• Open Access

Noninvasive Assessment of Systolic Left Ventricular Torsion by 2-Dimensional Speckle Tracking Imaging in the Awake Dog: Repeatability, Reproducibility, and Comparison with Tissue Doppler Imaging Variables

Authors

  • V. Chetboul,

    1. Unité de Cardiologie d'Alfort, Ecole Nationale Vétérinaire d'Alfort, 94 704 Maisons-Alfort cedex, France
    2. INSERM U841 (National Institute of Health and Medical Research), Ecole Nationale Vétérinaire d'Alfort, 94 704 Maisons-Alfort cedex, France
    Search for more papers by this author
  • F. Serres,

    1. Unité de Cardiologie d'Alfort, Ecole Nationale Vétérinaire d'Alfort, 94 704 Maisons-Alfort cedex, France
    Search for more papers by this author
  • V. Gouni,

    1. Unité de Cardiologie d'Alfort, Ecole Nationale Vétérinaire d'Alfort, 94 704 Maisons-Alfort cedex, France
    2. INSERM U841 (National Institute of Health and Medical Research), Ecole Nationale Vétérinaire d'Alfort, 94 704 Maisons-Alfort cedex, France
    Search for more papers by this author
  • R. Tissier,

    1. Unité de Cardiologie d'Alfort, Ecole Nationale Vétérinaire d'Alfort, 94 704 Maisons-Alfort cedex, France
    2. INSERM U841 (National Institute of Health and Medical Research), Ecole Nationale Vétérinaire d'Alfort, 94 704 Maisons-Alfort cedex, France
    Search for more papers by this author
  • J.L. Pouchelon

    1. Unité de Cardiologie d'Alfort, Ecole Nationale Vétérinaire d'Alfort, 94 704 Maisons-Alfort cedex, France
    2. INSERM U841 (National Institute of Health and Medical Research), Ecole Nationale Vétérinaire d'Alfort, 94 704 Maisons-Alfort cedex, France
    Search for more papers by this author

Corresponding author: Valérie Chetboul, DVM, PhD, Dipl-ECVIM-CA (Cardiology), Unité de Cardiologie d'Alfort, Ecole Nationale Vétérinaire d'Alfort, 7 Avenue du Général de Gaulle, 94704 Maisons-Alfort cedex, France; e-mail: vchetboul@vet-alfort.fr

Abstract

Background: Left ventricular (LV) torsional deformation plays an important role in myocardial function. However, it has never been assessed in the awake dog, because magnetic resonance imaging and sonomicrometry have been the only methods available so far.

Hypothesis: Two dimensional speckle tracking echocardiography (STE), a new ultrasound imaging technique, provides a repeatable and reproducible noninvasive assessment of systolic LV wringing motion in the awake dog.

Animals: Six healthy dogs were used to determine the repeatability and reproducibility of STE variables (study 1). These variables also were prospectively assessed in a population of 35 healthy dogs (study 2).

Methods: Peak LV basal and apical systolic rotations were measured by STE from right parasternal short-axis views using automatic frame-to-frame tracking of gray-scale speckle patterns. Systolic LV torsion (LVtor, °) was defined as apical rotation relative to the base.

Results: All within-day and between-day coefficients of variation were <20% (6.8–18.0%). Amplitude of apical systolic rotation was significantly higher (P < .001) than the basal value (5.4 ± 3.2° and −3.1 ± 1.3°, respectively). Global LVtor was significantly correlated with systolic longitudinal LV myocardial velocity gradient assessed by tissue Doppler (P < .05), but not with either systolic radial LV myocardial velocity gradient or the ratio of early mitral inflow velocity to early mitral annular velocity (Em/Ea).

Conclusions and clinical importance: Speckle tracking echocardiography is a repeatable and reproducible method for assessing systolic LV torsional deformation. The combination of these new STE indices with tissue Doppler variables could provide a new approach for quantifying canine LV systolic function.

