Magnetic Resonance Elastography of the in vivo Abdominal Aorta: A Feasibility Study for Comparing Aortic Stiffness between Hypertensives and Normotensives

Abstract

Purpose

The purpose of this study is to demonstrate feasibility of using magnetic resonance elastography (MRE) to identify hypertensive changes in the abdominal aorta when compared to normotensives based on the stiffness measurements.

Methods

MRE was performed on 8 volunteers (4 normotensives and 4 hypertensives) to measure the effective stiffness of the abdominal aorta. MRE wave images are directionally filtered and phase gradient analysis was performed to determine the stiffness of the aorta. Student’s t-test was performed to determine significant difference in stiffness measurements between normotensives and hypertensives.

Results

The normotensive group demonstrated an average abdominal aortic stiffness of 3.7 ± 0.8 kPa, while the controlled-hypertensive demonstrated an average abdominal aortic stiffness of 9.3 ± 1.9kPa. MRE effective stiffness of abdominal aorta in hypertensives was significantly greater than that of normotensives with p=0.02.

Conclusion

Feasibility of in vivo aortic MRE is demonstrated. Hypertensives have significantly higher aortic stiffness assessed through MRE than normotensives.

Introduction

Hypertension affects over 140 million people in North America and is one of the most important risk factors in the development of cardiovascular diseases leading to premature death (1–3).

Aortic stiffness (sometimes referred to as arterial stiffness) is a predictor of cardiovascular diseases, particularly in subjects with hypertension (4–6). Carotid-femoral pulse wave velocity (PWV) is widely used method to measure aortic stiffness (7–10). This technique requires measurement of peripheral wave velocity in peripheral arteries to calculate central pulse wave characteristics (11), which are indirectly used to obtain stiffness measurements (12). However, the peripheral pulse pressure is a poor reflection of central aortic pulse pressure (13,14). Therefore, there is a discrepancy between pulse readings at different sites (15).

Magnetic resonance elastography (MRE) is a novel imaging technique that can image the response of the tissue to externally-generated acoustic waves to obtain the intrinsic mechanical properties of the tissue (16–22). A phase contrast MRE sequence is used to synchronize external motion with motion sensitizing gradients in the phase of the MR images to obtain propagating waves. These wave images are then converted to stiffness map known as inversion (23,24). MRE has been shown to be clinically valuable in diseases such as liver cirrhosis where organ stiffness changes (16,25–27).

Our hypothesis is that MRE can be used to image early hypertensive changes occurring in the aorta which result in increased wall stiffness. The purpose of this study is to demonstrate feasibility of using MRE to identify hypertensive changes in the abdominal aorta when compared to normotensives based on the stiffness measurements and these MRE-derived stiffness measurements are not presented relative to a gold standard of direct mechanical measurement of the tissue which was not possible in our subjects.

Methods

Experiment Setup

In vivo aortic MRE was performed on 4 normotensives (25–45 years old Male) and 4 hypertensive (44 and 66 years old, 2 Females; 54 and 63 years old, 2 Males) volunteers with a long history of hypertension currently controlled on antihypertensive medications (Blood pressure in range of 118–127 systolic and 80–90 diastolic). Written consent was obtained from the volunteers with the permission of institutional review board. All imaging was performed in a 1.5-Tesla MRI scanner (Signa Excite, GE Health Care, Milwaukee, WI). The volunteers were positioned in the supine position and placed feet first into the scanner as shown in figure 1.

Figure 1.

Schematic showing MRE pneumatic passive driver placed on the abdomen of a volunteer. Sound waves are transmitted through a hollow plastic tube to the passive driver and into the volunteer’s abdominal aorta.

