Song-I Lim^1,2, Mark Stephan Widmaier^1,3, Yun Jiang⁴, and Lijing Xin^1,2
1CIBM Center for Biomedical Imaging, Lausanne, Switzerland, ²Animal Imaging and Technology, EPFL, Lausanne, Switzerland, ³Laboratory for Functional and Metabolic Imaging, EPFL,, Lausanne, Switzerland, ⁴Department of Radiology, University of Michigan, Ann Arbor, MI, United States

This study is to validate 31P MRS fingerprinting scheme at 7T. In this experiment, MRF scheme was validated using Pi phantom and in vivo brain data was acquired.

Ouri Cohen¹, Or Perlman², Christian T Farrar², and Ricardo Otazo^1
1Memorial Sloan Kettering Cancer Center, New York, NY, United States, ²Radiology, Massachusetts General Hospital, Charlestown, MA, United States

Development of a CEST-MR Fingerprinting (CEST-MRF) pulse sequence combined with a physics-based deep learning approach suitable for use on a 3T clinical scanner is described and its utility demonstrated in a healthy human brain. The acquisition is short (less than 2 minutes) and simultaneously yields 6 quantitative tissue parameters that can be used for tissue characterization.

Krishnapriya Venugopal¹, Esther A.H Warnert¹, Daniëlle van Dorth², Marion Smits¹, Juan Antonio Hernandez Tamames¹, Matthias J.P van Osch², and Dirk H.J Poot^1
1Radiology and Nuclear Medicine, Erasmus MC, Rotterdam, Netherlands, ²Radiology, Leiden University Medical Center, Leiden, Netherlands

This study uses a DSC based fingerprinting approach, monitoring the time evolution of a Hybrid (H) EPI sequence (HEPI) that simultaneously acquires GRE and SE. HEPI properties are incorporated from the scanner into a simulation of contrast agent extravasation and MR signal evolution. Signals simulated during bolus passage are used to construct GRE, SE and combined GRE-SE dictionaries in which vessel permeability (k), vessel radius (R), and cerebral blood volume fraction (rCBV) are varied. The dictionary is matched to in-vivo data of a brain tumor patient to retrieve information on the underlying microvasculature.

Gregory J. Wheeler¹ and Audrey P. Fan^1,2
1Biomedical Engineering, University of California Davis, Davis, CA, United States, ²Neurology, University of California Davis, Davis, CA, United States

Magnetic resonance vascular fingerprinting (MRvF) enables the simultaneous measurement of quantitative oxygen saturation, cerebral blood volume, and microvascular radii maps from a single scan. Accelerating image acquisition for MRvF would enable new dynamic investigations into cerebrovascular diseases. Acquisition acceleration will result in tradeoffs between parameter map accuracy, time resolution, and noise. We performed a simulation study in which five signal-to-noise ratios and five echo train lengths were used to generate 25 simulated datasets to assess limits required for accurate matching. Vascular parameter matching accuracy increases with increased SNR and echo train length and requires an SNR above at least 20.

Yilin Liu¹, Yong Chen², and Pew-thian Yap^1
1University of North Carolina at Chapel Hill, Chapel Hill, NC, United States, ²Case Western Reserve University, Cleveland, OH, United States

Existing techniques for 3D MR fingerprinting focus on accelerating either acquisition or template matching, with typically limited spatial resolution. We develop a 3D MRF sequence coupled with a novel end-to-end deep learning image reconstruction framework for rapid and simultaneous whole-brain quantification of T1 and T2 relaxation times with isotropic submillimeter spatial resolution. We demonstrate feasibility with both retrospectively and prospectively accelerated 3D MRF data.  

Charit Tippareddy¹, Walter Zhao², Andrew Sloan^3,4, Jeffrey Sunshine⁵, Jill Barnholtz-Sloan⁶, Mark Griswold^2,5, Dan Ma^2,5, and Chaitra Badve^5
1Case Western Reserve University School of Medicine, Cleveland, OH, United States, ²Department of Biomedical Engineering, Case Western Reserve University, Cleveland, OH, United States, ³Departments of Neurosurgery and Pathology, University Hospitals Cleveland Medical Center, Cleveland, OH, United States, ⁴Seidman Cancer Center and Case Comprehensive Cancer Center, Cleveland, OH, United States, ⁵Department of Radiology, University Hospitals Cleveland Medical Center, Cleveland, OH, United States, ⁶Department of Population and Quantitative Health Sciences, University Hospitals Cleveland Medical Center, Cleveland, OH, United States

The utility of MR fingerprinting (MRF) in characterization of non-enhancing tumor (NET) region in brain tumors has not been demonstrated. Quantitative characterization of NET (aka peritumoral white matter) in glioblastomas (GBM) is essential to identify imaging surrogates for tumor infiltration and predict future recurrence. Here we demonstrate the utility of pre and post contrast MRF to characterize and compare the NET region surrounding GBMs and metastases (METS). We identify NET radiomic features that are unique to each tumor type as well as features that can differentiate near (within 1 cm) versus far (beyond 1 cm) NET regions within each group.

