BMBF joint project MLgSA

Data- and simulation-based exploration, analysis and treatment of vasoconstrictions for the prevention of ischemic strokes.

Period: 15.09.2020- 14.09.2023

Status: Completed

Web: Collaborative project MLgSA

Project in Sciport

The aim of the joint project MLgSA is the development of informatic and mathematical analysis tools for the improvement of clinical treatment pathways in the exploration, analysis and treatment of vascular stenoses in order to reduce the resulting risk of ischemic stroke. To this end, we are developing new methods of blood flow simulation and machine learning as well as new optimization methods for fluid-structure interaction (FSI). Based on acquired image data, we will generate a 3D model of the carotid artery. Based on this, the aim is to prepare the model for blood flow simulation and to develop a program that allows the 3D model to be explored with the simulated blood flow data on the one hand, and to obtain a recommendation for surgical intervention using machine learning methods on the other. We will implement the latter with our clinical partners, who will provide us with a large number of clinical data sets which - supplemented by the simulation-based data - will form the basis for categorizing stroke and non-stroke patients. This is the basis for calculating a classifier.

Subproject 4 Hemodynamic modeling

Project leader Prof. Dr. Anna Hundertmark

The subproject of the University of Koblenz-Landau at the Landau campus develops numerical 3D models and simulation tools for hemodynamic flows in the carotid artery including fluid-structure interaction for non-Newtonian fluids. The simulation results complement the patient data and are exported for the training of ML algorithms. In addition, the numerical simulations are used in cooperation with the subproject at the Koblenz campus for the optimization methods for fluid-structure interaction problems.

MLgSA, subproject 4 in Sciport

Publications

events

  • ICIAM Congress in Tokyo 2023: Minisymposium with the title Mathematical modeling, simulation and optimization in stroke risk assessment (Hinze/Hundertmark)
  • TES Conference on Mathematical Optimization for Machine Learning, Berlin 2023: Minisymposium entitled Mathematical modeling, simulation and optimization in stroke risk assessment.

Conference contributions

  • GAMM Annual Meeting 2022: Numerical modeling of patient-based carotid artery flow with fluid-structure-interaction: hemodynamical risk parameters and compliance study (Richter)
  • Compact Seminar Numerics, JGU Mainz, 2022: Patient-based carotid artery flow and compliance modeling: hemodynamical risk evaluation using longitudinal shear stress (Hundertmark)
  • ICIAM 2023, Tokyo: Hemodynamic modeling of directional shear risk metrics in the carotid artery (Hundertmark)
  • ICIAM 2023, Tokyo: Physiological flow simulations for stroke assessement and importance of distensiblity (Richter)
  • TES Conference on Mathematical Optimization for Machine Learning, Berlin 2023: Modeling of carotid artery hemodynamics towards a CFD database for stroke risk assessment (Hundertmark)
  • Comsol Conference 2023, Munich: Utilizing COMSOL® in a Workflow to Assess Stroke Risks in a Large Set of Patient's Carotid Arteries, Poster Presentation (Probst)

Employees

Prof. Dr. Anna Hundertmark

Institute Director
Head of working group

AG Numerical Simulation

a.hundertmark[at]rptu.de
+49 (0)6341 280-34263

Read more

Kevin Richter

Former employee
Project MLgSA

k.richter[at]rPTU.de

Read more

Tristan Probst