Torsional deformation of the left ventricle (LV) results from helically oriented myofibers. This wringing motion has been shown to be a key factor of normal systolic and diastolic myocardial function in both dogs1–3 and humans.4,5 Systolic LV twist is characterized by opposite directed apical and basal rotations around the long heart axis.5,6 This torsion motion has an important complementary effect to that of radial and longitudinal systolic myocardial ejection forces. Moreover, it allows elastic energy to be stored in compressed titin.7 Therefore, the subsequent rapid LV untwist, mostly occurring during the isovolumic relaxation phase before mitral valve opening, results from the release of energy stored in elastic elements during the systolic deformation.8 This elastic torsional recoil leads to partial return of the LV to its pre-ejection configuration and contributes to efficient early diastolic filling because of a suction-aided effect.1,9

Assessment of LV torsional deformation seems, therefore, to represent an important approach to quantification of myocardial function. However, this component of heart motion is difficult to measure in vivo. To date, reports on the topic have been very limited, because sonomicrometry and magnetic resonance imaging (MRI) have been the only methods available so far. Sonomicrometry is an invasive technique that can be used to study myocardial motion in the dog but requires anesthesia. MRI tagging was demonstrated to be promising for studying LV torsion (LVtor, °) in humans10–12 and in dogs.1,13 However, its routine practical use in the latter is considerably limited by cost, availability, time consumption, technical complexity of data analysis, and the required anesthesia in dogs. Therefore, LVtor rarely has been assessed in either pathological or research settings until now, and to the best of our knowledge has never been reported and evaluated in the awake dog.

The recent development of ultrasound imaging techniques such as tissue Doppler imaging (TDI) and speckle tracking echocardiography (STE) has provided a new opportunity for the noninvasive study of LV torsional deformation. TDI enables both global and regional myocardial function to be quantified from measurements of myocardial velocities in real time and has been shown to be more sensitive than conventional echocardiography in detecting LV myocardial dysfunction in a dog model of Duchenne's cardiomyopathy.14,15 In humans, twist and untwist of the LV recently have been studied using TDI.4,16 The 2-dimensional (2D) STE principle is based on the formation of speckles caused by reflection, scattering, and interferences between tissue and ultrasound beams in gray scale 2D echocardiographic images.11,17 These speckles are homogeneously distributed within the myocardium on 2D ultrasound images and appear as small, bright elements. They represent natural acoustic tissue markers that can be tracked from frame to frame throughout the cardiac cycle.4,11 2D ultrasound speckle tracking imaging allows a non-Doppler assessment of cardiac motion by filtering out these random speckles, and then performing autocorrelations to evaluate the motion of stable structures using specific software.4 We hypothesized that STE also could be used in the awake dog to assess the LV twisting motion because of the unique feature of angle independence of this technique and the rapidity and ease of data acquisition associated with its use in unanesthetized patients.

The purposes of this prospective study therefore were (1) to determine the within-day and between-day variabilities of peak systolic apical and basal LV rotation and global LV systolic torsion in awake dogs using STE, (2) to quantify these STE variables in a population of healthy dogs and determine the correlation between these new systolic indices and heart rate, age, body weight, and conventional echocardiographic (eg, fractional shortening, %FS) or TDI indices, and last, (3) to provide preliminary results of potential rotation alterations in diseased dogs characterized by myocardial hypokinesia.

Materials and Methods

Experimental Design

Two separate studies were performed.

Study 1. The within-day and between-day variabilities of the STE variables were determined by performing 36 STE examinations on 6 healthy awake dogs (age: 3.1 ± 1.8 years [0.8–5.0 years]; weight: 22.7 ± 11.2 kg [11–40 kg]) on 4 different days over a 2-week period: 1 Cane Corso (neutered female), 1 Beagle (neutered female), 1 Cocker Spaniel (neutered female), 1 Labrador (neutered female), 1 Brittany Spaniel (neutered male), and 1 Boxer (neutered male). On a given day, 3 dogs were examined at 3 nonconsecutive times (therefore a total of 9 ultrasound examinations were performed per day). Each ultrasound examination included the 2 LV short-axis planes obtained at the base and the apical levels required for assessment of basal and apical rotation, respectively (see details on STE method below). Each STE variable was measured from 3 consecutive cardiac cycles using the same frame, and the mean values were used to determine the within-day and between-day variabilities.

