LAUSR

Recent work on high-performance lower-field MR systems has renewed the interest in assessing relative advantages and disadvantages of magnetic fields <1T. The objective of the present work was to investigate signal-to-noise ratio (SNR) scaling of point-resolved spectroscopy (PRESS) as a function of field strength and to test the feasibility of proton MRS of triglycerides (TG) in human in vivo myocardium at 0.75T relative to 1.5T and 3T. Measurements at 0.75T were obtained by temporarily ramping down a clinical 3T MR scanner. System configurations at 0.75, 1.5 and 3T featured identical hard- and software, except for differences in transmit/receive coil geometries and receive channel count, which were accounted for in SNR comparisons. Proton MRS was performed at 0.75T, 1.5T and 3T in ex vivo tissue and in vivo calf muscle to measure T1 and T2 values as a function of field strength, which in turn served as input to simulations of SNR scaling and field-dependent TG fit errors. Preliminary in vivo spectra of myocardium were acquired at 0.75T, 1.5T and 3T in healthy subjects. Measurements of both ex vivo tissue and in vivo muscle tissue at 0.75T versus 1.5T and 3T confirmed decreasing T1 and increasing T2 * for decreasing field strengths. Using measured T1 , T2 and T2 * as input and using field-dependent echo time and bandwidth scaling, simulated Cramér-Rao lower bounds of TG amplitudes at 0.75T were 2.3 and 4.5 times larger with respect to 1.5T and 3T, respectively. In vivo measurements demonstrate that human proton spectroscopy of triglycerides in cardiac muscle is feasible at 0.75T, supporting the potential practical value of lower-field high-performance MR systems.