Purpose: The goal of the study is to develop 31P spectroscopic MR fingerprinting (MRF) at 7T to measure T1 and T2 relaxation times simultaneously and to compare time efficiency and test-retest reproducibility of MRF with conventional inversion recovery and multi-TE methods. Methods: A 31P MRF scheme was designed based on a balanced steady-state free precession type sequence. Dictionary was generated using the Bloch equations. B0 map was acquired experimentally and incorporated into the dictionary. Simulations were performed to evaluate estimation bias. 7 phantoms with different T1 and T2 relaxation times were prepared for MRF validation. 7 volunteers were scanned twice using both MRF and the conventional methods to evaluate the reproducibility. Results: In phantom measurements, T1 and T2 values between MRF and conventional methods demonstrated a good agreement with Pearson's correlation coefficients of 0.99 and 0.97, respectively. In in vivo experiments, estimated T1 by MRF were in good agreement with those measured by the inversion recovery and in literatures. On the other hand, estimated T2 values by MRF were shorter than those measured by the multi-TE method. 31P MRF method can reduce the acquisition time by 15 min providing less than 10% of mean CV for T1 estimations and less than 20% of mean CV for T2 estimations of metabolites. Conclusion: Our results shows the feasibility of simultaneous T1 and T2 measurements of 31P metabolites in human brain using the MRF technique at 7T. High reproducibility can be achieved especially for T1 measurement with 40% time reduction over conventional methods.

Phase-cycled spectroscopic imaging (PCSI) method was implemented and validated for 31P PCSI imaging at 7T. The PCSI method uses a balanced steady-state free precession sequence with an ultra-low flip angle (<1°) to achieve sharp passband with 2.52-ms of TR, which enable to accelerate the acquisition. With prior knowledge of 31P spectra, it is feasible to acquire major 31P peaks by changing the frequency offset and non-uniform phase sweeping instead of acquiring full spectra uniformly. To investigate feasibility of the method, a multi-compartment KH2PO4 phantom with in vivo equivalent concentrations was prepared and 31P PCSI was compared with conventional FID-CSI.

A new and efficient magnetisation transfer 31P magnetic resonance fingerprinting (MT-31P-MRF) approach is introduced to measure the creatine kinase metabolic rate kCK between phosphocreatine (PCr) and adenosine triphosphate (ATP) in human brain. The MRF framework is extended to overcome challenges in conventional 31P measurement methods in the human brain in vivo by reducing specific absorption rate deposition, incorporating system imperfection into a model, and testing a range of values for each parameter to arrive at their estimation. To compute complex high-dimensional datasets, we propose a nested iteration interpolation method (NIIM). As the number of parameters to estimate increases, the size of the dictionary grows exponentially. NIIM can reduce the computational load by breaking dictionary matching into subsolutions of linear computational order. The MT-31P-MRF combined with the NIIM provides T1PCr , T 1ATP , and kCK estimates in good agreement with those obtained by the exchange kinetics by band inversion transfer (EBIT) method and literature values. Furthermore, the test-retest reproducibility results showed that the MT-31P-MRF achieves less than 10% of the coefficient of variation for T1PCr , T 1ATP , and kCK measurements in 4 min 15 s, which is more than four times faster than the EBIT method. We conclude that MT-31P-MRF in combination with the NIIM is a fast, accurate and reproducible approach for in vivo kCK assays in the human brain, which enables the potential for investigating energy metabolism in a clinical setting.

A new and efficient magnetisation transfer 31P magnetic resonance fingerprinting (MT-31P-MRF) approach is introduced to measure the creatine kinase metabolic rate kCK between phosphocreatine (PCr) and adenosine triphosphate (ATP) in human brain. The MRF framework is extended to overcome challenges in conventional 31P measurement methods in the human brain in vivo by reducing specific absorption rate deposition, incorporating system imperfection into a model, and testing a range of values for each parameter to arrive at their estimation. To compute complex high-dimensional datasets, we propose a nested iteration interpolation method (NIIM). As the number of parameters to estimate increases, the size of the dictionary grows exponentially. NIIM can reduce the computational load by breaking dictionary matching into subsolutions of linear computational order. The MT-31P-MRF combined with the NIIM provides T1PCr , T 1ATP , and kCK estimates in good agreement with those obtained by the exchange kinetics by band inversion transfer (EBIT) method and literature values. Furthermore, the test-retest reproducibility results showed that the MT-31P-MRF achieves less than 10% of the coefficient of variation for T1PCr , T 1ATP , and kCK measurements in 4 min 15 s, which is more than four times faster than the EBIT method. We conclude that MT-31P-MRF in combination with the NIIM is a fast, accurate and reproducible approach for in vivo kCK assays in the human brain, which enables the potential for investigating energy metabolism in a clinical setting.

A new and efficient magnetisation transfer 31P magnetic resonance fingerprinting (MT‐31P‐MRF) approach is introduced to measure the creatine kinase metabolic rate between phosphocreatine (PCr) and adenosine triphosphate (ATP) in human brain. The MRF framework is extended to overcome challenges in conventional 31P measurement methods in the human brain, enabling reduced acquisition time and specific absorption rate (SAR). To address the challenge of creating and matching large multiparametric dictionaries in an MRF scheme, a nested iteration interpolation method (NIIM) is introduced. As the number of parameters to estimate increases, the size of the dictionary grows exponentially. NIIM can reduce the computational load by breaking dictionary matching into subsolutions of linear computational order. MT‐31P‐MRF combined with NIIM provides , and estimates in good agreement with those obtained by the exchange kinetics by band inversion transfer (EBIT) method and literature values. Furthermore, the test–retest reproducibility results showed that MT‐31P‐MRF achieves a similar or better coefficient of variation (<12%) for and measurements in 4 min 15 s, than EBIT with 17 min 4 s scan time, enabling a fourfold reduction in scan time. We conclude that MT‐31P‐MRF in combination with NIIM is a fast, accurate, and reproducible approach for in vivo assays in the human brain, which enables the potential to investigate energy metabolism in a clinical setting.