High-frequency speckle tracking echocardiography in the assessment of left ventricular function and remodeling after murine myocardial infarction.

Am J Physiol Heart Circ Physiol 306: H1371–H1383, 2014. First published February 15, 2014; doi:10.1152/ajpheart.00553.2013.

Innovative Methodology

High-frequency speckle tracking echocardiography in the assessment of left ventricular function and remodeling after murine myocardial infarction Amit Bhan,1* Alexander Sirker,1* Juqian Zhang,1* Andrea Protti,1,2 Norman Catibog,1 William Driver,1 Rene Botnar,2 Mark J. Monaghan,1 and Ajay M. Shah1 1

Cardiovascular Division, King’s College London British Heart Foundation Centre, King’s College London, London, United Kingdom; and 2Imaging Sciences Division, King’s College London British Heart Foundation Centre, King’s College London, London, United Kingdom Submitted 22 July 2013; accepted in final form 11 February 2014

Bhan A, Sirker A, Zhang J, Protti A, Catibog N, Driver W, Botnar R, Monaghan MJ, Shah AM. High-frequency speckle tracking echocardiography in the assessment of left ventricular function and remodeling after murine myocardial infarction. Am J Physiol Heart Circ Physiol 306: H1371–H1383, 2014. First published February 15, 2014; doi:10.1152/ajpheart.00553.2013.—The objectives of this study were to assess the feasibility and accuracy of high-frequency speckle tracking echocardiography (STE) in a murine model of myocardial infarction (MI). STE is used clinically to quantify global and regional cardiac function, but its application in mice is challenging because of the small cardiac size and rapid heart rates. A high-frequency microultrasound system with STE (Visualsonics Vevo 2100) was compared against magnetic resonance imaging (MRI) for the assessment of global left ventricular (LV) size and function after murine MI. Animals subjected to coronary ligation (n ⫽ 46) or sham ligation (n ⫽ 27) were studied 4 wk postoperatively. Regional and global deformation were also assessed. STE-derived LV ejection fraction (EF) and mass correlated well with MRI indexes (r ⫽ 0.93, 0.77, respectively; P ⬍ 0.001), as did STE-derived mass with postmortem values (r ⫽ 0.80, P ⬍ 0.001). Higher STE-derived volumes correlated positively with MRI-derived infarct size (P ⬍ 0.01). Global strain parameters were significantly reduced after MI (all P ⬍ 0.001) and strongly correlated with LV mass and MRI-derived infarct size as promising surrogates for the extent of remodeling and infarction, respectively (both P ⬍ 0.05). Regional strain analyses showed that radial strain and strain rate were relatively preserved in anterior basal segments after MI compared with more apical segments (P ⬍ 0.001); however, longitudinal strain and strain rate were significantly impaired both basally and distally (P ⬍ 0.001). Strain-derived parameters of dyssynchrony were significantly increased in the MI group (P ⬍ 0.01). Analysis time for STE was 210 ⫾ 45 s with acceptable inter- and intraobserver variability. In conclusion, high-frequency STE enables quantitative assessment of regional and global function in the remodeling murine LV after MI. mouse; myocardial infarction; strain; echocardiography; left ventricular function COMPARED WITH CLINICAL IMAGING, echocardiography in mouse models remains relatively challenging, primarily because of the exceptional spatial and temporal resolution required to image such a small, rapidly beating heart. Research studies involving gene-modified mouse models commonly use clinical echocardiography systems (with transducer frequencies of up to 15 MHz) and tend to rely on rather simple measures of LV size and function, for example from M-mode tracings. These

* A. Bhan, A. Sirker, and J. Zhang contributed equally to this study. Address for reprint requests and other correspondence: A. M. Shah, Cardiovascular Division, The James Black Centre, King’s College London, 125 Coldharbour Ln., London, UK SE5 9NU (e-mail: [email protected]). http://www.ajpheart.org

measures provide relatively crude estimates of global LV size and function, especially in remodeled hearts (e.g., after MI) where the geometric assumptions of symmetry are likely to be inaccurate. In clinical practice, such measures have largely been superseded, and the last decade has seen the introduction of several advanced echocardiography techniques to improve quantification of LV structure and function (5, 15, 18, 30). Myocardial deformation (strain) imaging is one of the more robust of such techniques, having been extensively investigated clinically (4, 5, 15). It offers enhanced quantification of global as well as regional function, both of which are important in assessment of the remodeling LV. Deformation information can be derived either from color-coded tissue Doppler imaging or from 2-dimensional (2D) STE. STE uses patterns resulting from acoustic interference phenomena within the myocardium to track myocardial motion throughout the cardiac cycle (8, 17). The main advantage of STE over the Doppler-based technique is its lack of dependence on the ultrasound insonation angle (10). This may be of particular benefit in small animal imaging where the ability to align myocardial structures of interest with the transducer can be challenging. Recent advances in high-frequency microultrasound have led to the commercial availability of a dedicated small animal ultrasound system equipped with a high-frequency transducer (Vevo 2100, Visualsonics, Toronto, Canada) that allows quantification of global and regional LV function with speckle tracking-derived measurements. We evaluated the utility of STE-based assessment of global and regional LV function in mice undergoing LV remodeling after left coronary ligation or sham surgery. The study had three specific aims: 1) to compare data for LV volumes, ejection fraction, and mass (which are automatically obtained during STE analysis) against the reference standard of MRI; 2) to assess parameters of global deformation and their relationship to infarct size and LV mass (as a measure of extent of remodeling); and 3) to assess regional deformation and LV dyssynchrony after MI. MATERIALS AND METHODS

Surgery and general protocol. In vivo procedures were conducted in accordance with the Guidance on the Operation of the Animals (Scientific Procedures) Act, 1986 (UK Home Office), and with the approval of the local independent institutional review committee. A total of 73 adult female C57Bl6 mice (weight 18 –24 g) were studied, either after left coronary ligation (n ⫽ 46) or sham surgery (n ⫽ 27). Animals were anesthetized with 2–2.5% isoflurane/97.5–98% oxygen, endotracheally intubated, and ventilated using a small animal ventilator (Hugo Sacks Elektronic, Germany). Myocardial infarction was created, in a similar manner to that previously described by our group (13). However, a wider range of infarct sizes was produced in the