Study 2. The study population consisted of 35 prospectively recruited healthy dogs, free of medication, and with no history of heart or respiratory disease (Table 1). The owner's consent for each dog was obtained before its enrollment in the study. All dogs were determined to be healthy on the basis of a complete physical examination, ECG, and standard echo-Doppler examinationsa that were performed just before STE recording to confirm normal heart anatomy and myocardial function. Briefly, each echo-Doppler variable was measured from 3 consecutive cardiac cycles and averaged, and each variable was compared with data obtained from nonsedated dogs in order to confirm that it was within the reference ranges for healthy dogs.18 Conventional echocardiographic variables, all previously validated by our group,19 included ventricular measurements taken from the right parasternal short-axis view using the 2D-guided M-mode (left and right ventricular end-diastolic diameters, LV end-systolic diameter, left ventricular free-wall [LVFW], and interventricular septal thickness in diastole and in systole, the systolic right ventricular wall thickness, and %FS). They also included measurements of the aorta (Ao) and left atrial (LA) diameter by 2D method,19 and calculation of the LA/Ao ratio. The maximal systolic aortic and pulmonary velocities and the maximal early and late diastolic mitral flow velocities also were determined by pulsed-wave Doppler. Last, during the prospective recruitment period of the healthy animals, dogs diagnosed with hypokinesia using M-mode echocardiography (n = 7, %FS = 18 ± 10%, [4–28%]) and TDI at the Cardiology Unit of Alfort also were examined by STE. This decreased myocardial function was caused either by dilated cardiomyopathy (DCM, n = 6) or by patent ductus arteriosus (n = 1 dog treated surgically 2 years ago). This group included 3 Dobermans, 1 Beauce Shepherd, 1 Labrador Retriever, 1 Bull Terrier, and 1 English Bulldog (age = 4.1 ± 3.1 years, [1–9 years]; body weight = 30.8 ± 6.8 kg, [20–42 kg]). All dogs except two received one or a combination of the following drugs: angiotensin-converting–enzyme inhibitors such as benazepril or ramipril (n = 5), furosemide (n = 3), spironolactone (n = 2), and amiodarone (n = 1).

Table 1.   Age, body weight, heart rate, sex, breed, and main conventional echo-Doppler variables in the healthy canine population (n = 35).
 Healthy Canine Population
(n = 35)
Mean ± SD (Range)
Age (years)3.0 ± 2.1 (0.8–8)
Body weight (kg)18.4 ± 11.7 (2–52)
Heart rate (beats/min)104 ± 25 (60–165)
Sex
 Malen = 9
 Femalen = 26
BreedsAkita (n = 1)
Beagle (n = 1)
Boxer (n = 2)
Brittany Spaniel (n = 2)
Bull Terrier (n = 1)
Cairn Terrier (n = 1)
Cane Corso (n = 2)
Cavalier King Charles Spaniel (n = 4)
Cocker Spaniel (n = 3)
German shorthaired Pointer (n = 1)
Golden Retriever (n = 3)
Jack Russel Terrier (n = 2)
King Charles Spaniel (n = 2)
Labrador Retriever (n = 4)
Mixed breed dog (n = 3)
Pyrenean Shepherd (n = 1)
Saarlos Wolfhound (n = 1)
Staffordshire Terrier (n = 1)
Fractional shortening (%)38.4 ± 6.1 (30–49)
Left atrium/aorta0.9 ± 0.1 (0.5–1.1)
Peak aortic velocity (m/s)1.3 ± 0.3 (0.8–1.9)
Peak pulmonary velocity (m/s)0.9 ± 0.2 (0.6–1.6)
Mitral E wave velocity (m/s)0.8 ± 0.1 (0.6–1.0)
Mitral A wave velocity (m/s)0.5 ± 0.1 (0.2–0.8)
Mitral E/A ratio1.8 ± 0.4 (1.2–3.3)

Dogs from Study 1 and Study 2 were used in another procotol for the assessment of strain and strain rate by STE.20

Color TDI Examinations. 2D color TDI examinations were performed in awake standing dogs with continuous ECG monitoring by the same trained observer (V.C.) and using the same ultrasound unit as for standard echocardiography.a For each examination, the gray scale receive gain was set to optimize the clarity of the endocardial and epicardial boundaries of each myocardial wall. Real-time color Doppler images were superimposed on the grayscale with a frame rate ≥100 frames/s. The Doppler receive gain was adjusted to maintain optimal coloring of the myocardium, and the Doppler velocity range was set as low as possible while still avoiding aliasing artifacts. Digital images were obtained and stored for later review (performed on the same day) with image management software.b A 2 × 2 mm sampling was used, and a tissue velocity profile was displayed in each sample location. The LVFW velocities resulting from the radial LV motion were measured using the right parasternal ventricular short-axis view between the 2 papillary muscles, as previously described and validated by our group.21,22 Measurements were made in an endocardial and epicardial segment of the LVFW. Simultaneous endocardial and epicardial velocity profiles were obtained during the offline analysis. Radial myocardial velocities were determined in systole, and the radial systolic myocardial velocity gradient (MVG[cm/s]), defined as the difference between endocardial and epicardial systolic velocities, then was calculated. The LVFW velocities resulting from the longitudinal LV motion were measured using the standard left apical 4-chamber view, as previously described and validated by our group.21,22 Measurements were made of 2 myocardial segments in the internal midportion of the LVFW, a basal segment, and an apical segment. Simultaneous basal and apical velocity profiles were obtained during the offline analysis. Longitudinal myocardial velocities were determined in systole, and longitudinal MVG, defined as the difference between basal and apical systolic velocities, was then calculated. Using the same view, the early mitral annular velocity (Ea) also was assessed, and the ratio of early mitral inflow velocity to early mitral annular velocity (Em/Ea) was calculated. For each TDI variable, measurements were made on 3 consecutive cardiac cycles on the same frame, and the mean value was used for statistical analysis.

