Table 2 Comparison of 4D Flow MRI Validation Methods

In-vivo

Fidelity with respect to clinical or research use of the method

Physiological variability, commonly a lack of gold standard

Verification that method works in vivo

[83,84,85,86]

Phantoms

Controllable, easier to get a gold standard than in-vivo studies, even simple experiments can be of value

Realistic models challenging to construct and control

Evaluation variables/parameters in a repeatable setting in real MRI hardware

[87,88,89,90,91, 168]

Simulations

Very controllable, different sources of error can be separated, underlying numerical velocity data serves as ground truth

Fidelity uncertain, computational cost

Rapid feedback during development, when the desired evaluation cannot be realized in a phantom setup or in vivo

[92,93,94, 169]