Cardiac MR Elastography: Comparison with left ventricular pressure measurement
© Elgeti et al; licensee BioMed Central Ltd. 2009
Received: 04 June 2009
Accepted: 09 November 2009
Published: 09 November 2009
Purpose of study
To compare magnetic resonance elastography (MRE) with ventricular pressure changes in an animal model.
Three pigs of different cardiac physiology (weight, 25 to 53 kg; heart rate, 61 to 93 bpm; left ventricular [LV] end-diastolic volume, 35 to 70 ml) were subjected to invasive LV pressure measurement by catheter and noninvasive cardiac MRE. Cardiac MRE was performed in a short-axis view of the heart and applying a 48.3-Hz shear-wave stimulus. Relative changes in LV-shear wave amplitudes during the cardiac cycle were analyzed. Correlation coefficients between wave amplitudes and LV pressure as well as between wave amplitudes and LV diameter were determined.
A relationship between MRE and LV pressure was observed in all three animals (R2 ≥ 0.76). No correlation was observed between MRE and LV diameter (R2 ≤ 0.15). Instead, shear wave amplitudes decreased 102 ± 58 ms earlier than LV diameters at systole and amplitudes increased 175 ± 40 ms before LV dilatation at diastole. Amplitude ratios between diastole and systole ranged from 2.0 to 2.8, corresponding to LV pressure differences of 60 to 73 mmHg.
Externally induced shear waves provide information reflecting intraventricular pressure changes which, if substantiated in further experiments, has potential to make cardiac MRE a unique noninvasive imaging modality for measuring pressure-volume function of the heart.
Today several cardiac imaging modalities are available that provide excellent information on cardiac morphology and tissue structure with high temporal resolution [1–4]. However, as these modalities cannot measure forces, insight into myocardial physiology, particularly initiation of contraction and relaxation, is limited. To obtain information on cardiac elasticity and contractility in a clinical setting, it is still necessary to perform invasive catheterization for direct measurement of pressure in the chamber of interest . Such information is very important as abnormal alteration of ventricular pressure between systole and diastole can sensitively indicate a variety of diseases such as hypertrophy or dysfunctional diastolic relaxation [5–7].
Recently cardiac magnetic resonance elastography (MRE)  was introduced as a noninvasive means for detecting relative changes in myocardial elasticity during the cardiac cycle [9–11]. This method employs low-frequency shear waves induced in the heart by an external harmonic vibrator and measures the wave amplitude variation (WAV). This WAV results from changes in myocardial stiffness during the cardiac cycle. A linear elastic model was used to transform WAV to elasticity ratios. To further exploit cardiac MRE, a linear pressure-stiffness model was developed to utilize WAV-MRE for noninvasive measurement of left ventricular (LV) pressure ratios. A study of eight healthy volunteers showed good correlation between relative pressure values derived by MRE and data available from the literature . While this correlation confirmed the validity of the assumptions made for WAV-MRE  in principle, MRE was not correlated with LV pressure because invasive examination of pressure-time function is precluded in healthy volunteers.
Therefore, we here present a study investigating the correlation between wave amplitudes in MRE and LV pressure determined by catheter measurement in pigs. As both experimental measures are expected to vary widely due to individual heart morphology and function three different animals showing significant differences in weight, age and physiologic cardiac parameters were investigated. Our intention was to show the evolution of geometry, pressure and shear stimulus of the heart under different physiological conditions applying neither elastodynamic assumptions nor geometrical models. WAV-time functions are deduced from raw wave-phase data by an algorithm introduced here that identifies regions of contracting tissue. Then, the correlation of WAV-MRE to LV pressure and geometry is analyzed by linear regression.