0363-6135/14 Copyright © 2014 the American Physiological Society

H1371

Innovative Methodology H1372

SPECKLE TRACKING ECHOCARDIOGRAPHY IN MICE

present study, by ligating the left coronary artery at varying levels below the left atrial lower border; this permitted a broad range of subsequent left ventricular volumes and dysfunction to develop and be comparatively evaluated using echocardiography vs. MRI. Sham surgery involved an identical procedure except for coronary ligation. Animals were administered buprenorphine and flunixin analgesia and allowed to recover in a warmed chamber. Four weeks (⫾ 2 days) after surgery, all mice underwent transthoracic echocardiography and magnetic resonance imaging (MRI), the two modalities being performed within 24 h of each other. After animals were killed, hearts were

excised, the LV was dissected, and LV mass was measured using a fine balance. Echocardiography. Transthoracic echocardiography was performed using the Vevo 2100 ultrasound system (Visualsonics, Toronto, Canada) equipped with a high-frequency (30 MHz) linear array transducer. Animals were placed supine on an electrical heating pad at 37°C under light isoflurane anesthesia (usual maintenance level 1.5% isoflurane/98.5% oxygen). Continual ECG monitoring was obtained via limb electrodes. Heart rates were ⬎450 beats/min for the duration of the study (with adjustment of level of anesthesia as necessary). Any

Fig. 1. Representative images of speckle tracking echocardiography (STE)-based analyses of ventricular function. A: an example of the semiautomated endocardial and epicardial border detection in a parasternal long-axis view. B: segmental radial and longitudinal strain curves from a parasternal long-axis view in a mouse after myocardial infarction.

AJP-Heart Circ Physiol • doi:10.1152/ajpheart.00553.2013 • www.ajpheart.org

Innovative Methodology SPECKLE TRACKING ECHOCARDIOGRAPHY IN MICE

images obtained with suboptimal physiological parameters were excluded from analysis. A 2D echocardiographic study was performed predominantly in parasternal long-axis except where noted otherwise. Careful attention was paid to image depth, width, and gain settings in order to optimize image quality. Frame rates were consistently above 150 Hz with a mean of 201 Hz (⫾24). All views were digitally stored in cine loops consisting of 300 frames. Subsequent analysis was performed off-line on a workstation installed with Vevo 2100 software (version 1.4.0) by two experienced cardiologists blinded to the surgery type. Analyses of LV volumes and ejection fraction (EF) were performed using both the standard 2D quantification software and the dedicated STE software. For standard 2D analyses, the user marks out the endocardial border in end-diastole and end-systole. The software places an imag-

H1373

inary trapezoid around the LV, then calculates the difference between the traced border and the trapezoid using a method of disks technique. This is subtracted from the known volume of the trapezoid to obtain LV volumes. To calculate mass using the standard technique, borders were traced along the endocardium and epicardium and mass was calculated from these using the area-length method (33). Analyses in the STE software utilize semiautomated border tracking. First, the user places a number of tracking points on the endocardial and epicardial borders. These are used as a guide for border delineation and subsequent frame-by-frame tracking throughout the cardiac cycle (Fig. 1A). From these borders and their motion, LV volumes are calculated using a method of disks technique (25), and ejection fraction is subsequently obtained. Manual adjustments can be made as required in order to optimize border tracking. During STE

Fig. 2. Comparison of end-diastolic volume (EDV), end-systolic volume (ESV). and ejection fraction (EF) assessed by STE vs. MRI. A, C, E: linear correlations of the volumetric parameters assessed by STE vs. MRI. The 95% confidence band of the regression lines is enclosed by dotted lines. B, D, F: Bland-Altman analyses for each parameter. Dotted lines show the best estimate of bias as well as its limits of agreement (LoA). AJP-Heart Circ Physiol • doi:10.1152/ajpheart.00553.2013 • www.ajpheart.org



analysis, the LV is automatically divided into six segments. For the parasternal long-axis view, this means two basal, two mid, and two apical segments. Strain curves are made available for each analyzed segment (Fig. 1B), without the requirement for any additional user input. LV mass is calculated from the difference between the endocardial and epicardial borders at end diastole, and the myocardial density constant. Data were obtained for LV volumes, EF, LV mass, peak longitudinal and radial segmental strain (S) and strain rate (SR), and the time to peak strain (time from the onset of the QRS complex to peak strain value). The standard deviation of time to peak strain, corrected for the RR interval, was used as a measure of intraventricular dyssynchrony (34). Magnetic resonance imaging. MRI was performed in a 7-T horizontal scanner (Varian, Palo Alto, CA) with a gradient coil inner diameter of 12 cm, 1000 mT/m (100 G/cm) strength, and rise-time of

120 ␮s. A quadrature transmit/receive coil (RAPID Biomedical) with an internal diameter of 39 mm was used. Anesthesia and physiological parameters were maintained as for echocardiography. Eleven shortaxis slices were acquired from apex to base of the left ventricle with an in-plane resolution of 0.195 ⫻ 0.195 mm and a slice thickness of 1 mm. MRI acquisitions were performed over a period of ⬃8 min per animal and this was comparable to that for echocardiography. Off-line analyses of LV volumes, EF, and mass from MRI were performed using a semiautomated software program developed in-house (22). This calculated the left ventricular end-diastolic volume (LVEDV) and left ventricular end-systolic volume (LVESV) after estimating end-diastolic and end-systolic blood pool volumes for each MRI slice. The software also identified epicardial borders from which LV mass could be calculated as the end-diastolic LV myocardial volume multiplied by an MRI density factor (1.055 g/cm2). For the subgroup

Fig. 3. Comparison of end-diastolic volume (EDV), end-systolic volume (ESV), and ejection fraction (EF) assessed by standard 2D echocardiography vs. MRI. A, C, E: linear correlations of the volumetric parameters assessed by STE vs. MRI. The 95% confidence band of the regression lines is enclosed by dotted lines. B, D, F: Bland-Altman analyses for each parameter. Dotted lines show the best estimate of bias as well as its limits of agreement (LoA). AJP-Heart Circ Physiol • doi:10.1152/ajpheart.00553.2013 • www.ajpheart.org


of mice in which infarct size was estimated, this was done from the 11 consecutive short-axis views from base to apex. The infarct size was calculated as the percentage midline length of thinned myocardium against the total LV midline length (21, 22). Thinned myocardium was defined as a reduction in thickness by ⬎60% compared with healthy myocardium, and the motionless properties of the tissue in MRI images during the cardiac cycle. Our group and others have previously demonstrated good agreement with histologically determined infarct size using triphenyltetrazolium chloride (TTC) staining (22). Intra- and interobserver variability. Blinded repeat analyses of echocardiographic data from 20 randomly chosen studies were performed at an interval of several weeks by the original observer and a different observer, in order to obtain inter- and intraobserver variability (n ⫽ 8 and n ⫽ 12, respectively). To assess test-retest variability, in seven animals we undertook two echocardiograms 24 h apart.