Speckle Tracking Echocardiography. The STE analysis was performed by the same trained observer (V.C.) as for the TDI examinations but 4–24 weeks later using softwareb,23–25 and a separate personal computer workstation that did not include the TDI and conventional echocardiographic data. Therefore, the observer was blinded to the TDI and conventional echocardiographic examinations while assessing the STE variables. The 2D echocardiographic loops used for STE analysis were acquired and recorded in awake standing dogs with continuous ECG monitoring using the same ultrasound unit as for the standard echocardiography and TDI examinations.a At the end of the standard echocardiogram, 2 LV short-axis planes obtained at the base and the apical levels were acquired using 2nd harmonic gray scale imaging with sampling rates from 70 to 110 frames/s, as already described in humans23–25 using the same software.b,c The 2 short-axis image planes were standardized as follows: the basal level was defined as the one showing the mitral leaflets without visualization of the papillary muscles, whereas the apical level was acquired just below the papillary muscles. The LV cross-section for each short-axis acquisition was made as circular as possible. For each plane, 3 consecutive cardiac cycles were acquired and digitally stored in Raw Dicom format on a hard disk for offline analysis on a workstation. The rotation STE variables were measured in 5 steps: (1) In each short axis view, the endocardial border of the LV was manually traced at end-systole (Fig 1A). (2) A circular region of interest, in which the computer software automatically performed speckle tracking, then was drawn to include the entire myocardium, and the software defined the ventricular centroid for the midmyocardial line on each frame (Fig 1B). (3) The software algorithm then automatically segmented the LV short-axis into 6 equidistant segments (involving the interventricular septum and the LVFW) and selected suitable speckles for tracking (Fig 1C). Once completed, the software algorithm automatically searched for these speckles on a frame-by-frame basis using the sum of absolute difference algorithm. (4) The speckle-tracking algorithm then provided a tracking score, representing the reliability of tracking based on the degree of decorrelation of the block-matching algorithm. Values for individual regional tracking scores ranged from 1 (excellent) to 3 (poor). Myocardial segments with a score ≥2.5 were excluded from analysis. (5) Six LV rotation profiles then were obtained, corresponding to the average rotation of each myocardial segment (Fig 1D). For all of the rotation profiles, the starting point was manually placed just before the qRs complex on the concomitant ECG tracing to reset rotation angles to zero. A 7th rotation profile (dotted line, Fig 1D) was obtained, representing the averaged (or global) LV rotation profile of the 6 segments. This profile was used to perform all STE measurements, including peak systolic rotation (°) and time to peak systolic rotation (time between the starting point and the peak systolic rotation). The LVtor then was calculated as defined in previous studies23–27 (ie, LV apical rotation relative to the base [or absolute value of {peak systolic basal rotation–peak systolic apical rotation}]). As chosen in humans using the same software,23–25 counterclockwise rotation as viewed from the LV apex was expressed as a positive value, and clockwise rotation as a negative value. For each STE variable, measurements were made on 3 consecutive cardiac cycles, and the mean value (representing an STE analysis of 18 myocardial segments per dog) was used for statistical analysis. Heart rate was calculated during each STE examination from an ECG taken during the same 3 cardiac cycles used for the STE measurements.

Figure 1.

