Deep learning-based prediction of kinetic parameters from myocardial perfusion MRI

摘要: The quantification of myocardial perfusion MRI has the potential to provide a fast, automated and user-independent assessment ischaemia. However, due relatively high noise level low temporal resolution acquired data complexity tracer-kinetic models, model fitting can yield unreliable parameter estimates. A solution this problem is use Bayesian inference which incorporate prior knowledge improve reliability estimation. This, however, uses Markov chain Monte Carlo sampling approximate posterior distribution kinetic parameters extremely time intensive. This work proposes training convolutional networks directly predict from signal-intensity curves that are trained using estimates obtained inference. allows fast estimation with similar performance