H1375

Statistics. Data are presented as means ⫾ standard error of the mean (SE). A Student’s t-test was performed for comparison of continuous variables between the groups. P ⬍ 0.05 was considered statistically significant. Correlations were assessed using linear regression and correlation coefficients, and agreements between measures were shown using Bland-Altman plots. Variability is reported as the mean (⫾ SD) of the absolute differences between paired measurements and by the correlation coefficient (r). All statistical analyses were performed using Graphpad Prism v4 (Graphpad Software). RESULTS

LV volumes, mass, and global function. A comparison of STE-derived LV end-diastolic volume (EDV), end-systolic volume (ESV), and ejection fraction (EF) against MRI was performed on the entire cohort of 73 mice. This demonstrated

Fig. 4. Comparison of STE-derived vs. MRI-derived and necropsy LV mass (LVM). The dotted lines on the regression graphs (A, C, E) showed the 95% confidence band of the regression. Dotted lines on the Bland-Altman plots (B, D, F) show the best estimate of bias and the limits of agreement (LoA).




a good correlation between STE and MRI for all parameters, with r values between 0.86 and 0.93 (all P ⬍ 0.0001 for all) (Fig. 2). There was, however, a significant underestimation of all three parameters by STE with biases (average differences between MRI and STE) of 7.82 ␮l, 3.75 ␮l, and 4.78% for LVEDV, LVESV, and EF, respectively. Heart rates were not significantly different between the MI and sham groups (495 ⫾ 30 and 479 ⫾ 44 beats/min, P ⫽ 0.39). The mean EF using STE-derived data was 54.0 ⫾ 1.3% in the sham group vs. 22.5 ⫾ 1.4% in the MI group (P ⬍ 0.0001). We also compared standard 2D-derived parameters against MRI in the same cohort (Fig. 3). 2D-derived EDV, ESV, and EF also correlated quite well with MRI-derived data but with slightly lower correlation coefficients (0.84 to 0.89) than for STE-derived measures. Bland-Altman plots demonstrated larger biases with 2D measures than STE-derived measures, and these were less systematic: e.g., positive bias for EDV and negative bias for ESV. In a subgroup of 14 animals (7 sham, 7 MI), STE- and MRI-derived LV mass estimation were compared against each other and against the necropsy mass (Fig. 4). STE-derived mass correlated well with MRI (r ⫽ 0.77, P ⫽ 0.0007) and necropsy mass (r ⫽ 0.80, P ⫽ 0.0003), while MRI-derived mass also correlated with necropsy measures (r ⫽ 0.74, P ⫽ 0.0037). Mass by STE was slightly underestimated compared with necropsy mass (estimate of bias ⫽ 9.17 mg), while MRI-derived results were closer in value to necropsy mass (estimate of bias ⫽ ⫺1.00 mg). Using STE-derived data, LV mass was 56.5 ⫾ 3.0 mg in the sham group vs. 97.9 ⫾ 4.7 mg in the infarct group (P ⬍ 0.0001).

In a subset of 20 MI animals, we estimated in vivo infarct size from consecutive short-axis MRI images as described in MATERIALS AND METHODS and then assessed the correlation between infarct size and STE-derived volumetric parameters. There was a significant linear correlation between MRI infarct size and LV volumetric parameters obtained by STE in this group; i.e. EDV (r ⫽ 0.75, P ⫽ 0.0005), ESV (r ⫽ 0.82, P ⬍ 0.0001), and EF (r ⫽ ⫺0.81, P ⬍ 0.0001) (Fig. 5). Infarct size correlated with necropsy mass (as a parameter of total LV remodeling, i.e., amount of hypertrophy in remote myocardium), as would be expected (Fig. 5). Taken together, these results suggest that STE-derived volumetric parameters provide useful measures of adverse LV remodeling after MI in mice. Global strain parameters of myocardial deformation. We next analyzed strain parameters in 35 mice (10 sham, 25 MI). Global peak strain (PS) and strain rate (PSR) were derived by taking the mean of all 6 analyzed segments in longitudinal and radial directions (Fig. 1B). Both longitudinal and radial global strain and the corresponding strain rates were significantly lower in the MI group vs. sham controls (Fig. 6). In a subset of 14 MI mice where both postmortem LV mass and strain data were available, LV mass as a simple measure of total LV remodeling correlated negatively with parameters of both longitudinal and radial strain (Fig. 7). This correlation was stronger for longitudinal strain and strain rate than for radial strain/strain rate. In a subset of seven MI animals, we assessed global strain parameters both from a conventional parasternal long-axis view (involving 6 segments; Fig. 1) and from consecutive

Fig. 5. Linear regression of infarct size against parameters of LV remodeling. Correlation of infarct size with STE-derived end-diastolic volume (EDV; A), STE-derived end-systolic volume (ESV; B), STE-derived ejection fraction (EF; C), and necropsy mass (D). LVM ⫽ LV mass. Infarct size was measured from MRI images. Dotted lines show the 95% confidence range of the regression line. AJP-Heart Circ Physiol • doi:10.1152/ajpheart.00553.2013 • www.ajpheart.org


Fig. 6. Differences in global strain [longitudinal (A) and radial (C)] and strain rate [longitudinal (B) and radial (D)] between MI and control groups. **P ⬍ 0.005, ***P ⬍ 0.0001 by unpaired t-test.

