Simultaneous fat-free isotropic 3D anatomical imaging and T2 mapping of knee cartilage with lipid-insensitive binomial off-resonant RF excitation (LIBRE) pulses.

Abstract

Improved knee cartilage morphological delineation and T2 mapping precision necessitates isotropic 3D high-resolution and efficient fat suppression. To develop and assess an isotropic 3D lipid-insensitive T2 mapping technique of the knee for improved cartilage delineation and precise measurement of T2 relaxation times. Prospective. Phantoms (n = 6) used in this study were designed to mimic the T1 and T2 relaxation times of cartilage and fat. The study cohort comprised healthy volunteers (n = 7) for morphometry and T2 relaxation time measurements. A high-resolution isotropic 3D T2 mapping technique that uses sequential T2 -prepared segmented gradient-recalled echo (Iso3DGRE) images and lipid-insensitive binomial off-resonant radiofrequency (RF) excitation (LIBRE) at 3T. Numerical simulations and phantom experiments were performed to optimize the LIBRE pulse. Phantom studies were carried out to test the accuracy of the technique against reference standard spin-echo (SE) T2 mapping. Subsequently, T2 maps with and without LIBRE pulses were acquired in knees of healthy volunteers and the T2 relaxation time values in different cartilage compartments were compared. A two-tailed paired Student's t-test was used to compare the average T2 values and the relative standard deviations (inverse measurement of the precision) obtained with and without LIBRE pulses. A LIBRE pulse of 1 msec suppressed fat with an RF excitation frequency offset of 1560 Hz and optimal RF excitation angle of 35°. These results were corroborated by phantom and knee experiments. Robust and homogeneous fat suppression was obtained (a fat signal-to-noise ratio (SNR) decrease of 86.4 ± 2.4%). In phantoms, T2 values were found in good agreement when comparing LIBRE-Iso3DGRE with SE (slope 0.93 ± 0.04, intercept 0.11 ± 1.6 msec, R2 >0.99). In vivo, LIBRE excitation resulted in more precise T2 estimation (23.7 ± 7.4%) than normal excitation (30.5 ± 9.9%, P < 0.0001). Homogeneous LIBRE fat signal suppression was achieved with a total RF pulse duration of 1 msec, allowing for the removal of chemical shift artifacts and resulting in improved cartilage delineation and precise T2 values. 2 Technical Efficacy: Stage 1 J. Magn. Reson. Imaging 2019;49:1275-1284.