 Example of measurement process of left ventricular (LV) basal rotation by two-dimensional (2D) speckle tracking echocardiography in a healthy dog. At the basal level (A), mitral leaflets but no papillary muscles are visualized; the LV endocardial border is traced at end-systole (dotted blue line). A circular region of interest (dotted red lines) then is automatically drawn (B), and the operator sets its width so that it includes the entire myocardium. The software also automatically defines the ventricular centroid (red central dot). The software algorithm then automatically separates the LV short-axis into 6 equidistant myocardial segments (C) within the interventricular septum and the LV free wall. For each myocardial segment, the tracking score (here 1 for each segment) is displayed on the screen. (D) On the right, the 6 corresponding LV basal rotation versus time curves, and the orange dotted line the averaged LV rotation versus time curve of the 6 segments on which STE measurements are made. This representative case demonstrates that, as seen from the apex, the 6 LV segments homogeneously go through a systolic wringing motion with an initial minimal counterclockwise rotation (positive rotation) followed by a dominant clockwise rotation (negative rotation). This also may be observed on 2D color-coded views (left) showing a counterclockwise (blue) and then clockwise (red) rotation at early and end-systole (ES), respectively. (E) An example of LV apical rotation by 2D speckle tracking echocardiography in a healthy dog. As seen from the apex, the 6 LV segments go through a systolic wringing motion with an initial clockwise rotation (negative rotation) followed by a dominant counterclockwise rotation (positive rotation). Note the homogenous distribution of rotation between myocardial segments. This can also be observed on 2D color-coded views (left) showing a clockwise (red) and then counterclockwise (blue) rotation at early and ES, respectively.

Statistical Analysis. Data are expressed as mean ± SD. Statistical analyses were performed using computer software.d The following linear model (ANOVA method) was used for STE rotation and torsion variables in the within-day and between-day variability analysis (study 1):

image

where Yijk is the kth value measured for dog j on day i, μ is the general mean, dayi is the differential effect of day i, dogj is the differential effect of dog j, (day × dog)ij is the interaction term between day and dog, and ɛijk is the model error. Before statistical analysis, the normality of the data was tested using a Kolmogorov-Smirnov statistic, and all data were normally distributed. The SD of repeatability was estimated as the residual SD of the model and the SD of reproducibility as the SD of the differential effect of day. The corresponding coefficients of variation (CV) were determined by dividing each SD by the mean.

Correlations between heart rate, age, body weight, conventional echo-Doppler or TDI indices, and the STE variables were examined by applying the Pearson correlation analysis (study 2). Peak end-systolic rotation and time to peak end-systolic rotation were compared between base and apex by a paired Student's t-test. The same test was used to compare basal and apical peak early-systolic rotation and time to peak early-systolic rotation. An unpaired Student's t-test was used to compare STE variables between healthy and diseased dogs. P values < .05 were considered statistically significant.

Results

General Description of Normal Basal and Apical Rotation versus Time Profiles

As seen from the apex, LV basal segments (Fig 1D) performed a systolic twisting motion with an initial (or early systolic) counterclockwise rotation (positive rotation angle) followed by a clockwise rotation (negative rotation angle, called by convention the end-systolic rotation angle). Conversely, LV apical segments (Fig 1E) performed a systolic twisting motion with an initial (or early systolic) clockwise rotation (negative rotation angle) followed by a counterclockwise rotation (positive rotation angle, called by convention the end-systolic rotation angle). The initial counterclockwise basal rotation and the initial clockwise apical rotation were not observed in 24 and 20% of healthy dogs, respectively.

Within-Day and Between-Day Intraobserver Variabilities of STE Data (Study 1)

All within-day and between-day CV values (n = 6) were <20% (Table 2). The lowest CV values were obtained for peak systolic basal rotation (7.3%) and systolic LVtor (6.8%). A statistically significant interaction was observed between day and dog for only 1 variable (peak end-systolic apical rotation).

Table 2.   Within-day and between-day variability, expressed as SD and coefficients of variation (CV), of peak end-systolic apical rotation, peak end-systolic basal rotation (°), and systolic torsion (°) of the left ventricle assessed by 2D speckle echocardiography.
 Within-DayBetween-Day
SD (°)CV (%)SD (°)CV (%)
  • a

    CV < 15%, considered low.

  • b

    A statistical interaction between day and dog was observed (P < .05).

  • c

    Defined as absolute value of (peak end-systolic basal rotation – peak end-systolic apical rotation).