H1377

short-axis views from base to apex (i.e., involving all 16 LV segments). Both sets of global strain parameters were significantly correlated with MRI-derived infarct size (Fig. 8). In general, global strain derived from short-axis views showed slightly better correlation with infarct size than that based on a long-axis view (e.g., radial strain r ⫽ ⫺0.81 vs. r ⫽ ⫺0.63). Regional myocardial strain. Regional deformation was analyzed on a segmental basis (3 anterior and 3 posterior segments) in a subset of 31 animals (14 sham, 17 MI). The mean peak radial strain per segment was significantly lower in MI compared with sham hearts for all segments except the anterior basal segment (Fig. 9). Anterior mid and apical segments showed reduced strain in MI hearts compared with sham as did all three posterior segments. A similar pattern held true for radial strain rate. Longitudinal strain and strain rate, however, were reduced in all segments of MI hearts compared with sham. For all four strain parameters, the apical segments in MIs had lower peak strain/strain rates compared with mid and basal segments, corresponding to the apical location of infarct regions secondary to LAD occlusion. LV dyssynchrony. The data for radial and longitudinal strain and corresponding strain rates that was used for global strain analysis above (i.e., 35 mice: 10 sham, 25 MI) were also analyzed for dyssynchrony. This showed that the standard deviation of the time to peak strain (corrected for the RR interval) was significantly higher in the MI group compared with sham for all four parameters (Fig. 10A). Radial intraventricular strain dyssynchrony was 20.4 ⫾ 2% in the MI group vs. 8.5 ⫾ 1.4% in shams, while longitudinal strain dyssynchrony was 16.6 ⫾ 2.2% in the MI group vs. 7.5 ⫾ 1.3% in shams (P ⬍ 0.01 for all comparisons). The dyssynchrony index

Fig. 7. Linear regression of global strain (A, C) and strain rates (B, D) against postmortem LV mass (LVM) in infarcted hearts. The dotted lines show the 95% confidence range of the regression line. AJP-Heart Circ Physiol • doi:10.1152/ajpheart.00553.2013 • www.ajpheart.org



Fig. 8. Linear regression of global strain/strain rate derived from parasternal long-axis views (PSLAX) or sequential short-axis views (SAX) from base to apex against MRI-derived infarct size. The dotted lines show the 95% confidence range of the regression line.



H1379

Fig. 9. Regional segmental strain (A, C) and strain rate (B, D) in MI and control groups. Ant(B), anterior basal; Ant(M), anterior mid; Ant(A), anterior apical; Post(A), posterior apical; Post(M), posterior mid; Post(B), posterior basal. ns: not statistically significant; *P ⬍ 0.05; **P ⬍ 0.005; ***P ⬍ 0.0001.

correlated positively with MRI-derived infarct size for all strain parameters (Fig. 10B); i.e., larger infarcts were associated with greater dyssynchrony. Analysis time, and inter- and intraobserver and test-retest variability. The total offline analysis time required to obtain LV volumes, EF, LV mass, and regional strain curves was 3.5 ⫾ 0.75 min per study for echocardiography. MRI analysis time for volumes and mass was 9 ⫾ 2 min per study. Data for inter- and intraobserver variability and test-retest variability are shown in Table 1. These analyses show acceptable inter- and intraobserver and test-retest variability for most indexes. Estimation of dyssynchrony index was more reproducible when using strain than strain rate. DISCUSSION

Despite significant recent advancements in clinical echocardiography, studies in gene-modified mouse models generally employ relatively basic echocardiographic imaging methods and analyses with only a few exceptions (24). This is in stark contrast to the complex and innovative molecular biological techniques that are commonly used to create these models. In this study, we have undertaken a careful evaluation of STEbased imaging in mice using a dedicated high-frequency small animal ultrasound system. We focused on mice undergoing adverse LV remodeling after permanent coronary ligation, a commonly used experimental model that is challenging for

accurate estimation of LV volumes and mass from (1D) Mmode or 2D echocardiography due to the irregular geometric shape of the remodeled LV. Our results indicate that STEbased assessment of global LV volumes and function as well as regional function in mice is fast, simple, and reproducible, compares favorably with MRI, and is suitable for the assessment of LV remodeling after MI. The STE-based quantification of parameters of global LV structure and function (i.e., LV volumes, mass, and EF) were automatically calculated by the software when performing a deformation analysis, and therefore required minimal additional operator input. STE-derived parameters measured in this manner correlated very well with MRI-derived measures in the same animals. Furthermore, there was a good linear correlation between infarct size and STE-derived measures of LV dilatation and LV function (as would be expected from first principles), indicating that the technique can be applied across a group of mice with variable degrees of infarction and LV remodeling and provides biologically sensible data in this setting. STE-derived LV volumes and mass underestimated the values obtained by MRI. In the context of volumes, because both ESV and EDV were generally underestimated, the EF data were closer to the corresponding MRI values. Assessment of LV volumes and mass by standard 2D-echocardiography also correlated with MRI data, but there were larger biases than with STE-derived measures. This may at least in part reflect the




Fig. 10. Dyssynchrony derived from strain and strain rates. A: comparison of dyssynchrony derived from longitudinal and radial strain and strain rates between MI and sham control groups. **P ⬍ 0.005; ***P ⬍ 0.0001. B: linear regression of strain-derived dyssynchrony against MRI-derived infarct size.

different algorithms that were employed for STE vs. standard 2D-derived measures (see MATERIALS AND METHODS). A similar underestimation of LV volumes by STE compared with 3D echocardiography or MRI is also reported in human studies (14, 15). The underestimation is most probably a result of foreshortening of the true LV long axis and inaccuracies caused by geometrical assumptions. In addition, difficulty in differentiating trabeculations from the true endocardial boundary also plays a role. Therefore, MRI remains the gold standard

for absolute quantification of LV volumes but STE-derived measures may have significant utility in serial assessment and for comparison among groups. Another clear advantage of MRI remains the direct determination of infarct size, which is often critical in experimental models in ensuring equal stimuli to postinfarct LV remodeling across different study groups (21, 22). On the other hand, echocardiography has a number of other potential benefits over MRI, namely 1) greater availability to researchers in many institutions; 2) greater tolerability—




