Physics-informed neural networks for myocardial perfusion MRI quantification
This work provides a new method for more reliable quantification of myocardial perfusion parameters for clinicians, potentially improving diagnostic accuracy in cardiac imaging. It is an incremental improvement on existing fitting methods.
This paper addresses the challenge of quantifying kinetic parameters like blood flow from dynamic contrast-enhanced MR images, which is often limited by low signal-to-noise ratio and temporal resolution. The authors introduce Physics-Informed Neural Networks (PINNs) to fit observed perfusion MR data while adhering to multi-compartment exchange models, demonstrating an overall reduction in mean-squared error in silico compared to standard non-linear least squares fitting.
Tracer-kinetic models allow for the quantification of kinetic parameters such as blood flow from dynamic contrast-enhanced magnetic resonance (MR) images. Fitting the observed data with multi-compartment exchange models is desirable, as they are physiologically plausible and resolve directly for blood flow and microvascular function. However, the reliability of model fitting is limited by the low signal-to-noise ratio, temporal resolution, and acquisition length. This may result in inaccurate parameter estimates. This study introduces physics-informed neural networks (PINNs) as a means to perform myocardial perfusion MR quantification, which provides a versatile scheme for the inference of kinetic parameters. These neural networks can be trained to fit the observed perfusion MR data while respecting the underlying physical conservation laws described by a multi-compartment exchange model. Here, we provide a framework for the implementation of PINNs in myocardial perfusion MR. The approach is validated both in silico and in vivo. In the in silico study, an overall reduction in mean-squared error with the ground-truth parameters was observed compared to a standard non-linear least squares fitting approach. The in vivo study demonstrates that the method produces parameter values comparable to those previously found in literature, as well as providing parameter maps which match the clinical diagnosis of patients.