Basal rotation
 Peak basal rotation at end-systole0.469.07a0.377.33a
Apical rotation
 Peak apical rotation at end-systole0.6411.82a0.9718.02b
Torsionc1.0316.370.436.84a

Quantification of LV Systolic Apical and Basal Rotation, and LV Torsional Deformation in the Population of 35 Healthy Dogs (Study 2)

A total of 77/630 basal (12%) and 24/630 (4%) apical segments could not be analyzed by STE because of a score ≥2.5. Peak end-systolic rotation and time to peak end-systolic rotation measured at the base and the apex, and the deduced LVtor, are presented in Table 3. The characteristics of initial rotation motion (counterclockwise at the base and clockwise at the apex) also are shown. Peak end-systolic rotation at the apex (Fig 2) was significantly higher (P < .001) than the absolute value of peak end-systolic rotation at the base, and occurred significantly sooner (P < .001).

Table 3.   Peak systolic apical and basal rotation (°), time to peak systolic apical and basal rotation (ms), and systolic torsion (°) of the left ventricle assessed by 2D speckle echocardiography in a population of 35 healthy dogs by a single trained observer.
 Mean ± SD (Minimum–Maximum)
Early Rotation
Motion
Late Rotation
Motion
  • a

    P < .001, versus the corresponding absolute basal value.

  • b

    Defined as absolute value of (peak end-systolic basal rotation – peak end-systolic apical rotation).

Apical rotation
Heart rate during apical examination (bpm)107 ± 27 (66–179)
Peak apical rotation (°)−1.6 ± 0.8a (−3.4–−0.4)5.4 ± 3.2a (1.0–12.9)
Time to peak apical rotation (ms)70 ± 22a (34–126)189 ± 36a (99–280)
Basal rotation
Heart rate during basal examination (bpm)107 ± 29 (68–189)
Peak basal rotation (°)3.9 ± 2.8 (0.2–10.7)−3.1 ± 1.3 (−5.5–−1.0)
Time to peak basal rotation (ms)113 ± 45 (26–177)261 ± 43 (136–354)
Torsion at end systole (°)b 8.4 ± 3.8 (2.5–18)
Figure 2.

 Box plots representing 10th, 25th, 50th, 75th, and 90th percentiles of absolute peak end-systolic rotation (expressed in °, A) and time to peak end-systolic rotation (expressed in ms, B) measured in the population of 35 healthy dogs at the left ventricular (LV) base and LV apex, using speckle tracking echocardiography. *P < .001 versus corresponding absolute value.

No correlation was found among the 3 end-systolic STE indices and heart rate, body weight, Em/Ea, or systolic radial MVG assessed by TDI (Table 4). A positive correlation was found between age and peak end-systolic rotation at the base (P= .01, Fig 3A). A positive correlation also was demonstrated (P < .001) between %FS and peak end-systolic apical rotation and LVtor (Fig 3B and C, respectively). Systolic longitudinal MVG assessed by TDI was correlated only with LVtor (P < .05, Fig 3D).

Table 4.   Correlations between speckle tracking variables (peak end-systolic apical and basal rotation, and torsion of the left ventricle) and either heart rate, age, body weight, or conventional echo-Doppler and tissue Doppler variables (%FS, Em/Ea, and systolic radial and longitudinal MVG).
Correlation betweenHeart
Rate
AgeBody
Weight
%FS E m/EaSystolic Radial
MVG
Systolic Longitudinal
MVG
  1. E a, velocity of the mitral valve annulus assessed by tissue Doppler imaging; Em, mitral E wave assessed by pulsed-wave Doppler mode; %FS, fractional shortening; MVG, myocardial velocity gradient; NS, not significant.

  2. Results were obtained in a population of 35 healthy dogs by a single trained observer.

Peak basal rotation at end-systoleNS r= 0.43 (P= .01)NSNSNSNSNS
Peak apical rotation at end-systoleNSNSNSr = 0.59 (P < .001)NSNSNS
TorsionNSNSNSr = 0.60 (P < .001)NSNS r= 0.42 (P < .05)
Figure 3.

 Significant correlations between speckle tracking echocardiography variables measured in healthy dogs (n = 35) and age (A), fractional shortening (FS, B and C), and longitudinal systolic myocardial velocity gradient (MVG, D) assessed by 2D tissue Doppler imaging.