Table 1. Inter- and intraobserver variability for measured parameters Intraobserver (n ⫽ 12)

Standard EDV, ␮l Standard EF, % Standard LVM, mg STE EDV, ␮l STE EF, % STE LVM, mg Global radial strain, % Radial strain DI, % Radial strain rate, %/s Radial strain rate DI, % Global longit. strain, % Longit. strain DI, % Longit. strain rate, %/s Longit. strain rate DI, %

Interobserver (n ⫽ 8)

Test Retest (n ⫽ 7)

Mean difference (SD)

Correlation


Correlation


Correlation

6.9 (18.2) 2.3 (4.1) 4.6 (14.6) 2.7 (17.2) 2.6 (2) 3.3 (12.3) 0.8 (1.1) 0.9 (3.4) 1.1 (0.9) 2.5 (7.3) 1.4 (1.9) 4.2 (10.2) 1.8 (1.1) 3.7 (9.9)

0.94 0.95 0.93 0.94 0.93 0.94 0.98 0.90 0.92 0.68 0.96 0.74 0.85 0.68

4.3 (19) 4.8 (4.3) 17 (18.6) 3.6 (22.2) 4.8 (3.4) 5.4 (14.3) 0.7 (1.7) 1.8 (8.1) 1.2 (1.7) 2 (13.5) 0.8 (2.7) 6.6 (12.9) 3.2 (2.3) 5.2 (9.7)

0.93 0.89 0.88 0.94 0.94 0.94 0.94 0.82 0.83 0.71 0.90 0.69 0.81 0.64

7.3 (19.2) 5.2 (5.1) 14 (16.2) 1 (10.1) 3.8 (4) 4.6 (15.2) 1.7 (3.1) 2 (7.2) 0.3 (1.1) 1.3 (12.4) 0.6 (1.8) 1.8 (7.2) 0.4 (1.6) 3.1 (7.2)

0.87 0.90 0.88 0.92 0.91 0.92 0.85 0.81 0.82 0.78 0.86 0.79 0.82 0.76

EDV, end-diastolic volume; EF, ejection fraction; LVM, left ventricular mass; PS, peak strain; PSR, peak strain rate; STE, speckle tracking echocardiography; longit., longitudinal; DI, dyssynchrony index

some groups even use echo in conscious mice that have undergone training in order to allow this (32). whereas the remote positioning of mice far from the experimenter during MRI renders this impractical in MRI scanning; 3) better frame rates during scanning (in the present study, typically around 200 Hz for echocardiography vs. 83 Hz for MRI), permitting increased confidence in correctly acquiring end-diastolic and end-systolic time points; 4) lower equipment costs; 5) portability, allowing its use as an adjunct to other experiments, e.g., during pressure-volume loop assessment using conductance catheterization, echo can be used to allow calibration of volume in absolute terms by performing simultaneous measurement of LV cavity size on echo (while the conductance catheter is in place and measuring volume). This could not be achieved using MRI due to fixed nature of the equipment, issues relating to the magnetic field, and the remote positioning of the experimental animal in the scanner. A second aim of our study was to investigate parameters of global deformation. STE-derived deformation data readily differentiated between infarcted mice and control animals. More importantly, global strain parameters correlated well with MRI-derived infarct size or postmortem LV mass, as a determinant and a simple measure of the degree of LV remodeling, respectively. In other words, larger infarcts and hearts with higher mass (more remodeling) were associated with lower strain. Although global strain parameters derived from consecutive short-axis views which involved 16 segments showed slightly higher correlation with infarct size than those based on a single long-axis view involving 6 segments, strong correlations were still found in the latter case. Therefore, data based on a single long-axis view provide useful information but can be extended to a more time-consuming analysis involving multiple short-axis views. Clinical data show similar relations between global strain and abnormal LV function. For example, a recent study in patients with hypertrophic cardiomyopathy (HCM) that compared STE and delayed hyperenhancement MRI found a correlation between global longitudinal strain and the number of fibrotic segments, as well as total LV hypertrophy (19). Furthermore, Peng et al. (16) reported that circumferential strain accurately tracks the progression of heart failure induced by transaortic constriction in mice and that it was

reflective of myocardial fibrotic changes. Whether the relationship we observed between longitudinal strain and LV mass may also, in part, be attributable to the presence of myocardial fibrosis requires further study. Strain rate has also been shown to be useful in the assessment of transmurality of MI early after reperfusion or at follow up (29, 31, 35). Regional assessment is an important application of strain, and this has been validated in humans (3, 7, 8). Regional function was assessed in our study by comparing myocardial deformation at six different segments involving anterior and posterior walls. Comparison of mean strain from these segments revealed that radial strain parameters were relatively well preserved at the anterior basal wall of infarcted hearts compared with more distal regions, consistent with these segments being remote from the infarct. However, longitudinal strain was significantly impaired in basal segments as well as more distal ones, suggesting that it may be a more sensitive marker of subtle changes in function. Indeed, changes in longitudinal strain are among the first parameters to be affected during cardiac remodeling (6, 9, 23). Longitudinal strain is more sensitive than radial strain in acute and chronic ischemia in humans (6). In patients with hypertensive heart failure, longitudinal strain is impaired in those with milder symptoms whereas radial parameters are only affected with more severe limitation (9). The increased sensitivity of longitudinal strain may be related to the fact that longitudinal function is largely a result of endocardial fiber motion, which is altered earlier during postischemic remodeling. It is reported, however, that the relative contribution of long-axis function may be lower in small mammals than humans (20), and this is an interesting area for future studies. Interestingly, posterior basal segments (which are remote from the infarct) showed reduced radial as well as longitudinal strain compared with sham animals. This may reflect the complexity of the LV remodeling that occurs after MI in the mouse permanent ligation model, being influenced both by the regional properties of the myocardium (e.g., hypertrophy, fibrosis, contractile function) and by interactions among segments. Comparison with other studies after mouse MI is difficult because previous work either only reported strain measured in a single short-axis slice (28), or studied reperfused MI (11, 12). However, Li et al.




