Jakob Assländer¹, Steffen Glaser², and Jürgen Hennig^1
1Dept. of Radiology - Medical Physics, University Medical Center Freiburg, Freiburg, Germany, ²Dept. of Chemistry, Technische Universität München, Munich, Germany

This work discusses steady state issues in SSFP-based fingerprinting sequences. It is shown that variations of the flip angle destroy the steady state, causing instabilities with respect to intra-voxel dephasing. A pseudo steady state can be achieved by adapting TR and TE to a given flip angle pattern, restoring the typical SSFP behavior. Furthermore, an iterative reconstruction algorithm for fingerprinting data is proposed.

Karsten Sommer¹, Thomas Amthor¹, Peter Koken¹, Mariya Doneva¹, and Peter Börnert^1
1Philips Research Europe, Hamburg, Germany

A key question of magnetic resonance fingerprinting (MRF) is the appropriate choice of sequence parameters to achieve a high sensitivity to the tissue parameters of interest. In this contribution, different candidates for a measure of MRF sequence encoding capability are evaluated. While interpretation of measures that rely on local or global dot products proved difficult, a ‘brute force’ Monte Carlo approach showed good agreement with experimental results. By restricting this Monte Carlo method to small local dictionaries, substantial acceleration could be achieved.

Ouri Cohen^1,2, Mathieu Sarracanie^1,3, Matthew S. Rosen^1,3, and Jerome L. Ackerman^1,2
1Athinoula A. Martinos Center, Charlestown, MA, United States, ²Radiology, Massachusetts General Hospital, Boston, MA, United States, ³Physics, Harvard University, Cambridge, MA, United States

In this work we demonstrate an in vivo human brain application of a previously described schedule optimization method for rapid MR Fingerprinting. The method is validated in a phantom by comparison to a spin-echo sequence. The optimized schedule allowed acquisition of a single slice in 2.4 seconds without the use of any k-space undersampling. 

Jesse Ian Hamilton¹, Anagha Deshmane¹, Mark Griswold^1,2, and Nicole Seiberlich^1,2
1Biomedical Engineering, Case Western Reserve University, Cleveland, OH, United States, ²Radiology, University Hospitals, Cleveland, OH, United States

MR Fingerprinting with Chemical Exchange (MRF-X) is presented for in vivo quantification of relaxation times, volume fraction, and exchange rate for tissues with two compartments. Data are presented in healthy volunteers in both brain and leg skeletal muscle and compared with previously reported measurements.

Mariya Doneva¹, Thomas Amthor¹, Peter Koken¹, Karsten Sommer¹, and Peter Börnert^1
1Philips Research Europe, Hamburg, Germany

In this work, we present a method for reconstruction of undersampled Magnetic Resonance Fingerprinting (MRF) data based on low rank matrix completion, which is performed entirely in k-space and has low computational cost. The method shows significant improvement in the MRF parameter maps accuracy compared to direct matching from undersampled data, potentially enabling more robust highly accelerated MR Fingerprinting.

Guido Buonincontri¹, Rolf Schulte², Mirco Cosottini^3,4, Stephen Sawiak⁵, and Michela Tosetti^4,6
1INFN Pisa, Pisa, Italy, ²GE Global Research, Munich, Germany, ³Department of Radiology, University of Pisa, Pisa, Italy, ⁴IMAGO7 Foundation, Pisa, Italy, ⁵Clinical Neurosciences, University of Cambridge, Cambridge, United Kingdom, ⁶IRCCS Stella Maris, Pisa, Italy

MR fingerprinting (MRF) can be used to rapidly estimate quantitative parameters in MRI. However, the homogeneity of the transmission radiofrequency field (B1+) can introduce errors in the measurements. Here, we modified spiral MRF acquisitions and included the effects of B1+ directly in the reconstruction framework. We could obtain B1-corrected T1 and T2 maps without using an extra scan. These advances are demonstrated in human brain images at 7T.

Christopher C. Cline^1,2, Xiao Chen¹, Boris Mailhe¹, Qiu Wang¹, and Mariappan Nadar^1
1Medical Imaging Technologies, Siemens Healthcare, Princeton, NJ, United States, ²Biomedical Engineering, University of Minnesota, Minneapolis, MN, United States

We propose an accelerated iterative reconstruction for magnetic resonance fingerprinting (AIR-MRF) with comprehensive integration of temporal compression of fingerprints and accelerated dictionary matching with approximate nearest neighbor search. Faster and more accurate MRF reconstruction was achieved, as demonstrated by simulations with a numerical phantom.

Debra F. McGivney¹, Anagha Deshmane², Yun Jiang², Dan Ma¹, and Mark A. Griswold^1
1Radiology, Case Western Reserve University, Cleveland, OH, United States, ²Biomedical Engineering, Case Western Reserve University, Cleveland, OH, United States

Magnetic resonance fingerprinting (MRF) is a technique that allows us to produce quantitative maps of tissue parameters such as T1 and T2 relaxation times, however it is susceptible to artifacts due to the partial volume effect. The aim of this work is to provide a blind solution to the partial volume problem in MRF using the Bayesian statistical framework. A complete description of the algorithm is presented as well as applications to in vivo data.

Xiaodi Zhang^1,2, Rui Li¹, and Xiaoping Hu^2
1Center for Biomedical Imaging Research, Department of Biomedical Engineering, School of Medicine, Tsinghua University, Beijing, China, People's Republic of, ²The Wallace H. Coulter Department of Biomedical Engineering, Georgia Institute of Technology and Emory University, Atlanta, GA, United States

The reconstruction of MR fingerprinting currently relies on matching with a dictionary. In this paper, we describe an alternative method using Kalman filter instead of dictionary in the reconstruction. The method is shown to allow the reconstruction of MR Fingerprinting without the use of dictionary and achieves better results.

Taehwa Hong¹, Min-Oh Kim¹, Dongyeob Han¹, and Dong-Hyun Kim^1
1Electrical and Electronic engineering, Yonsei university, Seoul, Korea, Republic of

MR fingerprinting (MRF) is a rapid method for quantifying multiple tissue properties. However, estimation errors can increase when systematic imperfections including RF and gradient coils exist. In this study, we analyzed estimation errors from non-ideal slice profile and gradient delay by simulation. Our results showed that these systematic imperfections can cause significant errors in parameter estimation.