%0 Generic
%A Hermann, Ingo
%C Heidelberg
%D 2021
%F heidok:29978
%R 10.11588/heidok.00029978
%T Development of novel methods in quantitative magnetic resonance imaging
%U https://archiv.ub.uni-heidelberg.de/volltextserver/29978/
%X Quantitative magnetic resonance imaging (MRI) is a non-invasive and versatile tool for the assessment of anatomical structures. In recent years, MRI has evolved rapidly and is of high clinical interest because of its potential to distinguish diseased from healthy tissue. A variety of methods have been proposed for quantitative cardiac MRI, but insufficient precision and practicality limit its clinical use. One objective of this work was to analyze the effects of blood flow in relation to T1 relaxation times of blood for conventional inversion recovery (IR) and saturation recovery (SR) methods. Simulations, phantom, and in vivo  experiments were performed to validate the effects of flow. The in-flow of non-prepared spins resulted in decreased T1 times, and thus SR methods were found to be more resistant to flow effects. Based on this, a sequence was developed for simultaneous quantification of T1, T2, and T2*. Phantom measurements were performed with high accuracy in agreement with simulations and good visual image quality was observed in the myocardium compared to reference methods and in patients. In the second part of the work, a novel renal magnetic resonance fingerprinting (MRF) approach was developed for the simultaneous quantification of T1 and T2* within four slices. Simulations showed good agreement with phantom measurements and a convergence of the reconstructed relaxation times. In vivo measurements benefited from a 10-fold speedup compared to conventional methods and good reproducibility for repeated measurements. Additionally, this technique has been used in brain scans at two centers to study white matter lesions in patients with multiple sclerosis. Complex and computationally costly data processing was replaced by a neural network combining noise reduction, T1 and T2* reconstruction, distortion correction, and white matter, gray matter and lesion segmentation. Robust and accurate parameter maps provide reconstructions with a 100-fold speed up, and therefore ideal for clinical applications.