This section provides an overview on the developments performed in elastography.
Various elastography approaches can be employed to investigate mechanical properties of tissues. In collaboration with team 5, magnetic resonance elastography (MRE) was developed to access shear storage and loss moduli of biological media. Two axes were particularly investigated. First, a conventional approach, using oscillating gradients for tissue-motion encoding as well as direct inversion method for moduli estimation was developed. The MRE technique was assessed by comparing the resulting moduli with those from high-frequency rheometry, a technique allowing for a wider range of testing frequencies than classical rheometers (Fig.1, [1]). Then, the use of optimal control theory to design specific RF pulses that directly encode (i.e. without oscillating motion-encoding gradients) tissue displacements in phase images was investigated and implemented [2]. To our knowledge, this is the first time that optimal control is applied to MRE, and an improvement in the phase-to-noise ratio of images was observed compared to conventional MRE [3].
Figure 1: (a) Example of wave images obtained with our magnetic resonance elastography (MRE) technique in a homogeneous plastisol phantom, for different mechanical excitation frequencies. Shear (b) storage and (c) loss moduli of three homogeneous plastisol phantoms (A, B, C) measured with MRE and high frequency rheometry (HFR) [1]. The results provided by the two techniques are relatively close.
Developments were also performed in quasi-static ultrasound elastography, which produces strain images of biological tissues when subjected to some compression. Processing 2-D radiofrequency ultrasound data can be a limitation for this application, since information like out-of-plane (or elevational) motion is inaccessible. To improve tracking of tissue displacements while preserving the flexibility of the ultrasound examination, a prototype transducer (5x128 elements) has been developed that captures three adjacent imaging planes, regularly spaced. The idea is to acquire the only necessary data to still produce 2D elastograms of a medium but resulting from a 3D motion estimation. Designed in collaboration with Vermon S.A. (Tours, France), this transducer was experimentally assessed, showing notably strain images of improved quality and associated with higher correlation coefficients (CC) achieved during elastogram computation than those obtained with the classic 2D approach. An example of results obtained with a phantom experiment is presented in Figure 2, where the probe was moved 0.8 mm in 0.1-mm steps in elevation only (i.e. no compression was applied). It is therefore expected that the estimates of axial strains will be close to 0%, and those of elevational displacements, close to the programmed values. Figure 2 shows that results from the proposed approach remain in agreement with the experiment whatever elevational displacement, whereas those from the 2D method are affected by decorrelation noise with increasing out-of-plane motion [4,5].
Figure 2: Results obtained during a phantom experiment. From left to right are the results for programmed elevational displacements from 0.1 to 0.8 mm in 0.1-mm steps. (a) Axial strains and (b) associated correlation coefficients resulting from the use of the 2D method. (c) Axial strains, (d) elevational displacements and (e) associated correlation coefficients obtained with the proposed approach.
Finally, as tissue mechanical properties are investigated for diagnostic purposes, clinical assessment of the developments performed appears as an essential task. Our compression-induced tissue strain estimation method involving constrained maximization of a similarity criterion (correlation coefficient) and local regularization was first assessed in a 35 breast lesions study (Hospices Civils de Lyon collaboration, [6]) and a new study including comparison with a shear wave ultrasound elastography technique is ongoing at the Centre LĂ©on BĂ©rard (cancer treatment center), Lyon, France. Figure 3 provides some illustrations of breast elastograms.
Figure 3: Examples of breast elastograms obtained on (a) cancers and (b) fibroadenomas.
[1] P. Lefebvre, K. Tse Ve Koon, E. Brusseau, S. Nicolle, J.-F. Palierne, S. Lambert, D. Grenier, âComparison of viscoelastic property characterization of plastisol phantoms with magnetic resonance elastography and high-frequency rheometryâ, in Proc IEEE Eng Med Biol Soc, 2016;1216â1219, Orlando, Etats-Unis, 16-20 AoĂ»t 2016.
[2] P. Lefebvre, E. Van Reeth, H. Ratiney, O. Beuf, E. Brusseau, S. Lambert, S. Glaser, D. Sugny, D. Grenier, K. Tse Ve Koon, âActive control of the spatial MRI phase distribution with optimal control theory,â J Magn Reson, 281:82â93, 2017.
[3] E. Van Reeth, P. Lefebvre, H. Ratiney, S. Lambert, M. Tesch, E. Brusseau, D. Grenier, O. Beuf, S. Glaser, D. Sugny, K. Tse Ve Koon, âConstant gradient elastography with optimal control RF pulsesâ, J Magn Reson, 294:153â161, 2018.
[4] E. Brusseau, C. Meynier, G. FĂ©rin, O. Basset, A. NguyenâDinh, "Method and system for biological tissue strain imaging with dedicated ultrasound probe", U.S. Non-Provisional Utility Patent Application, filed December 19, 2018 (application no. 16/225,783)
[5] E. Brusseau, A. Bernard, C. Meynier, P. Chaudet, V. Detti, G. FĂ©rin, O. Basset, A. Nguyen-Dinh, âSpecific ultrasound data acquisition for tissue motion and strain estimationâŻ: initial results,â Ultrasound Med Biol, 43:2904â2913, 2017.
[6] E. Brusseau, V. Detti, A. Coulon, E. Maissiat, N. Boublay, Y. BerthezĂšne, J. Fromageau, N. Bush, J. Bamber, « In vivo response to compression of 35 breast lesions observed with a two-dimensional locally regularized strain estimation method », Ultrasound Med Biol, 40:300â312, 2014.