Last, peak end-systolic basal (−1.4 ± 1.1° [−3.9 to −0.6°]) and apical (1.9 ± 1.6° [0.1–4.6°]) rotations, and LVtor (3.3 ± 2.3° [0.7–6.7°]), were significantly decreased in dogs with hypokinesia compared with healthy dogs (P < .01). Times to peak end-systolic rotations were not significantly different between the 2 groups: 238 ± 80 ms (78–339 ms) at the base and 192 ± 36 ms (160–241 ms) at the apex in the diseased group (see Table 3 for values in healthy dogs).

Discussion

Quantitative assessment of systolic myocardial function is of great importance in the diagnosis, treatment, and management of dogs with heart disease. Standard echocardiography is commonly performed on dogs to noninvasively assess systolic myocardial function, and several 2D and M-mode measurements such as systolic LV diameter and index volume or %FS are often used as indices of myocardial performance.18,28 TDI and its derived modalities, strain and strain rate imaging, are newly developed ultrasound techniques permitting quantitative assessment of systolic function by calculating systolic myocardial velocities and the corresponding MVG, and systolic myocardial segmental deformation and the rate of deformation, respectively.21,22,29 Speckle tracking imaging represents an even more recent ultrasound modality based on 2D gray-scale echocardiographic images. One of its major advantages compared with Doppler-based techniques is independence of both the insonation angle and cardiac translation for the quantification of myocardial function, which allows the simultaneous assessment in several myocardial segments of velocity, strain, strain rate, and last, rotation as performed in the present study.17,30 Moreover,17 in contrast to Doppler-based techniques in which all measurements are done in reference to an external point (the transducer), STE offers direct measures of myocardial displacement because speckles, which are “fingerprints” of a given myocardial area, can be identified and followed within the myocardium in consecutive frames during the entire cardiac cycle on gray-scale 2D images, creating the opportunity for the construction of twist-displacement loops.23,25 Studies recently have validated the accuracy of STE versus tagged MRI or sonomicrometry for assessing wringing LV motion in humans and anesthetized dogs.11,26

Description of the normal STE rotation patterns and the determination of repeatability and reproducibility in the awake dog, as performed in this preliminary study, are 2 major prerequisites to use of this ultrasound technique in diseased dogs. The present study demonstrates the ability of 2D speckle tracking to quantify LV twist in the awake dog using basal and apical transventricular short-axis views. Our results show that the repeatability and reproducibility of the systolic STE measurements are adequate for routine clinical use: most within-day and between-day CV values (4/6) were <15%, and all values were <20%. The cut-off value was arbitrarily fixed at 15%, because we had previously demonstrated that, for a trained observer, almost all the within-day and between-day variability of conventional echocardiographic variables was below 15%.19 In the present study, the highest variability was observed for peak end-systolic apical rotation (18%). The variability of apical measurements has already been observed in the dog for the measurement of myocardial velocities using the 2D color TDI technique,22 and may be explained in part by the more variable positioning of the sample volume compared with the base. We hypothesize that this might also have been the case here, and that location of the apical planes varied from dog to dog, thereby inducing some measuring error. In some previously reported STE studies in humans, apical short-axis views were more precisely standardized, although the same anatomic landmarks as ours (papillary muscles not visible) were used, with LV luminal obliteration at end-systole as an additional criterion for proper apical short-axis level.25

The present study demonstrates that the rotation of all interventricular septal and LVFW segments is homogeneous in the awake healthy dog, and that when viewed from the apex, the LV performs a systolic twisting motion with a predominant clockwise rotation at the base and a predominant counterclockwise rotation at the apex. These results are similar to those previously reported using MRI, sonomicrometry, or STE in humans and anesthetized dogs.10–12 In the report of Helle-Velle et al,11 time to peak rotation and rotation variables assessed by STE in anesthetized dogs were higher and lower, respectively, than those obtained in the present study. For example, the mean values for end-systolic apical rotation and torsion were 4.1° and 6.0° (versus 5.4° and 8.4° in the present study, respectively). Similarly, time to peak end-systolic rotation at the apex was 262 ms (versus 189 ms in the present study). This slight discrepancy between the 2 studies may be because of anesthesia, which has been shown to decrease the amplitude of myocardial motion in the dog.22

Most of the dogs in the present study (76 and 80%, respectively) showed a small initial rotation, counterclockwise at the base and in the opposite direction at the apex. Similar early-systolic rotations have been described in humans and in dogs, and may be explained by the earlier activation of the subendocardial fibers (right-handed helix) rather than the subepicardial fibers (left-handed helix).11,23,24 The present results also show that in the normal LV, apical systolic rotation occurs significantly sooner and with a higher amplitude than basal rotation and, unlike the latter, is unaffected by aging. These results are consistent with those obtained in humans,25 confirming the key role of the apex in systolic myocardial performance and the relevance of LV apical rotation for assessing LV systolic function. However, all the dogs of the present study were <10 years old and the relative lack of older animals may have prevented observation of a potential alteration of apical rotation with aging.