Joon Yul Yul Choi¹, Siyuan Hu², Ting-yu Su^1,2, Yingying Tang¹, Ken Sakaie³, Ingmar Blümcke^1,4, Imad Najm¹, Stephen Jones³, Mark Griswold⁵, Dan Ma², and Zhong Irene Wang^1
1Epilepsy Center, Neurological Institue, Cleveland Clinic, Cleveland, OH, United States, ²Biomedical Engineering, Case Western Reserve University, Cleveland, OH, United States, ³Imaging Institute, Cleveland Clinic, Cleveland, OH, United States, ⁴Neuropathology, University of Erlangen, Erlangen, Germany, ⁵Radiology, Case Western Reserve University, Cleveland, OH, United States

We demonstrate in this study the sensitivity of multi-parametric magnetic resonance fingerprinting (MRF) results at 3T to differentiate cortical regions with different cyto- or myelo-architecture. The study investigated the quantitative T1 and T2 values in various Brodmann areas to verify the sensitivity of MRF in probing tissue properties of the human cortex. Additionally, the study explores the relationship between quantitative T1 and T2 values of gray and white matter in Brodmann areas.

Shahrzad Moinian^1,2, David Reutens^1,2, and Viktor Vegh^1,2
1Centre for Advanced Imaging, The University of Queensland, Brisbane, Australia, ²ARC Training Centre for Innovation in Biomedical Imaging Technology, The University of Queensland, Brisbane, Australia

We have previously shown the potential efficacy of a magnetic resonance fingerprinting (MRF) residual approach for human cerebral cortex parcellation. However, the classical Bloch equations, commonly used for MR signal simulation in the MRF dictionary, may not accurately describe the complex effect of the ensemble of microarchitectonic components of the gray matter tissue on MRF signal. This work benefitted from the more flexibility of the extended time-fractional order Bloch equations to improve MRF dictionary fitting accuracy. We demonstrated that the time-fractional order parameter α potentially associates with the effect of interareal architectonic variability, hypothetically leading to more accurate cortical parcellation.

Simran Kukran^1,2, Joely Smith^3,4, Luke Dixon^1,3, Ben Statton⁵, Sarah Cardona³, Lillie Pakzad-Shahabi^6,7, Matthew Williams^6,8, Dow-Mu Koh^2,9, Rebecca Quest^3,4, Matthew Orton², and Matthew Grech-Sollars^1,3
1Department of Surgery and Cancer, Imperial College London, London, United Kingdom, ²Department of Radiotherapy and Imaging, Institute of Cancer Research, London, United Kingdom, ³Department of Imaging, Imperial College Healthcare NHS Trust, London, United Kingdom, ⁴Department of Bioengineering, Imperial College London, London, United Kingdom, ⁵Medical Research Council, London Institute of Medical Sciences, Imperial College London, London, United Kingdom, ⁶Computational Oncology group, Institute for Global Health Innovation, Imperial College London, London, United Kingdom, ⁷John Fulcher Neuro-oncology Laboratory, Department of Brain Sciences, Imperial College London, London, United Kingdom, ⁸Radiotherapy Department, Charing Cross Hospital, London, United Kingdom, ⁹Department of Radiology, Royal Marsden Hospital, London, United Kingdom

MR Fingerprinting (MRF) was found to give highly repeatable T1 and T2 measurements in glioma and normal appearing contralateral brain tissue. Validity was investigated via comparison to standard mapping techniques: variable flip angle for T1 and multi-echo spin echo for T2. Biases were found between MRF and standard relaxometry methods, as in previous healthy volunteer studies. Statistically significant strong and moderate correlations were found between the MRF and standard mapping methods for T1 and T2 respectively, indicating MRF is comparably sensitive to changes in T1 and T2 as established mapping techniques in both cancerous and normal appearing contralateral brain tissue. 

pavan Poojar^1,2, Enlin Qian¹, and Sairam Geethanath ^1,2
1Columbia Magnetic Resonance Research Center, Columbia University, New York, NY, United States, ²Dayananda Sagar College of Engineering, Bangalore, India

