Computational Haemodynamics in Cerebral Aneurysms: the Eects of Modelled versus Measured Boundary Conditions

A. MARZO; P. SINGH; I. LARRABIDE; A. RADAELLI; M. GWILLIAM; I. D. WILKINSON; P. LAWFORD; P. REYMOND; U. PATEL; A. F. FRANGI; D. R. HOSE

ANNALS OF BIOMEDICAL ENGINEERING

SPRINGER

Lugar: Berlin; Año: 2010 vol. 39 p. 884 - 896

0090-6964

Modeling of flow in intracranial aneurysms (IAs) requires flow information at the model boundaries. In absence of patient-specific measurements, typical or modeled boundaryconditions (BCs) are often used. This study investigates the effects of modeled versus patient-specific BCs on modeled hemodynamics within IAs. Computational fluid dynamics (CFD) models of five IAs were reconstructed from threedimensional rotational angiography (3DRA). BCs wereapplied using in turn patient-specific phase-contrast-MR (pc-MR) measurements, a 1D-circulation model, and a physiologically coherent method based on local WSS at inlets. The Navier–Stokes equations were solved using the Ansys-CFX software. Wall shear stress (WSS), oscillatoryshear index (OSI), and other hemodynamic indices were computed. Differences in the values obtained with the three methods were analyzed using boxplot diagrams. Qualitative similarities were observed in the flow fields obtained with the three approaches. The quantitative comparison showed smaller discrepancies between pc-MR and 1D-model data, than those observed between pc-MR and WSS-scaled data. Discrepancies were reduced when indices were normalized tomean hemodynamic aneurysmal data. The strong similarities observed for the three BCs models suggest that vessel and aneurysm geometry have the strongest influence on aneurysmalhemodynamics. In absence of patient-specific BCs, a distributed circulation model may represent the best option when CFD is used for large cohort studies.