(12) reported significant reduction in regional function in remote segments after reperfused MI, which might be expected to be even worse after permanent coronary ligation. It would be useful in future studies to compare STE-derived strain data with MRI-derived measures, which was not possible in the current study for technical reasons. LV intraventricular dyssynchrony is an important therapeutic target in humans, and its assessment by echocardiography is topical. Methods that have been proposed in clinical practice include tissue Doppler imaging and STE (28, 34). The potential advantage of STE over tissue Doppler, especially in murine imaging, is its lack of dependence on the angle of incidence of the ultrasound beam (27). In the current study, STE-based measures of dyssynchrony reliably differentiated between animals with MI and controls. Furthermore, the extent of dyssynchrony correlated positively with infarct size. To our knowledge, this is the first study to use a commercially available software package designed for small animal imaging to assess LV dyssynchrony in mice. Other studies have reported on LV dyssynchrony measured with a commercially available system in a canine model (1). We show that assessment of LV dyssynchrony with the current system is straightforward in mice although additional studies are required to assess the reliability of such data in more detail. It is likely that data fidelity would be improved by averaging peak values over a number of cardiac cycles, which was not done in the current study. Other technical issues that may impact on the applicability of the technique include the ability to define peak strain values in infarcted segments with flattened strain curves, the definition of timing of aortic valve opening and closure, and limitations concerning maximal frame rates. The absolute values that we obtained for strain and strain rate appear to be lower than other published data (11, 26). In their paper from 2005, Sebag et al. (26) used color-coded tissue Doppler imaging to assess myocardial velocities and strain rates in mice at rest and during infusions of esmolol and dobutamine. They were only able to assess radial deformation but obtained radial strain rates in resting mice of around 12 s⫺1. Li et al. (11) assessed radial and circumferential strain in normal and post-MI mice using custom STE software and obtained radial strain values of up to 40% in normal mice. When directly comparing these results to our findings, several important issues should be considered. First, results using different techniques and different software packages cannot necessarily be extrapolated from one to another. Second, the values may be significantly affected by differences in animal sedation and heart rates. Third, one important trade-off for the angle independency of STE is a more limited frame rate capability. While a frame rate of ⱖ150 Hz is suggested by the manufacturer to be sufficient for reliable data, different frame rates could affect the absolute values of strain. In a recently published study using an identical system to that employed in our work, Bauer et al. (2) reported mean longitudinal and radial strain and strain rates in control mice that were similar to our results. In their sham-operated animals they found mean longitudinal strain and strain rate to be ⫺15.7% and ⫺7.4 s⫺1, respectively, compared with ⫺15.3% and ⫺6.6 s⫺1 in our study. Mean radial strain and strain rate derived from their long-axis images were 19.4% and 8 s⫺1, respectively, compared with our findings of 17.1% and 5.5 s⫺1. These authors also studied mice after MI and found mean longitudinal

strain and strain rate to be -3.4% and -2.3 s-1 compared to our values of ⫺5.7% and ⫺2.5 s⫺1, respectively. These figures for radial deformation were 8% and 5.2 s⫺1 for Bauer et al. and 9.1% and 3.6 s⫺1 in our animals. Their follow up scans, however, were performed at time intervals different from ours (3 wk rather than 4) making a direct comparison of these results a little more difficult. In general, strain data of the type reported here are probably best compared with measures that use identical systems and similar anesthetic protocols. The present study indicates that high-frequency 2D STE offers reliable and reproducible assessment of both global and regional function in post-MI remodeling mouse hearts, involving relatively straightforward and rapid analysis with dedicated software. STE-derived parameters of global LV structure and function correlate well with the noninvasive gold standard of MRI and appear suitable for distinguishing among varying extents of adverse LV remodeling and function. In addition to the traditional measures of LV structure and function, we confirm the feasibility of obtaining detailed STE-based data on global and segmental deformation. STE-based noninvasive assessment of cardiac structure and function therefore has the potential to significantly enhance studies investigating LV remodeling in mouse models. GRANTS This work was supported by The British Heart Foundation (BHF) and the Fondation Leducq. A. Sirker was funded by a British Cardiovascular Society Prize Fellowship. DISCLOSURES No conflicts of interest, financial or otherwise, are declared by the author(s). AUTHOR CONTRIBUTIONS Author contributions: A.B., A.S., J.Z., and A.P. performed experiments; A.B., A.S., J.Z., A.P., N.C., and W.D. analyzed data; A.B., A.S., J.Z., A.P., R.B., M.M., and A.M.S. interpreted results of experiments; A.B. and J.Z. prepared figures; A.B. and A.M.S. drafted manuscript; A.B., A.S., J.Z., A.P., W.D., R.B., M.M., and A.M.S. approved final version of manuscript; A.S., J.Z., and A.M.S. edited and revised manuscript; A.M.S. conception and design of research. REFERENCES 1. Arita T, Sorescu GP, Schuler BT, Schmarkey LS, Merlino JD, VintenJohansen J, Leon AR, Martin RP, Sorescu D. Speckle-tracking strain echocardiography for detecting cardiac dyssynchrony in a canine model of dyssynchrony and heart failure. Am J Physiol Heart Circ Physiol 293: H735–H742, 2007. 2. Bauer M, Cheng S, Jain M, Ngoy S, Theodoropoulos C, Trujillo A, Lin FC, Liao R. Echocardiographic speckle-tracking-based strain imaging for rapid cardiovascular phenotyping in mice. Circ Res 108: 908 –916, 2011. 3. Becker M, Bilke E, Kühl H, Katoh M, Kramann R, Franke A, Bücker A, Hanrath P, Hoffmann R. Analysis of myocardial deformation based on pixel tracking in two dimensional echocardiographic images enables quantitative assessment of regional left ventricular function. Heart 92: 1102–1108, 2006. 4. Bijnens B, Claus P, Weidemann F, Strotmann J, Sutherland GR. Investigating cardiac function using motion and deformation analysis in the setting of coronary artery disease. Circulation 116: 2453–2464, 2007. 5. Brown J, Jenkins C, Marwick TH. Use of myocardial strain to assess global left ventricular function: a comparison with cardiac magnetic resonance and 3-dimensional echocardiography. Am Heart J 157: 102. e101–e105, 2009. 6. Chan J, Hanekom L, Wong C, Leano R, Cho GY, Marwick TH. Differentiation of subendocardial and transmural infarction using twodimensional strain rate imaging to assess short-axis and long-axis myocardial function. J Am Coll Cardiol 48: 2026 –2033, 2006.