Image Acquisition

Single shot fast spin echo imaging was performed to obtain scout image in a sagittal plane. A gradient echo MRE sequence (28) was performed on this sagittal slice. Mechanical waves were introduced into the abdominal aorta by a pneumatic driver system as shown in figure 1. The passive driver was placed just inferior to the xiphisternum. Imaging parameters included TR/TE= 50,66,83/21,24,26 ms; FOV= 24,32,40 cm; α= 30°; slice thickness= 3,5,8 mm; acquisition matrix= 256x96; excitation frequency= 60 Hz (16.67 ms); 4 MRE phase offsets (i.e. the MRI acquisition is performed with multiple offsets of the phase of the externally applied wave to obtain images of the wave propagation over time; and the time shift for each phase offset is 4.17 ms (16.67/4)); and 16.67 ms duration (60 Hz) motion encoding gradients were applied separately in the x, y, and z directions to measure the in-plane and through plane motion. Additionally, the wave images were acquired without applying external motion, to verify that the wave images observed were from externally applied and not from physiologic motion. All images were acquired in free breathing.

Image Analysis

The sagittal images in all the volunteers were masked to obtain abdominal aorta for processing as shown in figure 2 indicated by red line. The x, y and z components of motion were first Fourier transformed in time to obtain the first harmonic displacement data and then filtered using 4^th order Butterworth band pass filter with cutoff values of 1 – 40 waves/FOV to remove longitudinal and high frequency displacements and additionally directionally filtered (23) in 8 directions to remove reflected waves. Then the filtered displacement data was analyzed using the phase gradient inversion algorithm (24) to obtain effective stiffness maps.

Figure 2. Normotensive Volunteer.

a) Sagittal magnitude image of the abdominal aorta indicated with red contour. b-e) Snap shot of four phases of the in-plane component of the propagating waves and f) Weighted stiffness map from 3 encoding directions with a mean stiffness of 4.2 kPa.

Statistical Analysis

Student’s t-test was performed on MRE-derived stiffness measurements to determine the significant difference between normotensives and hypertensives. Values were shown as mean ± standard deviation.

Results

The experimental results demonstrated that propagating mechanical waves could be visualized in the abdominal aorta in all the volunteers examined as shown in figures 2. When no external motion is applied no discernible waves were imaged indicating MRE is insensitive to the physiological motion of the aorta (not shown).

Figure 2 (a-e) shows an example of magnitude image of abdominal aorta with a red contour delineating aorta and the corresponding phase images of the in-plane component of propagating waves at four MRE phase offsets in one of the normotensive (i.e. normal) volunteer. Figure 2(f) shows the weighted stiffness map from 3 encoding directions using phase gradient inversion algorithm with a mean stiffness of 4.2 kPa.

Figure 3 (a-f) shows the magnitude image and phase images of propagating waves in the controlled hypertensive volunteer and the corresponding weighted stiffness map from 3 encoding directions using phase gradient inversion algorithm with a mean stiffness of 11.4 kPa.

Figure 3. Controlled Hypertensive Volunteer.

a) Sagittal magnitude image of the abdominal aorta indicated with red contour. b-e) Snap shot of four phases of the in-plane component of the propagating waves and f) Weighted stiffness map from 3 encoding directions with a mean stiffness of 11.4 kPa.

MRE effective stiffness of abdominal aorta in hypertensives was significantly greater than that of normotensives (Figures 4a) with p=0.02. Figure 4b shows the plot of mean effective stiffness from all the volunteers in each group. The normotensive group demonstrated an average abdominal aortic stiffness of 3.7 ± 0.8 kPa, while the controlled-hypertensive demonstrated an average abdominal aortic stiffness of 9.3 ± 1.9kPa.

Figure 4.

a) Plot of MRE-derived aortic stiffness measurements for normotensives and hypertensives. Hypertensives have significantly stiffer aorta than normotensives. b) Plot of mean aortic stiffness measurement in normotensives and hypertensives. Normotensives have mean aortic stiffness of 3.7 ± 0.8 kPa, while controlled hypertensives have mean aortic stiffness of 9.4 ± 1.9 kPa.

Discussion

This study demonstrated feasibility of using MRE to identify hypertensive changes in the abdominal aorta when compared to normotensives based on the stiffness measurements and showed controlled hypertensives have significantly higher stiffness compared to normotensives.

