Enhancing the predictive capability of magnetic resonance imaging using medical data-supervised cardiovascular flow...

Detailed hemodynamic parameters are essential for managing cardiovascular diseases, as they reveal blood flow dynamics that affect disease progression and treatment. However, even such advanced techniques as 4D Phase Contrast MRI face challenges in providing accurate, high-resolution data due to limitations in spatial and temporal resolution and image artifacts. Computational Fluid Dynamics (CFD) can estimate these parameters theoretically, but patient-specific accuracy may be compromised due to assumptions in boundary conditions and material properties.

Here, we aim to circumvent current limitations in medical imaging and CFD simulations by creating a comprehensive cardiovascular analytics model informed by clinical data. We develop a patient-specific simulation framework by deriving critical geometric parameters, boundary conditions, and aortic wall material properties directly from medical investigation and imaging data. This detailed information is subsequently integrated into Fluid-Structure-Interaction simulations to predict such key hemodynamic indicators as pressure distribution, wall deformation, time-averaged wall shear stress and oscillatory shear index to better assess individual vascular health. This approach effectively links imaging technology with computational modeling, as evidenced from our findings based on the medical imaging data of a representative human subject.

The results reveal that such amalgamation of patient-specific parameters enhances the simulation's accuracy, offering a more comprehensive and precise assessment of cardiovascular health than the traditional generic approaches. This comprehensive framework thus has potential to become an invaluable clinical tool, enhancing the accuracy of hemodynamic assessment, moving toward more personalized care and informing effective treatment decision-making.