Author + information
- Received October 20, 2010
- Revision received January 20, 2011
- Accepted February 3, 2011
- Published online June 28, 2011.
- Mathieu Pernot, PhD⁎,†,⁎ (, )
- Mathieu Couade, MSc⁎,‡,
- Philippe Mateo, PhD§,
- Bertrand Crozatier, MD, PhD§,
- Rodolphe Fischmeister, PhD§ and
- Mickaël Tanter, PhD⁎,†
- ↵⁎Reprint requests and correspondence:
Dr. Mathieu Pernot, Institut Langevin, ESPCI, 10 rue Vauquelin, 75005 Paris, France
Objectives The goal of this study was to assess whether myocardial stiffness could be measured by shear wave imaging (SWI) and whether myocardial stiffness accurately quantified myocardial function.
Background SWI is a novel ultrasound-based technique for quantitative, local, and noninvasive mapping of soft tissue elastic properties.
Methods SWI was performed in Langendorff perfused isolated rat hearts (n = 6). Shear wave was generated and imaged in the left ventricular myocardium using a conventional ultrasonic probe connected to an ultrafast scanner (12,000 frames/s). The local myocardial stiffness was derived from shear wave velocity every 7.5 ms during 1 single cardiac cycle.
Results The average myocardial stiffness was 8.6 ± 0.7 kPa in systole and 1.7 ± 0.8 kPa in diastole. Myocardial stiffness was compared with isovolumic systolic pressure at rest and during administration of isoproterenol (10−9, 10−8, and 10−7 mol/l, 5 min each). Systolic myocardial stiffness increased strongly up to 23.4 ± 3.4 kPa. Myocardial stiffness correlated strongly with isovolumic systolic pressure (r2 = [0.94; 0.98], p < 0.0001).
Conclusions Myocardial stiffness can be measured in real time over the cardiac cycle using SWI, which allows quantification of stiffness variation between systole and diastole. Systolic myocardial stiffness provides a noninvasive index of myocardial contractility.
Different indexes of ventricular contractility have been proposed (1,2), but most of them are dependent on load conditions (3) or cannot be measured noninvasively. The importance of ventricular stiffness has been shown through the powerful concept of time-varying elastance formalized by Suga and Sagawa (4), which provides a comprehensive description of the link between ventricular stiffness and cardiac function. However, despite the clinical importance of this concept, there is currently no noninvasive method for measuring time-varying myocardial stiffness.
Two-dimensional echocardiography has become the imaging modality of choice for noninvasive evaluation of regional myocardial function. New imaging modes such as tissue Doppler imaging (5) and strain or strain rate imaging (6) have been shown to provide good indicators for evaluation of abnormal left ventricular (LV) function (7,8). However, myocardial strain imaging is highly load dependent (8) and is therefore difficult to use as an index of myocardial contractility. Another method proposed by Hsu et al. (9) consists of inducing remotely a local strain in the myocardium using acoustic radiation force to assess myocardial stiffness under the assumption of uniform stress at the focus.
In this study, we propose a novel approach for measuring the local myocardial stiffness. Our method is based on shear wave imaging (SWI), an ultrasound-based technique for mapping quantitatively the stiffness of soft tissues characterized by the Young modulus defined by the slope of the stress/strain curve (10). This technique belongs to the field of multiwave imaging because it combines 2 different waves: one (shear wave) providing stiffness contrast and another (ultrasound) providing millimeter spatial resolution (11). The clinical potential of this approach has been recently demonstrated in the field of breast lesion imaging (12), as well as in the liver (13) and arteries (14) and for monitoring of thermal ablation (15).
The first goal of this study was to demonstrate the feasibility of measuring the dynamics of myocardial stiffness with high temporal resolution and good reproducibility. The second goal was to demonstrate that systolic stiffness can be used as an index of contractility by comparing it with conventional contractility indexes in the normal heart.
