O-022 CFD simulation for cerebral artery with patient-specific inflow conditions extracted from 4D-DSA

Introduction Hemodynamics is believed to be related to the pathology of cerebral aneurysms such as rupture and growth, so those phenomena have been investigated using computational fluid dynamics (CFD). In previous studies, we have used the same inflow condition to simulate blood flow regardless of patients’ differences, although the heartbeat cycle differed from patient to patient. In this study, we extracted the patient-specific heartbeat cycle and blood flow velocity from four-dimensional digital subtraction angiography (4D-DSA) data that contain spatial and temporal information of contrast media. Finally, we performed CFD simulations and evaluated the effect of patient-specific inflow conditions. Materials and methods Data of a 73-year-old man with hypertension were analyzed. We extracted the patient-specific heartbeat cycle and blood flow velocity from 4D-DSA data. Then, we performed CFD simulations using ANSYS CFX 18.1 under both general inflow conditions (healthy adult) and patient-specific inflow conditions extracted from 4D-DSA to reproduce the angiographic images. We calculated the cardiac cycle duration and the amplitude of voxel intensity values at one point on the vessel centerline over two cardiac cycles, for both the CFD simulation and 4D-DSA data sets. To assess the accuracy of the patient-specific flow condition extraction method, we compared the calculation results of the cardiac cycle duration and the amplitude obtained from both data sets. Finally, we compared the wall shear stress (WSS) differences caused between general inflow condition and patient-specific condition. Results The 4D-DSA analysis gave a heartbeat cycle duration of 1.11 s and an intensity amplitude of 0.856. The CFD simulation under general inflow condition gave a cycle duration of 0.825 s, and an intensity amplitude of 0.879. In contrast, the CFD simulation under patient-specific inflow condition gave a cycle duration of 1.03 s and an intensity amplitude of 0.854. The cycle duration error difference between 4D-DSA analysis and CFD simulation results under general inflow condition was 3.90%, and between intensities 25.3%. In contrast, the cycle duration error between 4D-DSA and CFD simulation under patient-specific condition was 0.945%, and the intensity amplitude error 7.24%. In addition, we found that the WSS value using general conditions was overestimated compared with that using patient-specific condition in this case (see figure 1). Conclusions We performed CFD simulation using patient-specific heartbeat cycle and blood flow velocity extracted from 4D-DSA data. The result showed that the CFD simulation under patient-specific inflow condition could simulate the actual blood flow more accurately. Disclosures H. Ohno: None. H. Takao: None. T. Suzuki: None. S. Fujimura: None. Y. Uchiyama: None. K. Tanaka: None. T. Okudaira: None. T. Ishii: None. K. Otani: None. T. Ishibashi: None. K. Fukudome: None. M. Yamamoto: None. Y. Murayama: None.