CompMath Seminar - Aratz Garcia-Llona, Ecole des Mines de Saint Etienne

Title: Compressive Stocking optimization for Lower-limb lymphedema treatment

Abstract:

Keywords: Lymphedema, Reduced order model, Image segmentation, Finite Element Method

The lymphatic system is the responsible to collect the lymph and return them to the bloodstream. Lymphedema is a chronic disease that causes swelling in the body’s tissues due to accumulation of lymph, which usually takes places in the extremities. It may occur due to genetic reasons (primary lymphedema) or as a result of damage to the lymphatic system, such as after cancer treatment (secondary lymphedema). This research focuses on the treatment of lymphedema at lower-limb calf through compression stockings, which is a widely applied measure to reduce the volume of edemas. Despite effective at a clinical level, the use of off-the-shelf stockings with predefined sizes, reduces the efficiency at the patient-specific level. With the long-term aim of designing patient-specific stockings, this study aims to develop real-time simulation tool, able to predict the efficiency of a given compression stocking for a given patient, as parametrized using geometrical and material parameters. For such purpose, the use of standard Finite Element techniques falls short, as typically large models are needed (in terms of degrees of freedom and patient-specific parameters) which implies high computational cost. For this purpose, a solution based on Reduced Order Model (ROM) is developed to com- pute, in real-time, the tissue hydrostatic stress distribution at the location of the lymphatic system. Hydrostatic stress is a compressing pressure that squeeze lymphatic vessels (hypertension), which improves lymph circulation and increase the drainage capacity. Starting from CT-scan of patients, the contours of the leg, muscle and bones are defined to build the 3D geometry of the leg. The geometry is registered through a reduced number of parameters applying Free-Form Deformation (FFD) algorithm, which are the parameters or inputs of the ROM together with material properties and compressive pressure. The ROM requires to solve a high number of FE models defined through the mentioned parameters to create a training dataset. Such pre-processing step (offline stage) is carried out through Gridap, an open source finite element method (FEM) library written in Julia.

This method enables to design the most efficient compressive stockings considering the particularities of each patient, which improves the drainage of lymph at lower-limb. It provides the required compressive pressure applied by the stocking which is an importantfact to design the patient-specific stockings.