Shear wave imaging
SWI (10,16) is based on the remote generation of shear waves in soft tissue by the acoustic radiation force at the focus of an ultrasound field. A short burst (300 μs) of focused ultrasound was transmitted by a diagnostic ultrasonic probe (linear array, 128 elements, 12-MHz central frequency, Vermon, Tours, France) to induce micrometric tissue displacements in a small zone of the myocardium as a result of the acoustic radiation force (Fig. 1A). This burst was composed of 3 transmissions of 100 μs each focused successively at 3 foci separated by 2 mm in depth. In response to that transient mechanical excitation, a shear wave was generated in the low kHz frequency range and propagated in the myocardium at velocities between 1 and 10 m/s, depending on tissue stiffness (Fig. 1B). The originality of SWI stems from imaging the shear wave propagation at an ultrahigh frame rate (12,000 images/s) using the same diagnostic probe connected to an ultrafast ultrasonic scanner (Aixplorer, SuperSonic Imagine, Aix-en-Provence, France). The whole acquisition of the shear wave propagation was performed within 5 ms. The raw data were transferred to a computer and processed offline. Tissue velocity maps were computed for each frame of the acquisition using in-phase and quadrature frame to frame cross-correlation. Myocardial wall motion was subtracted from the average wall motion during the acquisition, resulting in tissue motion solely induced by the shear wave (Figs. 1C to 1E). Shear velocity was computed at each depth of the image using the spatiotemporal data of the shear wave propagation. Finally, the shear modulus μ (i.e., stiffness) was derived at each location using the equation: where c is the shear velocity and ρ is the volumic mass of the tissue. A time-of-flight algorithm described in a previous study (13) was used to derive a complete stiffness map in the region reached by the shear wave.
Forty stiffness measurements were repeated at high rate over 2 cardiac cycles to investigate myocardial stiffness dynamics. Each stiffness measurement was achieved within 5 ms and was repeated every 7.5 ms (acquisition repetition frequency of 133 Hz). The ventricular pressure was recorded at the same time on an external analog to digital board (Usbamp, g.tec Medical Engineering GmbH, Schiedlberg, Austria), allowing synchronization of the acquisition in post-treatment.
SWI was performed in Langendorff perfused isolated adult rat hearts (heart/body weight ratio 3.7 ± 0.3 mg/g, n = 6). The isolated hearts were immersed into a saline bath (50 × 50 × 50 mm3), and the ultrasonic array was positioned through an acoustic window performed on the side of the water tank (Online Appendix). In all experiments, the probe orientation was long-axis view unless specified. The distance between the array and the heart was approximately 5 mm.
Ex vivo physiology
All experiments were carried out according to the European Community guiding principles in the care and use of animals (86/609/CEE). Ex vivo physiology experiments were performed in Langendorff perfused isolated rat hearts (n = 6) according to previously described methods (17). (Please see the supplemental methods section in the Online Appendix.) Pre-load was modified by varying balloon volume by 5-μl increments. Contractility was modified by changing Krebs extracellular calcium concentrations or by infusing increasing concentrations of isoproterenol.
Changes in myocardial stiffness were analyzed using a paired 2-tailed t test to evaluate the significance of the difference between individual mean values under different inotropic effects or pre-load conditions. Linear regression was used for correlation between systolic stiffness and contractility. Statistical significance was inferred for p < 0.05. Values are presented as mean ± SD.
Myocardial stiffness mapping
Figure 2A shows a typical stiffness map obtained at one period of the cardiac cycle. The maps were computed in the region reached by the shear wave, which was approximately 2 mm wide by 5 mm deep on each side of the pushing locations (Fig. 2A). A small region of interest of 1.4 mm (lateral) × 0.75 mm (depth) was chosen in the midwall region of the myocardium. In the following section, myocardial stiffness was averaged in the region of interest.
Myocardial stiffness dynamics
The variation of stiffness was measured, with good reproducibility in all animals. A typical time course of myocardial stiffness is shown in Figure 2B: a sharp stiffening (dμ/dtmax = 163 ± 21 kPa·s−1) is observed at the beginning of the systolic phase, reaches a maximum (μ = 8.4 ± 0.5 kPa), then decreases during end-systole and diastole, and finally returns to the initial state in the last part of the diastolic phase (μ = 1.2 ± 0.1 kPa). As shown in Figure 2B, the shape of the time-varying stiffness was similar to the LV isovolumic pressure curve, with stiffness peaks slightly preceding pressure peaks. In this study, we define the diastolic and systolic stiffness as the minimum and maximum values, respectively, of myocardial stiffness during a cardiac cycle. Measurement reproducibility was assessed for 5 acquisitions under the same experimental conditions for each heart. As shown in Figure 2, a good reproducibility was found for diastolic stiffness (SD 0.08 kPa), whereas the SD of stiffness was 0.5 kPa in systole, which corresponded to a maximum error of ±6%.
