Feasibility study of fluid-structure-growth modelling in the context of endogenous tissue restoration

In situ tissue engineering exploits natural inflammatory responses to turn a synthetic vascular graft into a regenerated, functioning, and living vessel through endogenous tissue restoration. Despite promising preclinical results, challenges remain, especially in high-pressure regimes. To better understand the underlying biological processes and predict long-term outcomes of graft implantation, we propose an in silico approach. We account for the solid mechanics of conduit grafts in high and low pressure regimes, the wall shear stresses, and long-term growth and remodeling through two reduced-order fluid-structure-growth models. The first employs an artificial neural network surrogate model to compute the time-averaged wall shear stress at every growth step, whereas the second evaluates a fully-coupled fluid-structure interaction model every 60 time steps. We thoroughly verify both models and confirm acceptable model-form errors, then calibrate them against data from animal experiments. A validation study comparing in silico predictions with unseen animal data of carotid and pulmonary conduit grafts confirms that the model remains reliable up to at least one month of prediction time, hence in the future enabling in silico optimization of short-term endogenous tissue restoration in tissue engineering vascular grafts.