#2795 In-silico assessment of arteriovenous fistula maturation with patient-specific computational hemodynamic models

Abstract

Abstract Background and Aims Native arteriovenous fistula (nAVF) is the preferred access for haemodialysis (HD) treatment. Both the quality and complications of the haemodialysis technique are closely related to vascular access. nAVF maturation process includes several morphological changes affecting haemodynamics, with previous studies suggesting that it is highly related to AVF failure. Computational Fluid Dynamics (CFD) simulations by using idealized morphologies or the analysis of one patient-specific geometry have been proposed to understand the pathophysiology of maturation failure. However, it is difficult to correlate with clinical practice and therefore limits its use. The aim is to develop a patient-specific modelling time pipeline, incorporating the haemodynamics of the blood flow to real anatomical and morphological characteristics. Method Design: prospective computational simulation analysis of nAVF in 3 patients at 3 different stages: after fistula creation, (1 week), maturation (1 month) and post-maturation (6 months). Analysis: morphological and haemodynamic. Inclusion criteria: >18 years and end-stage chronic kidney disease. Variables and data: Once morphology was extracted from nAVF magnetic resonance imaging, generated meshes were refined and compiled to obtain a volumetric mesh. The boundary conditions of the model were defined and extracted using ultrasound nAFV data. Data were analyzed at the 3 stages. Models: A 3-element Windkessel model was used to simulate the pressure waveform to feed the vein of the system. In this model, clinical data were also considered to classify the patient as normotensive or hypertensive to estimate waveform parameters. The Windkessel model estimates wall shear stress (WSS), velocity and flow for all the points located in the mesh by making an analogy from an electric circuit resistance model. As validation, the flow obtained from the Windkessel model was compared with the one obtained from the ultrasound scan of the patient. Results Different temporal points were analyzed for each nAVF. Results from the simulations in Fig. 1, illustrating the evolution over time and development. The geometrical characteristics were found to significantly impact on flow dynamics. The area most affected by flow was the juxta-anastomotic section which showed higher magnitude distribution of WSS and velocity for all time-points. According to the analysis of the simulations, a perpendicular junction between the vein and the artery potentially resulted in less damage in this area (in WSS terms), being important to mitigate the AVF failure. At maturation stage (1 month) the analysis shows WSS distribution mainly located in the juxta-anastomotic section and decreased in magnitude regarding one week; therefore, this can be marked as the stabilization point for nAFV maturation. Conclusion