The goal of this project is to develop a method to identify the vulnerability of plaques in the human body. Rupture of vulnerable plaques is the main cause of acute myocardial and cerebral infarctions. The pathology in the carotid arteries is directly related to cerebral events and is a marker for the pathology in the coronary arteries. As soon as the vulnerability of plaques in the carotid artery can be assessed using non-invasive ultrasound, identification of people that are at risk of having a cardiovascular infarct might be feasible on a routine basis. Based upon a pilot study [1], a VIDI grant proposal was provided by NWO and STW for this project.
Background
Atherosclerosis is a systemic disease in which lipid rich content is deposited in the arterial vessel wall. Myocardial infarction and stroke are two leading causes of death [2;3]. The primary trigger for these two causes is destabilization of atherosclerotic plaques [4-6]. Although atherosclerosis is a systemic disease, plaques usually develop in the conducting arteries (e.g. the coronary, carotid and femoral artery). Plaques start off as so called ‘fatty streaks’ on the inner layer (the intima) of the arterial wall: small accumulations of atherogenic lipoproteins, macrophages, white blood cells and smooth muscle cells. When people age these ‘fatty streaks’ grow and can either develop into stable plaques or into vulnerable plaques. Both types of plaques are schematically shown in Figure 1. Stable plaques contain a lot of fibrous tissue and little or no lipid content. Vulnerable plaques generally consist of a large pool of lipids and thrombogenic material covered by a thin fibrous cap [7]. Stable plaques tend to obstruct the blood flow, whereas vulnerable plaques initially develop outwards and do not limit the blood supply [8]. Since the reduction in blood flow caused by stable plaques is usually accompanied by clinical symptoms like pain at the chest, patients with stable plaques often see a physician before the occlusion becomes too severe and intervention is still possible. Vulnerable plaques are however much more dangerous, because there are no clinical warning signs before the sudden rupture of the thin cap. When this happens the lipid content comes into contact with the blood flow and a thrombogenic reaction causes a sudden occlusion of an artery, which may cause a heart attack or stroke in case a coronary or cerebral artery is blocked, respectively. Early identification of vulnerable plaques is therefore of crucial importance to prevent morbidity and mortality.
Figure 1: Schematic representations of a healthy vessel, a vessel with a stable plaque and a vessel with an early vulnerable plaque.
At the moment plaque severity is usually graded by measuring the thickness of the inner two layers of an artery (intima-adventitia) at the level of a plaque using ultrasound. However, intima-media thickness does not provide information about plaque vulnerability, but only about the severity of the occlusion. A technique which has been shown to be useful for identification of vulnerable plaques in the coronary arteries is intravascular ultrasound strain imaging or IVUS Elastography [9-14]. It was found that vulnerable plaques revealed more spots of high deformation/strain, than stable plaques and furthermore clinical outcome was related to measured strains [13]. However, until recently this technique could only be applied intravascularly and therefore was applicable only to patients that already were at the catheterization lab. With recent improvements in ultrasound and strain imaging methods, the development of a non-invasive version of the intravascular coronary technique that can be applied to scan the carotid arteries for vulnerable plaques has come within reach. A non-invasive version of the technique will allow early patient-friendly screening for vulnerable plaques in large populations.
In general this project can be divided into two subprojects, both dedicated to early detection of vulnerable plaques by using non-invasive ultrasound strain imaging. The first subproject aims at developing and validating a technique to measure strains in transverse cross-sections of a carotid artery. The second subproject will focus on the measurement of shear strain in the outer layer of the vessel wall (the adventitial layer). A hypothesis on the role of shear strain in the development of these fatty streaks into vulnerable plaques has been proposed by our group [15].
Figure 2: Imaging planes and measured strains.
Ultrasound strain imaging.
Strain imaging is based on the principle that soft tissue deforms more than hard tissue when an external force is applied. To measure strain, ultrasound acquisitions are acquired before (the pre-deformation state) and after (the post-deformation state) applying force to the tissue. From cross-correlation of the pre- and post-deformation images, local tissue displacements are estimated (see Figure 3) [16]. The location of the peak of the normalized cross-correlation corresponds to the time shift between the two ultrasound signals, caused by the displacement of the tissue. The time shift is translated into a local tissue displacement for the tissue of interest. The tissue strains can be derived from these estimates by first order spatial derivation. It should be noticed that the data below is the raw radiofrequency ultrasound signal and not the envelope or DICOM data. The use of Rf data instead of envelope/DICOM data allows a more precise estimation of the peak of the cross-correlation function and therefore also of the displacements and strains. Commercial ultrasound systems do normally not allow export of the raw radiofrequency data, however at our department we have several ultrasound systems from Philips and Medison with dedicated hardware that also output the Rf data.
