MulTiPLy

About 250,000 persons suffer from disorders of the lymphatic system in France, causing disabling limb swelling (lymphedema) whose treatment costs more than 100 million euros every year. There is no available treatment and it is only possible to attenuate lymphedema symptoms through massage sessions and compression therapies if lymphedema is detected early enough. If remained untreated or poorly managed, lymphedema can lead to severe complications, increasing the costly burden on the healthcare system with the need of prolonged care.

Personalized medicine in lymphology remains inexistent due to major challenges: complexity of the lymphatic system, multidisciplinary problems combining engineering, physiology and biology, large interindividual variability of lymphatic vessels, lack of reliable techniques to collect in vivo and in vitro data on interstitial fluid drainage to feed digital twins. To address these challenges, MulTiPLy will establish a ground-breaking digital tool combining neural networks and innovative sensors. For that, MulTiPLy gathers a multidisciplinary team built upon the complementarity of a strong consortium, gathering the major national experts in lymphology (IM2C, CHUR and CHUN), biomechanical modelling (SAINBIOSE) and biomedical sensors (INL). INL and SAINBIOSE are already collaborating within the framework of the I-Démo project “Symphonies” (2022-2025) which aims to develop a digital platform for lymphedema patients.

The objective of MulTiPLy is to elaborate the first multi-scale digital twin platform for lymphology, combining multimodal tissue imaging, machine learning and mechanistic computational modelling, aiming at long term to enable precision and personalized medicine in the management of lymphedema.

Our methodology will rely on 2 physics-informed neural network (PINN) models at 2 length-scales. A first PINN model of lymphatic capillaries at the cellular scale will simulate fluid transport, considering endothelial dysfunctions, pressures and tissue fibrosis. A second PINN model of soft tissue poroelasticity will simulate fluid and lymph drainage using a continuum approach. The two PINN models will be trained to satisfy the laws of physics and at the same time to match datasets acquired on every patient with different imaging modalities (multimodality): 4D MRI, electric impedance tomography, and tissue ultrasound elastography. A major effort will be devoted to build these datasets, especially at the cellular scale using lymphatic vessel on-chip experiments. This will especially require a technological breakthrough to enable high-resolution electric impedance tomography. The eventually coupled PINNs will consume available datasets to personalize the digital twin for further medical usage.

With more than 100 million euros every year spent by the French healthcare system, the societal and socio-economic impacts of lymphedema are massive. A significant gain would be achieved by reducing the frequency of medical consultations for patients with lymphedema thanks to the digital twin. Alerts generated by the numerical tool would be sent to the physician, who could adjust the patient’s treatment remotely. The goal is also to improve the patient’s daily life, autonomy, and consequently, their social

life.

Given the complexity of the lymphatic system, and globally of fluid drainage in the human body, and given the major role of this drainage for many diseases and for immunology in general, MulTiPLy will foster collaborations within the large community of digital modelling, offering perspectives for new problems with major potential impacts in human health and well-being. This will be highly valuable to innovate in a somehow forgotten pathology and to structure a larger community of research around this topic, involving not only academics but also stakeholders, particularly those in the socio-economic sectors.