Simulações computacionais do escoamento cardiovascular guiadas por ressonância magnética

Cardiovascular diseases represents today one of the leading causes of death worldwide. Thus, knowledge of the blood flow patterns in the human body is important for diagnosis and research of certain diseases. Using magnetic resonance (MRI), in vivo 3D blood flow patterns can be either measured directly using phase–contrast (PC) MRI or Fourier velocity encoding (FVE) MRI. On the other hand, blood flow patterns can be obtained from model-based computational fluid dynamics (CFD) calculations. PC–MRI suffers from long scan times, limited spatio–temporal resolution, partial–volume effects and low signal–to–noise ratio (SNR). Fourier velocity encoding (FVE) is a promising magnetic resonance imaging method for measurement of cardiovascular blood flow. FVE provides considerably higher SNR than phase contrast imaging, and is robust to partial–volume effects. FVE data is usually acquired with low spatial resolution, due to scantime restrictions associated with its higher dimensionality, (x, y, v, t), and provides the velocity distribution associated with a large voxel, but does not directly provides a velocity map. Since the acquisition time is a disadvantage of FVE, then, preferably, it should be acquired with rapid spiral trajectories. Finally, CFD provides arbitrarily high spatial and temporal resolution and reduced scan times, but its accuracy hinges on the model assumptions. The objective of this work is to present a method capable of integrate direct MRI (PC or FVE) measures inside a CFD solver on the way to reduce scan time in a clinical environment. This work presents two main original contributions: (i) a method to derive velocity maps with high spatial resolution from low spatial resolution FVE data; and (ii) a numerical framework for constructing a flow field that is influenced by both direct measurements (PC–MRI or FVE) and a fluid physics model.