FlowVN Trained on a Single Dataset Enables Rapid Reconstruction of Highly Accelerated 4D Flow MRI Across Multiple Sites.

Key Points

• This research aims to evaluate the effectiveness of FlowVN in reconstructing 4D flow MRI from a single dataset across multiple locations.
• Developed FlowVN model trained on one dataset.
• Tested model on multiple site data to assess reconstruction quality.
• Evaluated image quality with high acceleration factors.
• FlowVN successfully reconstructs 4D flow MRI from undersampled data.
• Maintains excellent image quality even with very high acceleration rates.