ALAIV: Assessing Lymphatic Flows through Artificial Intelligence Velocimetry

Fluid transport through the lymphatic network is vital for removing waste from tissues, coordinating immune responses, and maintaining fluid homeostasis in the body. Dysfunctional lymphatic flow with age leads to painful, incurable conditions like lymphedema and may also contribute to neurological disorders like Alzheimer's. Thus, accurate measurements of lymphatic flow are crucial for diagnosing lymphatic function and pathology. Current in vitro methods of measuring lymphatic flow dynamics are limited by using single artificially pressurized lymphatic vessels that do not mimic in vivo conditions, where multiple vessels coordinate their movement to transport fluid. Existing in vivo techniques that utilize particle tracking velocimetry (PTV) to quantify flow only offer sparse velocity measurements, limiting our ability to derive other hydrodynamic signatures of healthy and diseased flows, such as changes in wall shear stress and pressure gradients driving flow. In this talk, we present the capability of Artificial Intelligence Velocimetry (AIV) to integrate sparse velocity measurements with physics-informed neural networks to robustly quantify in vivo lymphatic flows and evaluate lymphatic function. We demonstrate that AIV provides high-resolution 2D velocity, pressure, and wall shear stress in lymphatic vessels of both young and aged mice. The efficacy of this approach is evidenced by training on a fraction of in vivo Particle Tracking Velocimetry (PTV) measurements and demonstrating close agreement with the remaining measurements. AIV's capability to infer in vivo flows promises to enhance fluid dynamic models and explore dysfunctional lymphatic flows as a disease biomarker.