This project aims at developing a multi-modality data fusion methodology for enhancing the accuracy of 4D Flow MR imaging of cerebral aneurysms. Obtaining reliable estimates of the hemodynamic parameters affecting cerebral aneurysm progression from in vivo imaging data will help in diagnostics and treatment decisions of these lesions. With the time-resolved phase-contrast MRI (4D Flow MRI) technique three-dimensional flow fields can be acquired in cerebral vessels, however the spatial and temporal resolution is insufficient for reliable assessment of the flow-derived metrics which trigger remodeling of the aneurysmal artery walls. Patient-specific computational and experimental models can provide superior resolution, but their accuracy depends on modeling simplifications and assumptions. We propose to overcome the limitations of existing flow quantification methods by developing an integrative multi-fidelity approach. In this approach, the flow in cerebral aneurysms will be imaged with 4D Flow MRI, simulated with patient-specific Computational Fluid Dynamics (CFD) and measured with Particle Image Velocimetry (PIV) methods. The velocity fields obtained with the different modalities will be merged by using a common mode decomposition methodology. A novel order reduction methodology will be used to identify the dominant hydrodynamic modes common across all three modalities. This approach is based on the notion that the common modes represent the governing flow physics while the remaining modes represent random and bias errors from imaging noise and limited resolution as well as modeling assumption errors and thus should be eliminated.
The aims of the proposed project are, therefore, to (1) Conduct error analysis of 4D Flow MRI velocimetry in comparison to CFD and PIV and (2) Integrate 4D Flow MRI, CFD and PIV modalities in order to obtain reliable assessment of the flow metrics related to aneurysm progression. This integrated multi-fidelity flow quantification framework will deliver the values and variances of clinically relevant flow metrics from in vivo 4D Flow MRI data. The successful completion of the project will enable the development of a clinical tool for predicting cerebral aneurysm growth and rupture on a patient basis and thereby for risk stratification of cerebral aneurysm patients. This project will be performed in close collaboration between the cardiovascular MR imaging group at Northwestern University, Feinberg School of Medicine and cardiovascular engineering group at Purdue University. This cross-disciplinary team will bring together experts in MRI velocimetry, patient-specific flow computations, and experimental fluid mechanics and uncertainty analysis. The superb medical imaging and engineering resources available at Northwestern University and Purdue University, existing data sharing agreement between the institutions, and ongoing collaboration between the PIs will ensure success of the proposed project.
Brain aneurysm stability is affected by local blood flow forces, which can trigger growth and rupture resulting in brain compression or stroke. The flow in cerebral aneurysms can be measured with 4D flow MRI, however, the accuracy of this method is inadequate for reliable estimation of the relevant flow descriptors. We propose to enhance the accuracy of the MRI blood flow assessment by developing a data augmentation approach, where MRI measurements are integrated with data from patient-specific computational and experimental models. This hybrid methodology deliver high-fidelity accuracy and reliability to evaluate clinically relevant hemodynamic parameters.