MR fingerprinting(MRF) has an advantage over quantitative MRI as it allows simultaneous acquisition of multi-parametric maps but does not generate multi-contrast images required for routine clinical studies. Tailored MRF(TMRF) allows simultaneous acquisition of two quantitative maps and six qualitative contrasts. A TMRF repeatability study was conducted for four days on one in vivo healthy human brain. Signal-to-noise ratio (SNR) and mean intensity values for grey matter(GM) and white matter(WM) were computed. Standard deviation of SNR for WM and GM were in the range of 0.1 to 0.75 and 0.2 to 1.1 respectively. This narrow range shows the repeatability of TMRF

Krishna Pandu Wicaksono¹, Yasutaka Fushimi¹, Satoshi Nakajima¹, Akihiko Sakata¹, Takuya Hinoda¹, Sonoko Oshima¹, Sayo Otani¹, Hiroshi Tagawa¹, Yang Wang¹, Tomohisa Okada¹, and Yuji Nakamoto^1
1Kyoto University, Graduate School of Medicine, Kyoto, Japan

3D MR Fingerprinting was introduced as a rapid quantitative MRI technique, enabling higher SNR efficiency and spatial resolution than the 2D counterpart. As a crucial element of MRF reconstruction, the impact of dictionary resolution on MRF performance is essential to be investigated. Following our phantom study, which determined equivalent accuracy and repeatability of 3D MRF using two different dictionary resolutions, a reproducibility study in healthy volunteers was performed. This study demonstrated a comparable 3D MRF reproducibility from two different dictionary resolutions in most brain parenchyma. Yet, lower reproducibility was evident in CSF measurement, more obviously in a higher resolution dictionary.

Mohammad Golbabaee¹ and Clarice Poon^1
1University of Bath, Bath, United Kingdom

We introduce a novel off-the-grid sparse approximation algorithm to separate multiple tissue compartments in image voxels and to estimate quantitatively their NMR properties and mixture fractions, given the MR fingerprinting (MRF) measurements. The proposed algorithm is an accurate and importantly a scalable alternative to the multicompartment MRF baselines because it does not rely on fine-gridded multiparametric MRF dictionaries. The method is theoretically described, and its feasibility is demonstrated and compared to other baselines on in-vivo healthy brain measurements.

Martijn A. Nagtegaal¹, Laura Nunez Gonzalez², Dirk H.J. Poot², Matthias J.P. van Osch³, Jeroen H.J.M. de Bresser⁴, Juan A. Hernandez Tamames^1,2, and Frans M. Vos^1,2
1Department of Imaging Physics, Delft University of Technology, Delft, Netherlands, ²Department of Radiology and Nuclear Medicine, Erasmus MC, Rotterdam, Netherlands, ³C.J. Gorter Center for high field MRI, Department of Radiology, Leiden University Medical Center, Leiden, Netherlands, ⁴Department of Radiology, Leiden University Medical Center, Leiden, Netherlands

Accurate proton density estimations are required to obtain tissue volume fractions from multi-component MR Fingerprinting data. We propose a method for estimating relative proton densities per tissue while taking the receiver sensitivity profile into account. In 20 different numerical brain phantoms this shows to improve tissue segmentations compared to conventional methods that use $$$T_1$$$ weighted images. Estimated proton density values for single slice in vivo data (7 scans for 4 subjects) were in range with literature values in particular for white and gray matter.

Zihan Zhou¹, Qing Li^1,2, Congyu Liao³, Xiaozhi Cao³, Ting Gong¹, Qiuping Ding¹, Hongjian He¹, and Jianhui Zhong^1,4
1Center for Brain Imaging Science and Technology, College of Biomedical Engineering and Instrumental Science, Zhejiang University, Hangzhou, China, ²MR Collaboration, Siemens Healthcare Ltd, Shanghai, China, ³Radiological Sciences Laboratory, Stanford University, Stanford, CA, United States, ⁴Department of Imaging Sciences, University of Rochester, Rochester, NY, United States

We proposed a 3D ultrashort echo time MR fingerprinting (3D UTE-MRF) method for whole brain myelin imaging in vivo and ex vivo. A series of dual-echo time images were acquired, and images optimized for long  T2 tissue suppression were found in the second echo image series, which were used to extract myelin signals. Non-negative least-square was used to map the ultrashort T2 myelin tissue fraction. 3D UTE-MRF sequence can achieve whole brain myelin mapping in 15 min with 0.87 mm isotropic resolution.