Animals and Anesthesia
All experiments were approved by the responsible authority. Three pigs (mini LEWE) with a body weight of 25-53 kg were used in this study. Before the experiments, anesthesia was induced by intramuscular injection of 15 mg/kg of ketamine hydrochloride (Ketamin® 100 mg/ml; Pfizer, Karlsruhe, Germany), 0.1 mg/kg droperidol (Droperidol 2.5 mg/ml; Hospira, Lake Forest, USA), and 0.2 mg/kg midazolam hydrochloride (Dormicum®; Hoffmann-La Roche AG, Grenzach-Whylen, Germany). To deepen anesthesia for intubation, propofol (propofol 1%, Fresenius Kabi, Bad Hombug, Germany) was administered into a lateral ear vein. After the anesthetics took effect, a 6.0-7.0 mm endotracheal tube (Mallinckrodt Laboratories, Athlon, Ireland) was inserted into the trachea to maintain anesthesia with a mixture of 2-3 vol% isoflurane (Forene; Abbott, Wiesbaden, Germany) and medical oxygen. Anesthesia was maintained using an electronic system for controlling ventilation and anesthesia (A.D.S. 1000, engler engineering corporation, Florida, USA). Directly prior to catheterization, 200 IU/kg heparin sodium (Liquemin® N 25.000, Hoffmann-La Roche AG, Grenzach-Wyhlen, Germany) was administered as an intravenous bolus. To maintain identical heart rates for LV pressure measurement and MRE experiments, the pigs received a continuous intravenous infusion of metoprolol tartrate (Beloc i.v., AstraZeneca GmbH, Wedel, Germany) at a rate of 50 ml/h.
Left Ventricular Pressure Measurement
The tip manometer (model PC-380, Millar Systems, Houston, Texas) was advanced into the heart via the right carotid artery using a guide-wire under fluoroscopic control. Tip manometer catheters were adjusted to atmospheric pressure prior to measurement. Pressure-time functions were stored together with electrocardiograph (ECG) signals in a digital oscilloscope (200 Hz sampling rate, 12 bit A/D converter). The time axis of the manometer data was subsequently shifted to match the R-waves of the MRE experiments using the ECG recordings.
Details of the acquisition protocol are described in [10, 11]. Standard short-axis cine SSFP images served to calculate LV volume and function in each animal. Directly after catheter intervention and recording of the LV pressure time course, MRE experiments were run in a clinical 1.5 T scanner (Magnetom Sonata, Siemens AG, Germany) using a standard 12-element body phased-array coil. A remote mechanical driver  was used to vibrate the left lateral chest of the animals at 48.3 Hz. An ECG-triggered, motion-sensitized, spoiled gradient echo sequence was run to acquire wave images with a frame rate of 193 Hz. (repetition time, TR: 5.18 ms; echo time, TE: 1.38 ms; flip angle, α: 15°, length of motion-encoding gradient (MEG): 2 ms; MEG amplitudes: 33 mT/m, 25 mT/m and 21 mT/m along the directions of read-out, phase-encoding and slice selection, respectively). Other imaging parameters were 200 × 400 mm field of view (FOV), 6 mm slice thickness in a short-axis view, and 96 × 128 matrix size. Segmented k-space acquisition was applied with one phase-encoding step per ECG trigger. Generalized autocalibrating partially parallel acquisition (GRAPPA; factor 2) was used. Two hundred eighty wave images were acquired in approximately 1 minute during ventilation stop. Left ventricular diameter was calculated from the magnitude images of the MRE sequence. Therefore the magnitude images of all three spatial directions are superimposed using custom-made software run on MATLAB (Version 7.0.4, The MathWork Inc.) and left ventricular diameter was measured repeatedly by a physician experienced in cardiovascular MR.
x is the location of a point within the ROI and ζ is an empirical factor between 0 and 1 that determines the optimal spatial extension of the ROI based on the point of maximum amplitude variation, . In the present study the WAV effect was evaluated for ζ = 0, 0.1, .1.0, and ζ = 0.9 was chosen after further discussion.
Physiological cardiac parameters.
LV ejection fraction [%]
LV end diastolic volume [ml]
The feasibility of amplitude-sensitive MRE for measuring tissue-elastic properties was demonstrated by phantom experiments for plane interfaces in  and for cylindrical inclusions in . This study compares, for the first time, MRE and cardiac pressure data. The synchronization of the MRE-derived wave amplitude alteration with invasively measured pressure changes corroborates the validity of a central assumption made previously in cardiac MRE: The change in shear wave amplitudes over the cardiac cycle is caused by elasticity changes in the myocardium and not driven by geometrical effects. Provided that stiffening and relaxation of the heart wall cause an immediate change in pressure, our results underscore that WAV-MRE is sensitive to LV pressure dynamics rather than to cardiac motion. An instantaneous response of LV pressure to myocardial contraction is well known from animal studies [14–16]. The comparison of WAV-MRE with LV pressure provides the missing link towards noninvasive assessment of cardiac pressure-volume variations.