Figure 3: Principle of rf-based strain estimation.
Strain imaging for transverse cross-sections of carotid arteries.
Methods: Contrary to intravascular applications, the ultrasound beam and radial strain are not aligned in non-invasive, percutaneous acquisitions of the carotid artery (Figure 2). Therefore, the radial strain cannot be measured in all regions directly. However, by combining axial (along the ultrasound beam) and lateral strain (perpendicular to the ultrasound beam), reconstruction of the radial strain for the whole artery cross-section is feasible [1]. This is illustrated in Figure 4 for a homogeneous vessel mimicking phantom that underwent a pressure increase of 4 mmHg. A positive strain corresponds to an expansion of the tissue. A negative strain corresponds to compression. As can be observed the axial and lateral strain images are very hard to interpret, whereas the radial strain image makes much more sense, because the strain is circular symmetric as expected for a homogeneous vessel.
Rf-data were acquired using a Philips SONOS 7500 live 3D ultrasound system, equipped with an L11 (3-11 MHz) linear array transducer and rf-interface. A homogeneous vessel mimicking phantom with an inner radius of 1.5 mm and an outer of 8 mm was made of 20% gelatin with 1% SiC scatterers (9-15 μm). The vessel phantom was positioned in a water tank and connected to a water column to simulate different intraluminal pressures. Longitudinal and cross-sectional recordings were made.
Figure 4: Axial, lateral and radial strain image for a homogeneous vessel phantom with a concentric lumen undergoing an intraluminal pressure change. The strain images were derived from ultrasound Rf data without beam steering.
As can be observed in Figure 4 the amount of noise for the estimated lateral strain component is much higher than the amount of noise for the axial component. This is mainly due to the lower resolution and the lack of phase information in the lateral direction. This lateral noise also deteriorates the radial strain image. Therefore, we have developed several techniques that circumvent the requirement for lateral information by combining radio-frequency ultrasound data obtained at multiple insonification angles [17-19]. It is possible to insonificate the tissue at different angles without translating the ultrasound transducer by electrical steering of the ultrasound beam, see Figure 5.
Figure 5: Images of the carotid artery can be acquired from different angles, without translating the transducer. This technique is called beam-steering.
In the most recent version of this multi-angle technique [17], axial data from two large but opposite insonification angles are used to derive the lateral information, instead of directly using the non-steered lateral component. Again, this information is combined with the axial information that was obtained without beam steering to derive the radial strain. As can be observed in Figure 6, the obtained lateral strain component and the resulting radial strain component are of a much higher precision than those obtained without beam steering.
Figure 6: Axial, lateral and radial strain image for a homogeneous vessel phantom with a concentric lumen undergoing an intraluminal pressure change. The strain images were derived from ultrasound Rf data obtained at beam-steering angles of -30°, 0° and 30° using a projection method [17].
The improvement in strain imaging precision obtained by beam-steering can also be quantified by calculating the root mean square error (RMSE), elastographic signal-to-noise ratio (SNRe) and elastographic contrast-to-noise ratio (CNRe). As can be observed in Table 1 the RMSE (error) decreases and the SNRe and the CNRe increase, thus the increase in image quality obtained by beam steering is confirmed by these parameters.
To confirm that the technique also improves the image quality for vessels with an eccentric lumen and for heterogeneous vessels, two phantoms with eccentric lumens were created from various gelatin-agar solutions [18;20]. One of the phantoms had a softer layer inside which represented the soft lipid pool of a vulnerable plaque. The other phantom was homogeneous. Figure 7 shows the B-mode images of the phantoms, the theoretical radial strain images, and the radial strain images as constructed without and with the beam-steering method.
Figure 7: B-mode images of two pressurized phantoms with an eccentric lumen. The corresponding radial strain images obtained without and with beam-steering at -30°, 0° and 30° using a projection method [17], and the corresponding theoretical radial strain images.
This image clearly shows that strain images reveal other information than echo images, whereas there is no clear difference between the echo image of the homogeneous phantom and the two layers of the heterogeneous phantom, the strain images clearly show a difference. For the homogeneous phantom the largest strains are observed where the vessel wall is thinnest, because the pressure gradient is largest over there. For the heterogeneous phantom the highest strains are present where the soft region is located. Again it is proven that beam-steering provides more precise strain estimates than can be obtained without beam-steering. Especially at the 3 and 9 o’clock region, where lateral information contributes most, the beam steered strain images resemble the theoretical strain images much better than those estimated without beam steering.
