Robust geometrical parametrization of the lower limb for the digital twin of patients under compression therapy

Compression therapy is extensively used for the treatment of venous deficiency and edema in the lower limb. The applied compression improves fluid circulation, which reduces the interstitial fluid accumulation. However, the effect of compression garments on soft tissues is poorly understood due to the high variability among the patients. With the long-term goal of designing patient-specific compression therapy, this study aims to develop a real-time simulation tool, able to predict the mechanical response of soft tissues of the lower limb under compression.

For such purpose, the use of standard Finite Element Analysis (FEA) falls short, as typically large numerical models needed for such simulations imply high computational cost. For this purpose, a solution based on Reduced Order Modeling (ROM) is developed to compute, in real-time, the mechanical effects of compression therapy. More specifically, this work focuses on the geometrical parametrization of lower limbs and the definition of the ROM through an adaptive sampling, which reduces the computational cost of the offline stage.

The implemented ROM reduces the dimensionality of the problem, which enables to carry out simulation in real time (online stage). Moreover, the proposed adaptive-sampling method offers an acceptable trade-off between computational cost and accuracy on the offline stage. Once the ROM is developed, it is used to perform a sensitivity analysis, which gives a detailed insight on the mechanical effects in the soft tissues of the lower limb.

The proposed methodology to develop ROM of the lower limb serves as a valuable numerical tool to foresee the effects of the compression therapy on the lower limb.