anterior and posterior circulations.
Longitudinal deformation and prediction of functional recovery
Efforts have been made to develop quantitative measures of myocardial function that could be incorporated during DbE to enhance its accuracy and reproducibility (4,14). Although no previous study has evaluated accuracy of STE measurements of longitudinal deformation for viability testing with DbE, their correlation with the transmural extent of infarct tissue has been studied (15). In 80 patients with chronic ischemic LV dysfunction, Chan et al. (15) found no difference in resting STE longS and longSR between subendocardial and transmural infarct segments. However, at LDD, both longS and longSR were significantly lower in segments with transmural compared with those with subendocardial infarcts (15). Even resting longitudinal TVI strain may be significantly impaired in transmural infarct segments (16).
In the present study, although STE longS and longSR were significant predictors of functional recovery, their accuracy was only modest and much lower than the TVI-based longS and longSR at LDD. Moreover, only TVI measurements and none of the STE measurements were found to have incremental value over wall-motion analysis. These limitations of STE at stress may relate to its dependence on gray-scale image quality or the low frame rate of this technique.
Radial and circumferential deformation and prediction of functional recovery
The LV wall is composed of multiple layers of myocardial fibers, responsible for shortening in different directions. Whereas longitudinal contraction is primarily a property of subendocardial fibers, contraction in radial and circumferential directions is brought about predominantly by fibers in the outer layers (17–19). Consequently, short-axis function is likely to be less affected in subendocardial than transmural infarction (15,20) and has been proposed as a marker of myocardial viability (21).
Unlike the previous studies, we showed that the augmentation in short-axis contractile function in response to low-dose dobutamine showed circS but not radS to remain a predictor of viability. This finding may reflect cross-fiber shortening, whereby passive radial thickening of segments with nontransmural infarction is generated by viable epicardial fibers (18). The finding that resting circS was the most accurate STE-strain measurement is congruent with circS offering the greatest contrast between subendocardial and transmural infarcts (14).
The radS and circS were not measured in the apical and basal segments, and this may explain why only circS was found to have independent predictive value. The lack of apical measurements could have been particularly important, as STE performed much better in the anterior than the posterior circulation.
This is a study of relatively small sample size using a wall-motion end point, which restricts the findings to this center's experience. The results may be different with a different patient sample with different degrees of transmural versus subendocardial scar, different revascularization outcome success, and other different confounding variables.
Strain and SR measurement with both TVI and STE are feasible during DbE and can predict recovery of regional contractile function after revascularization. However, TVI strain and SR measurements at low-dose dobutamine are more accurate for this purpose and can predict viability in both anterior and posterior circulations. In contrast, STE measurements appear to be most effective in the anterior circulation.
Although the authors' research group has received research support from General Electric Medical Systems, Horten, Norway, there was no specific support for this study. Sherif Nagueh, MD served as guest editor for this paper.
- Abbreviations and Acronyms
- area under the receiver operator characteristic curve
- circumferential strain
- circumferential strain rate
- dobutamine echocardiography
- low-dose dobutamine
- longitudinal strain
- longitudinal strain rate
- left ventricular
- radial strain
- radial strain rate
- receiver-operator characteristic
- strain rate
- speckle-tracking echocardiography
- tissue-velocity imaging
- Received July 21, 2009.
- Revision received September 16, 2009.
- Accepted September 23, 2009.
- American College of Cardiology Foundation
- Allman K.C.,
- Shaw L.J.,
- Hachamovitch R.,
- Udelson J.E.
- Bax J.J.,
- Wijns W.,
- Cornel J.H.,
- Visser F.C.,
- Boersma E.,
- Fioretti P.M.
- Picano E.,
- Lattanzi F.,
- Orlandini A.,
- Marini C.,
- L'Abbate A.
- Hanekom L.,
- Jenkins C.,
- Jeffries L.,
- et al.
- Urheim S.,
- Edvardsen T.,
- Torp H.,
- Angelsen B.,
- Smiseth O.A.
- Langeland S.,
- D'Hooge J.,
- Wouters P.F.,
- et al.
- Becker M.,
- Bilke E.,
- Kuhl H.,
- et al.
- Hanekom L.,
- Cho G.Y.,
- Leano R.,
- Jeffriess L.,
- Marwick T.H.
- Armstrong W.F.,
- Pellikka P.A.,
- Ryan T.,
- Crouse L.,
- Zoghbi W.A.
- Hoffmann R.,
- Altiok E.,
- Nowak B.,
- et al.
- Chan J.,
- Hanekom L.,
- Wong C.,
- Leano R.,
- Cho G.Y.,
- Marwick T.H.
- Zhang Y.,
- Chan A.K.,
- Yu C.M.,
- et al.
- Rademakers F.E.,
- Rogers W.J.,
- Guier W.H.,
- et al.
- Clark N.R.,
- Reichek N.,
- Bergey P.,
- et al.
- Becker M.,
- Hoffmann R.,
- Kuhl H.P.,
- et al.
- Becker M.,
- Lenzen A.,
- Ocklenburg C.,
- et al.