Congyu Liao¹, Xiaozhi Cao¹, Ting Gong², Zhe Wu³, Zihan Zhou², Hongjian He², Jianhui Zhong^2,4, and Kawin Setsompop^1
1Radiological Sciences Laboratory, Stanford University, Stanford, CA, United States, ²Center for Brain Imaging Science and Technology, College of Biomedical Engineering & Instrument Science, Zhejiang University, Hangzhou, China, ³Techna Institute, University Health Network, Toronto, ON, Canada, ⁴Department of Imaging Sciences, University of Rochester, Rochester, NY, United States

In this work, we developed ViSTa-MRF, which combined Visualization of Short Transverse relaxation time component (ViSTa) technique with MR Fingerprinting (MRF), to achieve whole-brain myelin-water fraction (MWF) and T₁/T₂/PD mapping at 1mm isotropic resolution in 10 minutes on a clinical 3T scanner. To achieve this fast acquisition, the ViSTa-MRF sequence also leverages an efficient 3D-spiral-projection acquisition along with spatiotemporal subspace reconstruction. With the proposed ViSTa-MRF method, direct MWF mapping was achieved without a need for multicompartment fitting. In comparison to conventional myelin-water imaging, the ViSTa-MRF method can provide improved-SNR and faster acquisition with high image-quality.

Cem Gultekin¹, Jakob Assländer², and Carlos Fernandez-Granda^3
1Mathematics, Courant Institute of Mathematical Science, New York, NY, United States, ²Radiology, New York University Grossman School of Medicine, New York, NY, United States, ³Mathematics and Data Science, Courant Institute of Mathematical Science and Center for Data Science New York University, New York, NY, United States

This work presents a new robust black-box numerical solver that can reliably solve many MRI typical ordinary differential equations. Our adaptive Petrov-Galerkin (PG) method can solve challenging MRI problems with additional complexities such as B₀- and B₁- inhomogeneities, RF pulses, chemical exchange, and magnetization transfer (MT). We apply the proposed technique to solve an ODE-constrained optimization problem for pulse design via gradient descent. Our method reduces the time required to compute the gradients by three orders of magnitude.

Andrew Mao^1,2,3, Sebastian Flassbeck^1,2, Cem Gultekin⁴, Xiaoxia Zhang^1,2, and Jakob Asslaender^1,2
1Center for Biomedical Imaging, Department of Radiology, New York University School of Medicine, New York, NY, United States, ²Center for Advanced Imaging Innovation and Research, New York University School of Medicine, New York, NY, United States, ³Vilcek Institute of Graduate Biomedical Sciences, New York University School of Medicine, New York, NY, United States, ⁴Courant Institute of Mathematical Sciences, New York University, New York, NY, United States

This work extends the recently introduced hybrid-state model to myelin water imaging. We numerically optimize a hybrid-state MR fingerprinting sequence for SNR efficiency and demonstrate in vivo the ability to quantify the myelin water fraction and T2 of axonal/extra-axonal water. Our preliminary results show myelin water fraction maps in agreement with literature and demonstrate the feasibility of simultaneous myelin water fraction, T1 and T2 quantification with high spatial resolution (1.2mm isotropic), full brain coverage with a 14 minute scan.

Pavan Poojar^1,2, Enlin Qian¹, and Sairam Geethanath ^1,2
1Columbia Magnetic Resonance Research Center, Columbia University, New York, NY, United States, ²Dayananda Sagar College of Engineering, Bangalore, India

MR Fingerprinting (MRF) allows simultaneous acquisition of multi-parametric maps but the synthetic contrast images suffer from artifacts due to incomplete simulations. This work provides rapid (~4min), natural contrast (non-synthetic), quantitative (T₁ and T₂ maps), and qualitative images (T₁-weighted, T₁-FLAIR, T₂-weighted, STIR, water,fat) simultaneously. Tailored MRF (TMRF) was demonstrated on four volunteers on 3T GE 750w scanner. It was compared with gold standard (GS) and MRF by computing SNR and mean intensity values of white matter (WM) and grey matter (GM) contrast. The SNR of GS>TMRF>MRF and the contrast for TMRF was greater than MRF and GS.

Di Cui¹, Edward S. Hui², and Peng Cao^1
1Diagnostic Radiology, The University of Hong Kong, Hong Kong, China, ²Rehabilitation Science, The Hong Kong Polytechnic University, Hong Kong, China

A k-space compression strategy is proposed in a 3D alternating direction method of multipliers (ADMM) framework in this study, with data and image series compressed, and intermediate computation simplified.

Peng Li¹ and Yue Hu^1
1Harbin Institute of Technology, Harbin, China

Due to the capability of fast multi-parametric quantitative imaging, magnetic resonance fingerprinting has become a promising quantitative magnetic resonance imaging (QMRI) approach. However, the highly undersampled and noise-contaminated k-space data will cause critical spatial artifacts, which subsequently lead to inaccurate estimation of the quantitative parameters. In this paper, we introduce a novel framework based on structured low-rank approximation and subspace modeling to recover temporal MRF data from its highly undersampled and noisy Fourier coefficients.