44 - "Enabling the Medical Digital Twin by advanced mesh morphing and high fidelity patient-specific simulations"
Abstract:
The in-silico approach is now well-established in various healthcare applications, including cardiovascular and human airways models. High-fidelity simulations enable the prediction of the flow field, interactions between moving walls and the flow, as well as interactions with particles. While the readiness of the in-silico approach is currently at a good level, there is still a need for the implementation of tools to integrate such approaches into daily clinical use. This would facilitate the creation and easy accessibility of a digital twin, i.e., a digital replica of the patient to be treated, by the medical staff.
A potential pathway to enable medical digital twins involves the extensive use of upfront high-performance computing (HPC) simulations to make high-fidelity results accessible in real time. The adopted strategy involves conducting a comprehensive scan of numerous variations (in the range of hundreds to thousands), compressing the results using techniques such as Singular Value Decomposition (SVD) and Proper Orthogonal Decomposition (POD), correlating input parameters with a reduced model through Artificial Intelligence and Machine Learning (AI-ML), and subsequently engaging in real-time interaction with the reduced model.
These variations could pertain to the patient, providing the ability to quickly personalize a treatment for a specific patient. Alternatively, variations could be related to the treatment itself, allowing, for instance, parametric surgery using patient-specific data to inform the procedure.
The adopted approach is comprehensively explained, illustrating how advanced mesh morphing based on Radial Basis Functions can be applied in two typical applications recently demonstrated within the framework of the EC project FF4EuroHPC. The first application, known as the DiTAiD experiment, involves the development of a parametric model of the human airways. This model can be swiftly adapted to a specific patient, facilitating the decision-making process for the most suitable treatment. The second application, named the Copernicus experiment, focuses on defining a virtual surgery for a cyanotic infant who is to undergo the Modified Blalock-Taussig Shunt procedure.
Short bio:
Dr. Marco Evangelos Biancolini is an associate professor of Machine Design at the University of Rome "Tor Vergata". His primary research interest lies in the engineering applications of fast radial basis functions, mesh morphing, and their application in the field of shape optimization, as well as reduced order models for the definition of digital twins. He is the owner and developer of the industrial software RBF Morph, a partner of Ansys, and an honorary member of the Technet Alliance. Dr. Biancolini serves as the principal investigator of the 4-year ITN-EID H2020 project "Medical Digital Twins for aneurysm prevention and treatment (MeDiTATe)" since 2020.
