Objective: The purpose of our study was to compare the accuracy of 3D fat-suppressed isotropic turbo spin-echo (TSE) sequences using sampling perfection with application-optimized contrasts using different flip angle evolution (SPACE) with 2D conventional MRI at 3 T in determining meniscal tear types.
Materials and methods: Sixty-nine patients with arthroscopically confirmed meniscal tears underwent MRI with 2D sequences and 3D TSE SPACE. Images were retrospectively analyzed by two reviewers and correlated with arthroscopic findings. Meniscal tears were classified into one of eight types: horizontal, longitudinal, radial, root, flap, oblique, complex, and bucket-handle. For every type of tear, sensitivity, specificity, and interobserver agreement were calculated.
Results: Mean sensitivities and specificities on 3D TSE SPACE versus 2D sequences were as follows: for radial tears, 77% and 93% versus 68% and 90%; for flap tears, 73% and 96% versus 54% and 85%; for horizontal tears, 85% and 91% versus 78% versus 87%; for longitudinal tears, 50% and 97% versus 72% and 96%; and for root tears, 88% and 99% versus 81% and 99%. The specificity for flap tears was statistically higher on 3D TSE SPACE than 2D sequences. Interobserver agreements were higher on 3D TSE SPACE than 2D sequences for radial, flap, and longitudinal tears.
Conclusion: In categorizing meniscal tears, 3D TSE SPACE has higher specificity for flap tears than 2D conventional sequences. However, there is no significant difference between 2D conventional sequences and 3D TSE SPACE except for flap tears.