The present study was focused on the principal relationship between MRE signal and invasively measured pressure. The most important result of this study is that there is a clear distinction between the coherences of WAV pressure and WAV geometry functions: While wave amplitudes vary (reciprocally) in synchrony with pressure changes (fig. 4) there is a clear delay between MRE and LV-diameter (fig. 5). Despite the small number of animals investigated this result is significant since the physiological differences seen in our vivo model covers a broad spectrum of pig heart physiology. Most noticeable, animal #1 was about twice the weight of the other two animals. However, Table 1 does not indicate a correlation between weight and heart rate, ejection fraction or LV volume. Instead, the increase in pressure parameters from animal #1 to #3 is reflected in the U(dia)/U(sys). Thus, the distinct offset of absolute wave amplitudes seen in animal #1 (Table 1, column U(systole)) is most probably attributable to different wave conditions resulting from the greater body extent imposing different wave conditions and not to effects of cardiac physiology. Yet, it is an encouraging result that this offset appears not to influence relative ratios U(diastole)/U(systole), which increase monotonically from animal #1 to #3. This observation suggests that WAV-MRE is a suitable technique for determining relative changes in left ventricular pressure. The diagnostic value of WAV-MRE has to be addressed in future studies. Additionally, ex vivo cardiac models (including tissue models) are needed for clarifying principles of shear waves in tissue exhibiting a wide variation in elasticity. In this respect, it is important to note that despite the good agreement in the timing of WAV and LV pressure changes, their dynamics are clearly different (see Figure 4). Differences in slopes and level times may result from nonlinear elastic effects in myocardial tissue. Moreover, the site from which LV pressure data are sampled relative to the position where the WAV effect is quantified may impact the coherence of both measures. The proposed method for selection of the appropriate ROI is a semi-automatic way to identify the maximum ratio U(diastole)/U(systole). It was found that the greatest WAV effect occurs in regions well exposed to shear vibrations in the vicinity of the mechanical transducer. Provided that the linear elastic model of Eq.A2 applies, the degree of amplitude variations at ζ = 0.9 indicates relative pressure ratios from 17 (animal #1) to 60 (animal #3). This large range is assumed to result from a forth-power law in Eq.A2 rather than reflecting physiological variability.
In summary, mechanical stimulation of the heart by low-frequency shear vibrations enables measurement of time-dependent ratios of wave amplitudes using cardiac MRE. We demonstrated that such ratios correlate with ventricular pressure and hence provide a noninvasive measure of ventricular pressure ratios. Our results suggest that the degree of wave amplitude alterations agrees with individual differences in cardiac physiology in terms of LV pressure variation.
Here, N is the number of points in U (35 in our experiments), and S is a step function defined on the basis of the duration of systole and diastole. becomes nonzero for points within contracting tissues and zero for points corresponding to noncontracting tissues. Therefore is an appropriate function for threshold-based segmentation of regions of different WAV characteristics. For example, based on the definition of S, the highest ratio U(diastole)/U(systole) is theoretically determined from the region of maximum . Such a region tends to be small, however, and because noise is present, averaging over such a small region may lead to inaccurate results. In principle, the proper region of interest should be determined based on the two competing factors of high values and large region size. For this purpose, we define the ROI within the heart such that its size and position are linked to by the factor ζ (Eq.1). Figure 1 illustrates that ζ = 1 results in a small ROI at maximum whereas ζ < 1 values cause the ROI to increase in size, reducing the effect of noise in the spatially averaged function U(t).
The authors acknowledge financial support from the German Research Foundation (DFG Sa 901/3) and would like to thank Bettina Herwig for language editing.
