Objectives: To investigate optimised isotropic 3D turbo spin echo (TSE) and gradient echo (GRE)-based pulse sequences for visualisation of articular cartilage lesions within the knee joint.
Methods: Optimisation of experimental imaging sequences was completed using healthy volunteers (n=16) with a 3-Tesla (3T) MRI scanner. Imaging of patients with knee cartilage abnormalities (n=57) was then performed. Acquired sequences included 3D proton density-weighted (PDW) TSE (SPACE) with and without fat-suppression (FS), and T2*W GRE (TrueFISP) sequences, with acquisition times of 6:51, 6:32 and 5:35 min, respectively.
Results: One hundred sixty-one confirmed cartilage lesions were detected and categorised (Grade II n=90, Grade III n=71). The highest sensitivity and specificity for detecting cartilage lesions were obtained with TrueFISP with values of 84.7% and 92%, respectively. Cartilage SNR mean for PDW SPACE-FS was the highest at 72.2. TrueFISP attained the highest CNR means for joint fluid/cartilage (101.5) and joint fluid/ligament (156.5), and the lowest CNR for cartilage/meniscus (48.5). Significant differences were identified across the three sequences for all anatomical structures with respect to SNR and CNR findings (p-value <0.05).
Conclusion: Isotropic TrueFISP at 3T, optimised for acquisition time, accurately detects cartilage defects, although it demonstrated the lowest contrast between cartilage and meniscus.
Key points: • Cartilage is better visualised with 3D TrueFISP than 3D SPACE sequences. • 3D TrueFISP is a reliable sequence for detecting low- and high-grade cartilage defects. • 3D TrueFISP at 3T provides excellent contrast between cartilage and joint fluid.
Keywords: Articular; Cartilage; Evidence-based practice; Knee joint; Magnetic resonance imaging.