In this study the external motion was applied to generate waves in the aorta. There was the possibility that the MRE scan could detect interference motion from the heart and lungs. MRE wave images were acquired with no external motion applied. When no external motion was applied, no discernible waves were seen. When external motion was applied we could observe discernible waves. This supports the fact that the motion sensitization of the MRE imaging filters out the extraneous motion noise.

There was also the possibility that the aortic wall would be too thin to detect on the MRE examination. However, the wave propagation was observed in the aorta because of the wave guided effect (22,29). When an aorta is vibrated, the aortic wall and the adjacent blood vibrate with the same frequency. Therefore, aortic wall and resulting motion of the adjacent blood in the lumen are used in the analysis of MRE to obtain the stiffness of aortic wall.

MRE method used in this study is different from all previous methods, in particular PWV. PWV indirectly measures aortic stiffness by assuming the ratio of aortic wall thickness and radius remains constant based on the Moens-Korteweg equation (12). The Moens-Korteweg equation states that PWV is proportional to the square root of the incremental elastic modulus of the vessel wall given constant ratio of the wall thickness to the vessel radius (30). Aortic wall thickness and radius varies from one location to the other in the aorta(31). However, clinically, PWV is measured how fast the wave travels a specified distance of the vascular bed, i.e. which requires measuring the length of the aorta and is subject to inter-observer variability, and is a limitation of PWV (7–10,31–33). The current study measures the stiffness based on the wavelengths of propagating waves without incorporating the thickness of the aorta. Other imaging strategies such as ultrasound (33,34) and MR (35,36) are in use to estimate the PWV to indirectly report the stiffness of the aorta. However, each technology applies completely different procedures to estimate the PWV and has its own limitations and therefore cannot be compared against each other and to the current study.

The MRE method used in this study also is different from MRE methods used in previous studies (22,29). Previously, MRE stiffness measurements in ex vivo models were reported by expressing stiffness in terms of product of Young’s modulus and wall thickness of the aorta. Therefore, the effective aortic stiffness reported in this study cannot be compared against previous studies’ results, (7–10,22,29,32) as they have implemented completely different procedure to estimate stiffness of the aorta. However, our new noninvasive measure of effective aortic stiffness has the advantage of not requiring measurement of aortic wall thickness.

There are several limitations to this study. First, the inversion used in this study is a two-dimensional (2D) inversion. Wave propagation within aorta is not planar and very complex associated with waveguide effects therefore a true three-dimensional (3D) inversion with the application of curl operator may offer further information to minimize the effects form waveguide. However, a 3D inversion requires a 3D volumetric data acquisition, which in turn requires a much longer scan time and becomes more difficult for the patient to tolerate. Therefore to incorporate a 3D inversion, data acquisition should be accelerated, so volumetric data can be obtained in a single breathhold. Furthermore, the obtained stiffness measurement for each volunteer is a measurement across the cardiac cycle, but it would be useful to be able to measure the stiffness in systole and diastole which might be possible with cardiac gating. Additionally, our stiffness measurement is a combination of intrinsic properties of the aorta plus luminal pressure, therefore the stiffness estimates are effective. The final limitation was that the patients were not age-matched in their comparison due to the small sample size. Despite of these limitations, we have observed significant aortic stiffness differences between controlled hypertensives and normotensives volunteers.

Future work will include implementation of 3D acquisition in the aorta during a single breathhold by applying parallel imaging acquisition strategies using gradient echo or echo planar imaging, which is feasible to acquire wave information in one encoding direction. This strategy is currently being implemented in liver MRE scans in our research studies. Further, we will apply cardiac gated cine MRE sequence to determine the stiffness of the aorta across the cardiac cycle, as aorta is a non-linear viscoelastic material and has influence of pressure (i.e. pressures changes from 80–120mmHg across the cardiac cycle) on stiffness. Future applications of in vivo aortic MRE will be investigated to diagnose different cardiac or aortic disease states, and more elaborate studies will be performed to attempt to separate the influence of load in measuring stiffness.

In conclusion, we have demonstrated feasibility of performing in vivo aortic MRE. We have also showed that aortic stiffness is significantly higher in controlled hypertensives when compared to normotensives. However, extensive work needs to be done to establish this truth.