High resolution in-vivo measurement of sodium T1 of human knee cartilage

Introduction: Sodium MRI is a promising diagnostic technique for assessing cartilage health in-vivo [1]. The measurement of sodium T1 for tissues in the knee is important for the development of simulations modelling the behaviour of sodium and the creation of optimized pulse sequences. T1 differences can be used to suppress fluid signal (akin to FLAIR for H) as has been shown for the brain [2] and knee [3,4]. Previous work has reported measurements of T1 for bovine [3,5,6] and ex-vivo human cartilage at 7T [4]. However, there are no human in-vivo measurements. Isolating T1 in-vivo for fluid sacs and thin cartilage requires good resolution inversionrecovery images that can be difficult to obtain due to rapid T2 decay. The purpose of this abstract is to measure sodium T1 in-vivo for cartilage, blood, and synovial fluid. We used the T1 values to obtain a high-resolution fluid-suppressed image of human knee cartilage. Methods: In five healthy subjects, inversion-recovery (rectangular π pulse, TI = 3,7,15,25,40,70 ms) sodium imaging was performed on a 4.7T MRI system, with a custom sodium birdcage coil. Twisted projection imaging (p=0.17, projections=3000, readout/TR/TE=12.95/(150+TI)/0.261ms, Nav = 1, res~1.5 mm iso) was used to obtain an image with a 120 mm field of view. Total imaging time for all acquisitions was 42 minutes. The acquired signal was regridded into an 80x80x80 matrix. The images were coregistered and the image data was fit to M=P1(1+P2exp(-TI/T1)) to create a T1 map. In the fit, M is the signal magnitude attained with inversion time TI; P1 is the initial magnitude of the signal; P2 accounts for incomplete inversion from relaxation due to the short T2. The patellar cartilage, femoral/tibial cartilage, popliteal and synovial fluid ROIs were identified on the T1 map. The timing for minimal fluid signal was identified from the inversion recovery curves, and used to obtain a fluid-suppressed sodium image (p=0.17, projections = 3000, readout/TR/TI/TE=12.95/100/24/0.261ms, Nav = 3, total time = 18.9 min, res~ 1.5 mm iso). Results/Discussion: Fits to the inversion curves (Figure 1) were excellent for all subjects and tissues (r > 0.9999) and yielded individual T1 maps (Figure 2). The T1’s of the patellar and femoral/tibial cartilage (Table 1) were similar, ~22 ms, whereas the T1’s for both the popliteal and synovial fluid were longer at ~47 ms. Twisted-projection imaging for the readout enabled the acquisition of the range of inversion times within one scanning session. Our sodium T1 values are a bit smaller than those of ex-vivo human cartilage (26±6 ms) measured at 7T [4]. The T1 value for both the fluids was slightly less than the T1 of saline (53 ms) at 4.7T [2]. This may be due to a difference in composition between saline and synovial fluid or possibly partial volume effects given the small tissue structures in the knee. The fluid suppressed sodium image, with a nominal resolution of 1.5 mm isotropic, demonstrated sharp delineation of the patellar cartilage as well as the thin femoral/tibial cartilage (Figure 3). When compared to previous fluid suppression techniques for sodium which produced images with a resolution of 3.6 mm isotropic in 17 min at 7T [4], twisted projection yielded superior images. This may assist in accurate quantification of cartilage sodium in-vivo and permit the non-invasive evaluation of cartilage degeneration.