Isotropic Three-Dimensional T2 Mapping of Knee Cartilage: Development and Validation

Purpose: 1) To implement a higher-resolution isotropic 3D T-2 mapping technique that uses sequential T-2-prepared segmented gradient-recalled echo (Iso3DGRE) images for knee cartilage evaluation, and 2) to validate it both in vitro and in vivo in healthy volunteers and patients with knee osteoarthritis. Materials and Methods: The Iso3DGRE sequence with an isotropic 0.6 mm spatial resolution was developed on a clinical 3T MR scanner. Numerical simulations were performed to optimize the pulse sequence parameters. A phantom study was performed to validate the T-2 estimation accuracy. The repeatability of the sequence was assessed in healthy volunteers (n = 7). T-2 values were compared with those from a clinical standard 2D multislice multiecho (MSME) T-2 mapping sequence in knees of healthy volunteers (n = 13) and in patients with knee osteoarthritis (OA, n = 5). Results: The numerical simulations resulted in 100 excitations per segment and an optimal radiofrequency (RF) excitation angle of 15 degrees. The phantom study demonstrated a good correlation of the technique with the reference standard (slope 0.9 +/- 0.05, intercept 0.2 +/- 1.7 msec, R-2 >= 0.99). Repeated measurements of cartilage T-2 values in healthy volunteers showed a coefficient of variation of 5.6%. Both Iso3DGRE and MSME techniques found significantly higher cartilage T-2 values (P < 0.03) in OA patients. Iso3DGRE precision was equal to that of the MSME T-2 mapping in healthy volunteers, and significantly higher in OA (P = 0.01). Conclusion: This study successfully demonstrated that high-resolution isotropic 3D T-2 mapping for knee cartilage characterization is feasible, accurate, repeatable, and precise. The technique allows for multiplanar reformatting and thus T-2 quantification in any plane of interest.