General guidelines for the performance of viscoelastic property identification in elastography: A Monte‐Carlo analysis from a closed‐form solution

• Published in: International journal for numerical methods in biomedical engineering Vol. 39; no. 8; pp. e3741 - n/a
• Main Authors: Van Houten, Elijah, Geymonat, Giuseppe, Krasucki, Françoise, Wattrisse, Bertrand
• Format: Journal Article
• Language: English
• Published: Hoboken, USA John Wiley & Sons, Inc 01.08.2023; Wiley Subscription Services, Inc; John Wiley and Sons
• Subjects: Bioengineering; Biomechanics; closed‐form viscoelasticity; Damping ratio; Elasticity; Elasticity Imaging Techniques - methods; elastography; Engineering Sciences; Exact solutions; Imaging techniques; Least squares; Least-Squares Analysis; Life Sciences; Magnetic resonance; Magnetic resonance imaging; Mechanical properties; Mechanics; Minima; Monte‐Carlo; Motion; Objective function; Phantoms, Imaging; Ultrasonography; viscoelastic waves; Viscoelasticity; Viscosity; Wave equations; Monte-Carlo; closed-form viscoelasticity; elastography; viscoelastic waves
• ISSN: 2040-7939; 2040-7947
• DOI: 10.1002/cnm.3741

Identification of the mechanical properties of a viscoelastic material depends on characteristics of the observed motion field within the object in question. For certain physical and experimental configurations and certain resolutions and variance within the measurement data, the viscoelastic properties of an object may become non‐identifiable. Elastographic imaging methods seek to provide maps of these viscoelastic properties based on displacement data measured by traditional imaging techniques, such as magnetic resonance and ultrasound. Here, 1D analytic solutions of the viscoelastic wave equation are used to generate displacement fields over wave conditions representative of diverse time‐harmonic elastography applications. These solutions are tested through the minimization of a least squares objective function suitable for framing the elastography inverse calculation. Analysis shows that the damping ratio and the ratio of the viscoelastic wavelength to the size of the domain play critical roles in the form of this least squares objective function. In addition, it can be shown analytically that this objective function will contain local minima, which hinder discovery of the global minima via gradient descent methods. A Monte‐Carlo based analysis of the 1D viscoelasticity inverse problem.