LAWS Webinar 5

Short Early Career Talk (10 Minutes)

Advancing Cell Analysis: Leveraging Opto-Acoustofluidic Systems for Bioevent Monitoring

Dr. Luiz Vasconcelos

Mayo Clinic

Main Talk (30 Minutes)

Cortical Bone Assessment Using Ultrasonic Guided Waves: From Lab to Clinics

Dr. Jean-Gabriel Minonzio

Universidad de Valparaíso

Short Early Career Talk

Convolutional Neural Network Regression for Viscoelastic Parameter Estimation in Ultrasound Shear Wave Elastography

Ultrasound shear wave elastography (SWE) techniques have been very useful for the analysis of tissue rheological properties by evaluating induced shear wave propagation in a given tissue. Convolutional Neural Networks (CNNs) have been established as one of the main neural network architectures for image processing. Most implementations have relied on CNNs for their classification capabilities in both binary and categorical tasks. In this talk, I will demonstrate that CNNs are also capable of performing regression tasks based on simulated shear wave propagation images. Staggered-grid finite difference simulations based on a Kelvin-Voigt rheological model were used to generate wave motion images with shear elasticity values ranging from 1-25 kPa, shear viscosities ranging from 0-10 Pa·s. The CNN architecture was able to estimate viscoelastic parameters with mean absolute error of less than 0.079 kPa and 0.052 Pa·s, for elasticity and viscosity, respectively. The method evaluated might enable simpler and more reliable non-invasive evaluation of tissue injuries that alter rheological parameters such as liver and kidney fibrosis. These results are also an example of CNN’s exceptional regression capabilities, providing novel tools for continuous parameters estimation based on imaging inputs.

Dr. Luiz Vasconcelos

Dr. Luiz Vasconcelos, Ph.D. is a Postdoctoral Research Fellow in the Ultrasound Laboratory at Mayo Clinic in Rochester, MN, USA. He conducts research with Dr. Matthew Urban to develop novel ultrasound technologies and applications. His research interests include biomedical signal processing, data processing and machine learning for novel ultrasound elastography diagnostic applications, such as, detection of liver and kidney allograft rejection. He earned his doctorate degree in Bioinformatics and Computational Biology at the University of Minnesota and his bachelor’s in Electronic and Computational Engineering at the Federal University of Rio de Janeiro, Brazil.

Main Talk

Cortical Bone Assessment Using Ultrasonic Guided Waves: From Lab to Clinics
Osteoporosis is still a worldwide problem, particularly due to associated fragility fractures at the spine or hip. Patients at risk of fracture are detected using the current X-Ray gold standard DXA (Dual XRay Absorptiometry), based on a calibrated 2D image. However, a majority of patients are still difficult to classify correctly to this day. Different alternatives have been proposed, such as 3D X-Rays, Magnetic Resonance Imaging (MRI), or Ultrasound, the latter having advantages of being portable, without radiation, less expensive, and sensitive to mechanical properties.

Among ultrasonic approaches, Bi-Directional Axial Transmission (BDAT) has been used to classify between fractured and non-fractured patients firstly using classical ultrasonic parameters, such as velocities or cortical thickness and porosity, obtained from an inverse problem using the SVD-based method. The classification performance has not yet been clearly improved compared to the gold standard. Recently, novel parameters obtained from structural analysis-guided wave spectrum images (GWSI) have been introduced. Compared to inverse problems, limited by solution ambiguities, these parameters can be automatically calculated. The aim of this talk is to present the traveled paths from lab experiments and wave modeling to clinical measurements and application. Differences and similarities between Latam, US, and Europe will be discussed.

Dr. Jean-Gabriel Minonzio

Jean-Gabriel Minonzio (Member, IEEE) received a B.S. degree in engineering physics from the Ecole Supérieure de Physique et de Chimie Industrielles (ESPCI) de la Ville de Paris, Paris, France, in 2003, and the M.S. and Ph.D. degrees in physical acoustics from University Denis Diderot, Paris, in 2003 and 2006, respectively. He has been involved in the clinical measurement of guided waves in cortical bones with the Laboratoire d’Imagerie Biomédicale (LIB) associated with Sorbonne Université, CNRS, and INSERM and has been a Co-Founder of the startup Azalée, Paris. He is currently a Full Professor with the School of Informatics Engineering, Universidad de Valparaíso. In this context, he is also the Head of the Ph.D. Program in Applied Informatics, the Co-Head of the Center for Research and Development in Health Engineering (CINGS-UV), and a member of the Ph.D. Program Sciences and Engineering for Health. His main research interests include array signal processing, wave propagation modeling, and inverse problem in applied ultrasound. In 2021, he was also part of the Organizing Committee of the first IEEE Latin America Ultrasonics Symposium (LAUS) and the Latin America Outreach Initiative.