References
1. Ribbers H, Lopata RG, Holewijn S, et al. Noninvasive two-dimensional strain imaging of arteries: validation in phantoms and preliminary experience in carotid arteries in vivo. Ultrasound Med Biol 2007;33(4):530-40.
2. Thom T, Haase N, Rosamond W, et al. Heart disease and stroke statistics--2006 update: a report from the American Heart Association Statistics Commit-tee and Stroke Statistics Subcommittee. Circulation 2006;113(6):e85-151.
3. Weintraub HS. Identifying the vulnerable patient with rupture-prone plaque. Am J Cardiol 2008;101(12A):3F-10F.
4. Carr S, Farb A, Pearce WH, Virmani R, Yao JS. Atherosclerotic plaque rupture in symptomatic carotid artery stenosis. J Vasc Surg 1996;23(5):755-65.
5. Davies MJ, Thomas AC. Plaque fissuring--the cause of acute myocardial in-farction, sudden ischaemic death, and crescendo angina. Br Heart J 1985;53(4):363-73.
6. Falk E, Shah P, Fuster V. Coronary plaque disruption. Circulation 1995;92(3):657-71.
7. Schaar JA, Muller JE, Falk E, et al. Terminology for high-risk and vulnerable coronary artery plaques. Report of a meeting on the vulnerable plaque, June 17 and 18, 2003, Santorini, Greece. Eur Heart J 2004;25(12):1077-82.
8. Pasterkamp G, Schoneveld AH, van der Wal AC, et al. Relation of arterial geometry to luminal narrowing and histologic markers for plaque vulnerability: the remodeling paradox. Journal of the American College of Cardiology 1998;32(3):655-62.
9. de Korte CL, Pasterkamp G, van der Steen AFW, Woutman HA, Bom N. Cha-racterization of plaque components using intravascular ultrasound elastogra-phy in human femoral and coronary arteries in vitro. Circulation 2000;102(6):617-23.
10. de Korte CL, Carlier SG, Mastik F, et al. Morphological and mechanical infor-mation of coronary arteries obtained with Intravascular elastography: a feasibility study in vivo. European Heart Journal 2002;23(5):405-13.
11. de Korte CL, Sierevogel M, Mastik F, et al. Identification of atherosclerotic pla-que components with intravascular ultrasound elastography in vivo: a Yucatan pig study. Circulation 2002;105(14):1627-30.
12. Ryan LK, Foster FS. Ultrasonic measurement of differential displacement and strain in a vascular model. Ultrasonic Imaging 1997;19:19-38.
13. Schaar JA, de Korte CL, Mastik F, et al. Vulnerable plaque detection with intravascular elastography: A sensitivity and specificity study. Circulation 2001;104(17):II-459.
14. Shapo BM, Crowe JR, Skovoroda AR, et al. Displacement and strain imaging of coronary arteries with intraluminal ultrasound. IEEE Transactions on Ultra-sonics, Ferroelectrics, and Frequency Control 1996;43(2):234-46.
15. Idzenga T, Pasterkamp G, De Korte C. Shear strain in the adventitial layer of the arterial wall facilitates development of vulnerable plaques. Bioscience Hy-potheses 2009;2(5):339-42.
16. Ophir J, Céspedes EI, Ponnekanti H, Yazdi Y, Li X. Elastography: a method for imaging the elasticity in biological tissues. Ultrason Imaging 1991;13:111-34.
17. Hansen HHG, Lopata RGP, Idzenga T, de Korte CL. Full 2D displacement vector and strain tensor estimation for superficial tissue using beam steered ultrasound imaging. Physics in Medicine and biology 2010;5(11):3201-18.
18. Hansen HHG, Lopata RGP, Idzenga T, de Korte CL. An angular compounding technique using displacement projection for non-invasive ultrasound strain imaging of vessel cross-sections. Ultrasound in Medicine and biology 2010;in press.
19. Hansen HHG, Lopata RGP, de Korte CL. Noninvasive Carotid Strain Imaging Using Angular Compounding at Large Beam Steered Angles: Validation in Vessel Phantoms. IEEE Trans Med Imaging 2009;28(6):872-80.
20. de Korte CL, Céspedes EI, van der Steen AFW, Lancée CT. Intravascular elasticity imaging using ultrasound: feasibility studies in phantoms. Ultrasound Med Biol 1997;23(5):735-46.