Three-dimensional Super-resolution Left Ventricular Vector Flow, Pressure, & Clotting Risk Mapping by Multi-Physics-Informed Neural Network*

Color-Doppler echocardiography and cardiac magnetic resonance imaging (MRI) are widely used to assess left ventricular (LV) flow. However, 2D echo is limited to the velocity component parallel to the ultrasound beam and 4D Flow MRI only provides velocity fields in limited planes. Vector Flow mapping (VFM) technique and its 3D variant infer missing velocity component, but they rely solely on mass conservation and are sensitive to truncated data. Moreover, VFM does not compute pressure or clotting risk; those involve secondary analyses complicating clinical translation.

 

We present AI-VFM, a multi-physics-informed vector flow mapping method that applies artificial intelligence (AI) to partial or complete clinical imaging data. Its underlying models are continuity, Navier-Stokes, de-aliasing, and transport equations to infer flow fields and clotting risk. We analyze AI-VFM on CFD-generated ground-truth data vs. imaging parameters like spatial and temporal resolution, probe angle, aliasing, and Doppler sector size. We apply AI-VFM to clinical Color-Doppler sequences and 4D flow MRI to reconstruct three-dimensional super-resolution LV velocities and pressure maps.