Department of Electrical and Computer Engineering Ph.D. Public Defense
Dynamic Optical Coherence Elastography
Supervised by Professor Kevin J. Parker and Co-Supervised by Jannick P. Rolland
Thursday, December 12, 2019
Hopeman Building, Room 224
Dynamic optical coherence elastography (OCE) is an attractive functional imaging modality that leverages the propagation of mechanical waves for estimating mechanical properties of tissues since it does not require (1) a priori knowledge of force/stress, (2) mandatory direct contact between the mechanical loading source and the tissue under study, and (3) large deformation of tissues. Therefore, dynamic OCE shows promise for the non-contact, in situ, and in vivo mechanical characterization of tissues with applications in laboratory and clinical studies.
In this context, the impact of using transient/continuous temporal excitation produced by one/multiple loading sources on the effectiveness of OCE methods for the elastic characterization of tissues has not been yet fully explored. In addition, due to the small field of view of OCT (~ 2 mm along depth), the complexity of boundary conditions of tissues (such as heterogeneous, composite, or plate-shaped media), and excitation wavelength (typically between 0.1 mm 10 mm), surface acoustic waves (SAW) are the dominant perturbation, diminishing the capabilities of OCE methods in estimating depth- dependent tissue elasticity information (such as in layered or composite materials). Moreover, the estimation of viscous parameters in addition to the classic elastic modulus is of great interest since it can provide useful information of disease stages. However, most OCE methods assume a rheological model of tissues and utilize only frequency-dependent wave speed measurements disregarding valuable information given by the wave attenuation process. Finally, the exploration of novel dynamic OCE techniques to address the aforementioned issues, including the aberration of waves in heterogeneous and anisotropic media and the multiple wave reflections produced by irregular boundary conditions, is highly desired.
In this thesis, a comparative study of transient and continuous dynamic OCE methods and estimators when using one or two vibration sources is conducted. Resolution, accuracy error, precision error, and contrast-to-noise ratio are the selected metrics for evaluating the performance of each method in numerical simulations and experiments in tissue- mimicking phantoms.
Furthermore, a novel OCE method based on reverberant shear wave fields in elastic media is presented. This method arose as a response to the limitations found in the comparative study developed previously. Numerical simulations and experiments in ex vivo porcine cornea demonstrate the capabilities of the proposed technique for layered material elasticity characterization compared to SAW-based OCE methods.
Subsequently, a novel technique for the viscoelastic characterization of tissues by propagating Gaussian-shape transient waves using acoustic radiation force (ARF) excitation is developed. Experimental results in viscoelastic tissue-mimicking phantoms validated the proposed approach, accounting for dispersion, distortion, and attenuation of transient waves while not assuming any rheological model for tissue characterization.
In addition, a study on longitudinal shear waves generated using a disk-shaped glass plate for the elastic characterization of heterogeneous distributed media is presented. Numerical simulations and experiments in tissue-mimicking phantoms demonstrate the capabilities and difficulties of such waves in detecting vertically and horizontally distributed materials. Preliminary experiments in mouse brain tissue are conducted followed by a discussion on future implementations of this method for in vivo mouse brain elastography studies.
Finally, at the end of this thesis, the elastography of small substructures in rabbit cornea produced by the localized laser-induced refractive index change (LIRIC treatment) by using OCE techniques proposed before is explored. In addition, the study of anisotropic tissues (chicken tibialis muscle and porcine brain) using non-contact ARF excitation is conducted. In these studies, a transverse isotropic mechanical model is used and validated.
In summary, we have (1) conducted a series of studies using numerical simulations and tissue-mimicking phantoms to understand the nature, properties, and capabilities of mechanical waves for the elastography of tissues; (2) developed a set of novel dynamic OCE approaches that are capable of detecting subtle changes (>13% change of elastic properties) within tissues, and with a spatial resolution as low as 120 𝜇m along lateral axis and 55.5 𝜇m along depth axis; and (3) applied these approaches to the study of tissues with different boundary conditions (composite plate-shaped media such as cornea, mouse in situ brain tissue, etc. ) and diverse mechanical properties (viscoelastic, heterogeneous and anisotropic tissues).