- Sengupta PP, Krishnamoorthy VK, Korinek J, Narula J, Vannan MA, Lester SJ, Tajik JA, Seward JB, Khandheria BK, Belohlavek M: Left ventricular form and function revisited: applied translational science to cardiovascular ultrasound imaging. J Am Soc Echocardiogr. 2007, 20 (5): 539-551. 10.1016/j.echo.2006.10.013.PubMed CentralView ArticlePubMedGoogle Scholar
- Axel L: Biomechanical dynamics of the heart with MRI. Annu Rev Biomed Eng. 2002, 4: 321-347. 10.1146/annurev.bioeng.4.020702.153434.View ArticlePubMedGoogle Scholar
- Finn JP, Nael K, Deshpande V, Ratib O, Laub G: Cardiac MR imaging: state of the technology. Radiology. 2006, 241 (2): 338-354. 10.1148/radiol.2412041866.View ArticlePubMedGoogle Scholar
- Beek AM, Bondarenko O, Afsharzada F, van Rossum AC: Quantification of late gadolinium enhanced CMR in viability assessment in chronic ischemic heart disease: a comparison to functional outcome. J Cardiovasc Magn Reson. 2009, 11 (1): 6-10.1186/1532-429X-11-6.PubMed CentralView ArticlePubMedGoogle Scholar
- Kass DA: Assessment of diastolic dysfunction. Invasive modalities. Cardiol Clin. 2000, 18 (3): 571-586. 10.1016/S0733-8651(05)70162-4.View ArticlePubMedGoogle Scholar
- Strobeck JE, Sonnenblick EH: Myocardial and ventricular function. Part II: Intact heart. Herz. 1981, 6 (5): 275-287.PubMedGoogle Scholar
- Mandinov L, Eberli FR, Seiler C, Hess OM: Diastolic heart failure. Cardiovasc Res. 2000, 45 (4): 813-825. 10.1016/S0008-6363(99)00399-5.View ArticlePubMedGoogle Scholar
- Muthupillai R, Lomas DJ, Rossman PJ, Greenleaf JF, Manduca A, Ehman RL: Magnetic resonance elastography by direct visualization of propagating acoustic strain waves. Science. 1995, 269 (5232): 1854-1857. 10.1126/science.7569924.View ArticlePubMedGoogle Scholar
- Rump J, Klatt D, Braun J, Warmuth C, Sack I: Fractional encoding of harmonic motions in MR elastography. Magn Reson Med. 2007, 57 (2): 388-395. 10.1002/mrm.21152.View ArticlePubMedGoogle Scholar
- Elgeti T, Rump J, Hamhaber U, Papazoglou S, Hamm B, Braun J, Sack I: Cardiac magnetic resonance elastography. Initial results. Invest Radiol. 2008, 43 (11): 762-772. 10.1097/RLI.0b013e3181822085.View ArticlePubMedGoogle Scholar
- Sack I, Rump J, Elgeti T, Samani A, Braun J: MR elastography of the human heart: noninvasive assessment of myocardial elasticity changes by shear wave amplitude variations. Magn Reson Med. 2009, 61 (3): 668-677. 10.1002/mrm.21878.View ArticlePubMedGoogle Scholar
- Klatt D, Asbach P, Rump J, Papazoglou S, Somasundaram R, Modrow J, Braun J, Sack I: In vivo determination of hepatic stiffness using steady-state free precession magnetic resonance elastography. Invest Radiol. 2006, 41 (12): 841-848. 10.1097/01.rli.0000244341.16372.08.View ArticlePubMedGoogle Scholar
- Papazoglou S, Xu C, Hamhaber U, Siebert E, Bohner G, Klingebiel R, Braun J, Sack I: Scatter-based magnetic resonance elastography. Phys Med Biol. 2009, 54 (7): 2229-2241. 10.1088/0031-9155/54/7/025.View ArticlePubMedGoogle Scholar
- Burkhoff D, Mirsky I, Suga H: Assessment of systolic and diastolic ventricular properties via pressure-volume analysis: a guide for clinical, translational, and basic researchers. Am J Physiol Heart Circ Physiol. 2005, 289 (2): H501-512. 10.1152/ajpheart.00138.2005.View ArticlePubMedGoogle Scholar
- Magorien DJ, Shaffer P, Bush CA, Magorien RD, Kolibash AJ, Leier CV, Bashore TM: Assessment of left ventricular pressure-volume relations using gated radionuclide angiography, echocardiography, and micromanometer pressure recordings. A new method for serial measurements of systolic and diastolic function in man. Circulation. 1983, 67 (4): 844-853.View ArticlePubMedGoogle Scholar
- Oh JK, Tajik J: The return of cardiac time intervals: the phoenix is rising. J Am Coll Cardiol. 2003, 42 (8): 1471-1474. 10.1016/S0735-1097(03)01036-2.View ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.