Innovative Methodology SPECKLE TRACKING ECHOCARDIOGRAPHY IN MICE 7. Donal E, Tournoux F, Leclercq C, De Place C, Solnon A, Derumeaux G, Mabo P, Cohen-Solal A, Daubert JC. Assessment of longitudinal and radial ventricular dyssynchrony in ischemic and nonischemic chronic systolic heart failure: a two-dimensional echocardiographic speckle-tracking strain study. J Am Soc Echocardiogr 21: 58 –65, 2008. 8. Korinek J, Wang J, Sengupta PP, Miyazaki C, Kjaergaard J, McMahon E, Abraham TP, Belohlavek M. Two-dimensional strain—a Dopplerindependent ultrasound method for quantitation of regional deformation: validation in vitro and in vivo. J Am Soc Echocardiogr 18: 1247–1253, 2005. 9. Kosmala W, Plaksej R, Strotmann JM, Weigel C, Herrmann S, Niemann M, Mende H, Störk S, Angermann CE, Wagner JA, Weidemann F. Progression of left ventricular functional abnormalities in hypertensive patients with heart failure: an ultrasonic two-dimensional speckle tracking study. J Am Soc Echocardiogr 21: 1309 –1317, 2008. 10. Langeland S, D’Hooge J, Wouters PF, Leather HA, Claus P, Bijnens B, Sutherland GR. Experimental validation of a new ultrasound method for the simultaneous assessment of radial and longitudinal myocardial deformation independent of insonation angle. Circulation 112: 2157– 2162, 2005. 11. Li Y, Garson CD, Xu Y, Beyers RJ, Epstein FH, French BA, Hossack JA. Quantification and MRI validation of regional contractile dysfunction in mice post myocardial infarction using high resolution ultrasound. Ultrasound Med Biol 33: 894 –904, 2007. 12. Li Y, Garson CD, Xu Y, Helm PA, Hossack JA, French BA. Serial ultrasound evaluation of intramyocardial strain after reperfused myocardial infarction reveals that remote zone dyssynchrony develops in concert with left ventricular remodeling. Ultrasound Med Biol 37: 1073–1086, 2011. 13. Looi YH, Grieve DJ, Siva A, Walker SJ, Anilkumar N, Cave AC, Marber M, Monaghan MJ, Shah AM. Involvement of Nox2 NADPH oxidase in adverse cardiac remodeling after myocardial infarction. Hypertension 51: 319 –325, 2008. 14. Mor-Avi V, Jenkins C, Kuhl HP, Nesser HJ, Marwick T, Franke A, Ebner C, Freed BH, Steringer-Mascherbauer R, Pollard H, Weinert L, Niel J, Sugeng L, Lang RM. Real-time 3-dimensional echocardiographic quantification of left ventricular volumes: multicenter study for validation with magnetic resonance imaging and investigation of sources of error. JACC Cardiovasc Imaging 1: 413–423, 2008. 15. Nishikage T, Nakai H, Mor-Avi V, Lang RM, Salgo IS, Settlemier SH, Husson S, Takeuchi M. Quantitative assessment of left ventricular volume and ejection fraction using two-dimensional speckle tracking echocardiography. Eur J Echocardiogr 10: 82–88, 2008. 16. Peng Y, Popovic ZB, Sopko N, Drinko J, Zhang Z, Thomas JD, Penn MS. Speckle tracking echocardiography in the assessment of mouse models of cardiac dysfunction. Am J Physiol Heart Circ Physiol 297: H811–H820, 2009. 17. Perk G, Tunick PA, Kronzon I. Non-Doppler two-dimensional strain imaging by echocardiography–from technical considerations to clinical applications. J Am Soc Echocardiogr 20: 234 –243, 2007. 18. Picard MH, Popp RL, Weyman AE. Assessment of left ventricular function by echocardiography: a technique in evolution. J Am Soc Echocardiogr 21: 14 –21, 2008. 19. Popovic ZB, Kwon DH, Mishra M, Buakhamsri A, Greenberg NL, Thamilarasan M, Flamm SD, Thomas JD, Lever HM, Desai MY. Association between regional ventricular function and myocardial fibrosis in hypertrophic cardiomyopathy assessed by speckle tracking echocardiography and delayed hyperenhancement magnetic resonance imaging. J Am Soc Echocardiogr 21: 1299 –1305, 2008. 20. Popovic ZB, Sun JP, Yamada H, Drinko J, Mauer K, Greenberg NL, Cheng Y, Moravec CS, Penn MS, Mazgalev TN, Thomas JD. Differences in left ventricular long-axis function from mice to humans follow allometric scaling to ventricular size. J Physiol 568: 255–265, 2005.