As expected, in the current study a positive (although weak) correlation was observed in healthy dogs between %FS and peak end-systolic apical rotation and LVtor, and also between the latter and longitudinal systolic MVG assessed by TDI. Similar results have been obtained in humans.25 Moreover, dogs of the present study with a hypokinetic myocardial tissue had significantly decreased basal and apical LV rotation and global LV twist. Similar alterations of the LV systolic wringing motion also have been identified in humans with DCM and myocardial infarction and could contribute to a certain extent to the reduction in stroke volume in these pathologic settings.31,32 Conversely, in human patients, dobutamine has been shown to augment systolic twist.33 These results again confirm that LV torsional deformation is a reflection of systolic myocardial performance, and that its assessment provides additional information beyond that obtained from conventional echocardiographic and TDI variables.

One interesting feature of STE variables is that all regional rotation and torsion values were independent of heart rate and probably partially independent of preload, because no correlation was identified between the 3 STE variables and Em/Ea, which has been shown to be a marker of mean left atrial pressure in an experimental canine model of mitral regurgitation.34 Consistently, in conscious human patients, volume loading has been shown not to affect systolic LV twist.33 Additional studies, however, are required in diseased dogs with spontaneous mitral regurgitation to better study such interaction.

One limitation of the present study is that our protocol was limited to normal systolic LV rotation. Studies have shown that the diastolic untwisting predominantly occurring during isovolumic relaxation is a major determinant of LV filling properties,1,2,9 and that quantification of untwisting and of the untwisting rate should provide novel markers for the assessment of LV relaxation. Another limitation of the present study is the small number of diseased dogs, precluding an accurate estimation of the correlation between LVtor and hypokinesia assessed by conventional echocardiography and TDI. Another limitation also is that the timing of specific events such as aortic valve closure and opening was not assessed concomitantly with rotation STE profiles. This assessment would have been useful to compare the timing of myocardial torsional deformation with that of hemodynamic events. Moreover, because timing of the peak basal rotation was not temporally concordant with that of apical rotation (which occurred sooner), the LVtor results presented here might differ to a certain extent from the actual LVtor, and such differences between LVtor measured by STE and the actual LVtor would probably be higher in diseased dogs with myocardial dyssynchrony. Last, STE in itself exhibits several technical limitations. One of the problems associated with STE when using short-axis images (as in the present study) is that longitudinal myocardial motion causes speckles to move in or out of the image plane, thereby decreasing the reliability of the speckle tracking process,11 particularly at the base, because the amplitude of basal myocardial motion is higher than that of the apex.21 Moreover, in the present report, reliable rotation measurements could not be performed in all segments, particularly at the base (12% of noninterpretable segments because of unreliable traceable speckles). Similar failure to obtain reliable basal rotation already has been described in human patients.11,25 In 1 study,25 reliable LVtor measurement was possible in only one-third of patients, and this low feasibility was largely a result of the failure to obtain reliable basal rotation values, particularly in LV anterolateral and inferolateral myocardial segments because of the dropouts and reverberation artifacts frequently observed in these myocardial zones.

In conclusion, the systolic LV wringing motion can be noninvasively quantified in real time from basal and apical recordings by automatic speckle tracking, with good repeatability and reproducibility in the awake dog. The combined use of these novel STE indices with other systolic parameters such as those obtained using Doppler-based techniques (TDI, strain, and strain rate imaging) offers a new ultrasound approach to assessment of systolic function in this species, and may complement conventional echocardiographic measures of myocardial performance. This simple and rapid noninvasive assessment of LV twisting motion could be used for both clinical and research assessment of the failing heart. Investigations therefore are now warranted in diseased dogs to determine the relevance of each STE variable and identify suitable cut-off values for their correct interpretation.

Footnotes

aVivid 7, GE Healthcare, Waukesha, WI

bEcho Pac PC 6.3 software for Vivid 7, GE Healthcare

cM7S, M5S, M3S probes, GE Healthcare

dSystat version 10.0, SPSS Inc, Chicago, IL

Ancillary