Pre-load dependence of myocardial stiffness
The variation of stiffness with pre-load was investigated by inflating progressively the latex water-filled balloon inserted into the LV. Figure 3 shows diastolic and systolic myocardial stiffness as a function of LV systolic pressure in one control heart. Before inflation of the balloon, systolic pressure was 77.1 ± 5.9 mm Hg. The balloon was then progressively inflated by 5-μl increments to a maximal volume of 70 μl, which increased systolic pressure to 132.4 ± 12.5 mm Hg owing to the Starling law. Systolic myocardial stiffness was found to increase by approximately 20%, from 10.1 ± 0.6 kPa to a maximum of 12.3 ± 0.7 kPa (p < 0.005). In contrast, diastolic stiffness did not change significantly with pre-load and was 0.98 ± 0.2 kPa. Similar results were found in all 6 hearts (Table 1). The average relative increase in systolic myocardial stiffness was 26 ± 4%. These results show that myocardial stiffness is relatively independent of pre-load over a large range of pressures. The effect of heart rate was also investigated (Online Appendix).
Calcium concentration dependence of myocardial stiffness
Figure 4 shows myocardial stiffness as a function of calcium concentration in Ringer solution. For each concentration, 3 LV volumes were investigated: the balloon was completely deflated (0 μl), inflated to 10 μl, and finally to 25 μl to study the dependence on pre-load. The increase in calcium concentration led to a strong increase in systolic stiffness, whereas the pre-load conditions did not significantly change the stiffness. Finally, diastolic stiffness did not vary significantly in any of these experiments.
Myocardial stiffness during beta-adrenergic stimulation
The variation of myocardial stiffness during isoproterenol stimulation is shown as a function of time in Figure 5. Systolic stiffness increased immediately after stimulation and followed the evolution of systolic pressure. Figure 6 shows myocardial systolic stiffness response at different concentrations of isoproterenol (10−9, 10−8, and 10−7 mol/l). Systolic stiffness was found to increase strongly upon administration of isoproterenol (up to 23.4 ± 3.4 kPa) and followed a similar increase as (dP/dt)max. Diastolic stiffness was unchanged in all hearts (not shown). The response of myocardial stiffness to isoproterenol was also evaluated in an LV hypertrophy model (Online Appendix).
To evaluate the ability of myocardial stiffness to reflect inotropic changes, systolic stiffness was compared with peak systolic pressure (Pmax) values. Under isovolumic conditions such as in the Langendorff model, Pmax gives directly access to an approximation of the end-systolic pressure-volume relation, which is a conventional index of contractility. Figure 7 shows the variation of systolic stiffness as a function of Pmax during isoproterenol stimulation in one heart. SWI was performed approximately every 20 s during 3 min to measure the change in stiffness during transitory response to isoproterenol stimulation. Myocardial stiffness appeared to be linearly related to Pmax in all hearts (0.94 < r2 < 0.98, p < 0.0001). Table 2 shows good reproducibility in the relationship between Pmax and stiffness in all 6 hearts. A mean slope of 0.22 ± 0.2 kPa/mm Hg and a mean intercept of −8.8 ± 0.3 kPa were found.
In this study, we showed the feasibility of measuring myocardial stiffness locally and its dynamics over the cardiac cycle using SWI in Langendorff perfused rat hearts. Myocardial stiffness was measured with good reproducibility (SD <6%). Peak systolic stiffness was found to be relatively independent of pre-load conditions but strongly dependent on inotropic state, as evidenced by the stimulatory effects of calcium or a beta-adrenergic agonist. A strong correlation was found between systolic myocardial stiffness and a reference contractility index. These results demonstrated that peak systolic myocardial stiffness has the potential to be used as an index of myocardial contractility.