H1383

21. Protti A, Dong X, Sirker A, Botnar R, Shah AM. MRI-based prediction of adverse cardiac remodeling after murine myocardial infarction. Am J Physiol Heart Circ Physiol 303: H309 –H314, 2012. 22. Protti A, Sirker A, Shah AM, Botnar R. Late gadolinium enhancement of acute myocardial infarction in mice at 7T: cine-FLASH versus inversion recovery. J Magn Reson Imaging 32: 878 –886, 2010. 23. Reant P, Labrousse L, Lafitte S, Bordachar P, Pillois X, Tariosse L, Bonoron-Adele S, Padois P, Deville C, Roudaut R, Dos Santos P. Experimental validation of circumferential, longitudinal, and radial 2-dimensional strain during dobutamine stress echocardiography in ischemic conditions. J Am Coll Cardiol 51: 149 –157, 2008. 24. Scherrer-Crosbie M, Thibault HB. Echocardiography in translational research: of mice and men. J Am Soc Echocardiogr 21: 1083–1092, 2008. 25. Schiller NB, Shah PM, Crawford M, DeMaria A, Devereux R, Feigenbaum Gutgesell H, Reichek N, Sahn D, Schnittger I. Recommendations for quantitation of the left ventricle by two-dimensional echocardiography. American Society of Echocardiography Committee on Standards, Subcommittee on Quantitation of Two-Dimensional Echocardiograms. J Am Soc Echocardiogr 2: 358 –367, 1989. 26. Sebag IA, Handschumacher MD, Ichinose F, Morgan JG, Hataishi R, Rodrigues AC, Guerrero JL, Steudel W, Raher MJ, Halpern EF, Derumeaux G, Bloch KD, Picard MH, Scherrer-Crosbie M. Quantitative assessment of regional myocardial function in mice by tissue Doppler imaging: comparison with hemodynamics and sonomicrometry. Circulation 111: 2611–2616, 2005. 27. Sivesgaard K, Christensen SD, Nygaard H, Hasenkam JM, Sloth E. Speckle tracking ultrasound is independent of insonation angle and gain: an in vitro investigation of agreement with sonomicrometry. J Am Soc Echocardiogr 22: 852–858, 2009. 28. Thibault H, Gomez L, Donal E, Augeul L, Scherrer-Crosbie M, Ovize M, Derumeaux G. Regional myocardial function after myocardial infarction in mice: a follow-up study by strain rate imaging. J Am Soc Echocardiogr 22: 198 –205, 2009. 29. Thibault H, Gomez L, Donal E, Pontier G, Scherrer-Crosbie M, Ovize M, Derumeaux G. Acute myocardial infarction in mice: assessment of transmurality by strain rate imaging. Am J Physiol Heart Circ Physiol 293: H496 –H502, 2007. 30. Thomas JD, Popovi´c ZB. Assessment of left ventricular function by cardiac ultrasound. J Am Coll Cardiol 48: 2012–2025, 2006. 31. Weidemann F, Dommke C, Bijnens B, Claus P, D’Hooge J, Mertens P, Verbeken E, Maes A, Van de Werf F, De Scheerder I, Sutherland GR. Defining the transmurality of a chronic myocardial infarction by ultrasonic strain-rate imaging: implications for identifying intramural viability: an experimental study. Circulation 107: 883–888, 2003. 32. Yang XP, Liu YH, Rhaleb NE, Kurihara N, Kim HE, Carretero OA. Echocardiographic assessment of cardiac function in conscious and anesthetized mice. Am J Physiol Heart Circ Physiol 277: H1967–H1974, 1999. 33. Youn HJ, Rokosh G, Lester SJ, Simpson P, Schiller NB, Foster E. Two-dimensional echocardiography with a 15-MHz transducer is a promising alternative for in vivo measurement of left ventricular mass in mice. J Am Soc Echocardiogr 12: 70 –75, 1999. 34. Yu CM, Gorcsan J, Bleeker GB, Zhang Q, Schalij MJ, Suffoletto MS, Fung JW, Schwartzman D, Chan YS, Tanabe M, Bax JJ. Usefulness of tissue Doppler velocity and strain dyssynchrony for predicting left ventricular reverse remodeling response after cardiac resynchronization therapy. Am J Cardiol 100: 1263–1270, 2007. 35. Zhang Y, Chan AK, Yu CM, Yip GW, Fung JW, Lam WW, So NM, Wang M, Wu EB, Wong JT, Sanderson JE. Strain rate imaging differentiates transmural from non-transmural myocardial infarction: a validation study using delayed-enhancement magnetic resonance imaging. J Am Coll Cardiol 46: 864 –871, 2005.


Two-dimensional speckle-tracking echocardiography assessment of left ventricular remodeling in patients after myocardial infarction and primary reperfusion.

Evaluation of left and right ventricular myocardial function after lung resection using speckle tracking echocardiography.

2D speckle tracking echocardiography for the assessment of regional contractile reserve after myocardial infarction.

Echocardiographic Predictors for Left Ventricular Remodeling after Acute ST Elevation Myocardial Infarction with Low Risk Group: Speckle Tracking Analysis.

Cyclosporin A in left ventricular remodeling after myocardial infarction.

Left ventricular twist during dobutamine stress echocardiography after acute myocardial infarction: association with reverse remodeling.

Assessment of left ventricular longitudinal function in cats with subclinical hypertrophic cardiomyopathy using tissue Doppler imaging and speckle tracking echocardiography.

Diacerein improves left ventricular remodeling and cardiac function by reducing the inflammatory response after myocardial infarction.

Two-dimensional speckle tracking echocardiography prognostic parameters in patients after acute myocardial infarction.

Quantitative assessment of systolic left ventricular function with speckle-tracking echocardiography in adult patients with repaired aortic coarctation.

Assessment of left ventricular function by two-dimensional speckle-tracking echocardiography in small breed dogs with hyperadrenocorticism.

Outcome prediction in sepsis: speckle tracking echocardiography based assessment of myocardial function.

Methodological Gaps in Left Atrial Function Assessment by 2D Speckle Tracking Echocardiography.

Right ventricular function assessment using tissue Doppler imaging and speckle tracking echocardiography.

Ventricular remodeling and prognosis after myocardial infarction.

Testosterone suppresses ventricular remodeling and improves left ventricular function in rats following myocardial infarction.

The relationship between time from myocardial infarction, left ventricular dyssynchrony, and the risk for ventricular arrhythmia: speckle-tracking echocardiographic analysis.

Evaluation of left ventricular function by three-dimensional speckle-tracking echocardiography in patients with myocardial bridging of the left anterior descending coronary artery.

Predicting Left Ventricular Dysfunction after Surgery in Patients with Chronic Mitral Regurgitation: Assessment of Myocardial Deformation by 2-Dimensional Multilayer Speckle Tracking Echocardiography.

Association between multilayer left ventricular rotational mechanics and the development of left ventricular remodeling after acute myocardial infarction.

Three-dimensional speckle tracking echocardiography and cardiac magnetic resonance for left ventricular chamber quantification and identification of myocardial transmural scar.

The role of speckle tracking echocardiography in assessment of lipopolysaccharide-induced myocardial dysfunction in mice.

Long-Term Cardiac Remodeling in Elite Athletes: Assessment by Tissue Doppler and Speckle Tracking Echocardiography.

Post-myocardial infarction left ventricular pseudoaneurysm diagnosed incidentally by echocardiography.