In silico assessment of aortic hemodynamic sensitivity to inlet boundary conditions: Comparative analysis of 4D MRI,...

The use of generic waveforms as inlet boundary conditions (BCs) in computational fluid dynamics (CFD) simulations of the aorta greatly expands the applicability of CFD beyond the limited number of cases with patient-specific 4D MRI datasets. However, the validity of this approach has not been fully investigated.

This study evaluates the sensitivity of five hemodynamic indices in the thoracic aorta-time-averaged wall shear stress (TAWSS), oscillatory shear index (OSI), relative residence time (RRT), cross-flow index (CFI), and normalized topological shear variation index (normalized_TSVI)-to the choice of inlet BC. Simulations are performed for six patients with aortic root aneurysms using four different inlet conditions: spatiotemporal 4D MRI data (S1), patient-specific flow waveforms (S2), and two modified generic waveforms (S3 and S4), with S3 statistically derived and S4 extracted from a single patient. Hemodynamic metrics are analyzed at nodal resolution and through circumferential averaging at 5 mm intervals along the aorta.

Accurate simulation in the ascending aorta (AA) requires full spatial and temporal inlet resolution (errors: >100% nodal, >40% circumferential). In contrast, the descending aorta (DA) shows lower sensitivity to inlet BCs. In the absence of 4D MRI data, S2 is the most reliable surrogate. Among the generic waveforms, the statistically derived S3 consistently outperforms S4 in the DA. With S3, TAWSS and OSI remain within 25% of 4D MRI values, both nodally and circumferentially. RRT shows moderate sensitivity: regional values are acceptable, but nodal and circumferential errors exceed 100% and 30%, particularly in the proximal DA. CFI and normalized TSVI are highly sensitive to inlet conditions and remain unreliable without 4D MRI data.

Full spatiotemporal inlet resolution is essential for accurate CFD simulations in the AA. In the DA, statistically derived generic waveforms such as S3 provide practical alternatives for evaluating robust metrics like TAWSS and OSI. However, sensitive indices such as RRT, CFI, and TSVI still necessitate patient-specific data. We recommend developing S3-like waveforms using larger datasets to further improve the accuracy and clinical utility of CFD in the absence of 4D MRI data.