Ideally, an index of contractility should only reflect the intrinsic contractile state of the myocardium and therefore be independent of load conditions. The end-systolic elastance has been shown to correctly describe the intrinsic contractile state, but its determination requires invasive catheterization for measurement of LV pressure. Therefore, a noninvasive technique that would allow evaluation of the time-varying stiffness or elastance would be of great interest. In this study, we proposed such a technique for local measurements of myocardial stiffness and its time-dependent variations. Similar to end-systolic elastance, we showed that end-systolic myocardial stiffness was relatively load independent and could be used as an index of contractility. The theoretical relationship between the Young's modulus and elastance has already been investigated by Sagawa et al. (18). In this study, we provided an experimental demonstration of this relationship. Because SWI can measure quantitatively the local myocardial stiffness in a small region of the ventricle (of millimetric dimensions), this technique has the potential of providing a regional index of contractility. However, further investigations are required to demonstrate this regional capability.
Contrary to strain-based techniques, this method allows direct measurement of the local stress/strain relationship in myocardium (i.e., myocardial stiffness). In the field of soft-tissue rheology, the assessment of stress/strain relationship of a biological tissue is necessary to fully characterize its intrinsic mechanical properties (19). This is even more important for cardiac applications because myocardial stiffness provides important insights into the intrinsic mechanical properties of the beating heart. Myocardial stiffness is relatively independent of external stress and accurately reflects myocardial properties in their active or passive states. Although previous attempts to estimate the regional elastance using strain imaging have failed because of the difficulty of measuring regional stress, our method may offer a novel tool for quantitative mapping of the regional elastance and myocardial contractility. Another advantage of this technique is its relative low sensitivity to motion artifacts as a result of the use of very high frame rates.
One limitation of our study is that it was performed ex vivo. The Langendorff set-up offered stabilized experimental conditions that allowed us to investigate the link between systolic stiffness and contractility. However, like any other conventional ultrasonic-based techniques, SWI could be fully implemented in vivo for noninvasive imaging of the human heart, and our efforts will focus in this direction. In vivo implementation will require the use of an ultrasonic probe dedicated to echocardiography, with the correct choice of frequency and geometric design for imaging through intercostal spaces. A lower imaging frequency (approximately 3 MHz) would be used for human heart imaging. We have already shown the technical feasibility of SWI in vivo in open-chest animals (20).
Another limitation is that local stiffness does not necessarily reflect the elastance and contractile function of the whole ventricle, which may vary regionally. As shown in this study, SWI can be used to derive a map of the local myocardial elastic properties in the region where shear waves propagate. It is possible to extend this region to map the entire ventricle by generating additional shear waves at different locations. The stiffness mapping of the entire ventricle could have important clinical applications for the diagnosis of cardiac pathologies with important heterogeneities of myocardial elastic properties, such as regional ischemia or dilated cardiomyopathy. Diastolic dysfunctions may also benefit from this technique by quantification of myocardial diastolic stiffness.
For supplemental Methods, figures, and a table, please see the online version of this article.
Mr. Couade is an employee of SuperSonic Imagine. Dr. Tanter is cofounder of SuperSonic Imagine. All other authors have reported that they have no relationships to disclose.
- Abbreviations and Acronyms
- left ventricle/ventricular
- peak systolic pressure
- shear wave imaging
- Received October 20, 2010.
- Revision received January 20, 2011.
- Accepted February 3, 2011.
- American College of Cardiology Foundation
- Little W.
- Su J.,
- Crozatier B.
- Suga H.,
- Sagawa K.
- D'hooge J.,
- Heimdal A.,
- Jamal F.,
- et al.
- Edvardsen T.,
- Skulstad H.,
- AaKhus S.,
- Urheim S.,
- Ihlen H.
- Fink M.,
- Tanter M.
- Bercoff J.,
- Pernot M.,
- Tanter M.,
- Fink M.
- Abi-Gerges A.,
- Richter W.,
- Lefebvre F.,
- et al.
- Sagawa K.,
- Maughan L.,
- Suga H.,
- Sunagawa K.
- Fung Y.