Differentiation between grade 3 and grade 4 articular cartilage defects of the knee: Fat-suppressed proton density-weighted versus fat-suppressed three-dimensional gradient-echo MRI

Purpose: To retrospectively compare the diagnostic efficacy between FS PD-weighted and FS 3D gradient-echo MRI for differentiating between grade 3 and grade 4 cartilage defects of the knee with arthroscopy as the standard of reference. Material and Methods: Twenty-one patients who had grade 3 or 4 cartilage defects in medial femoral condyle at arthroscopy and knee MRI were included in this study: grade 3, > 50% cartilage defects; grade 4, full thickness cartilage defects exposed to the bone. Sagittal FS PD-weighted MR images and FS 3D gradient-echo images with 1.5 T MR images were independently graded for the cartilage abnormalities of medial femoral condyle by two musculoskeletal radiologists. Statistical analysis was performed by Fisher's exact test. Inter-observer agreement in grading of cartilage was assessed using kappa coefficients. Results: Arthroscopy revealed grade 3 defects in 17 patients and grade 4 defects in 4 patients in medial femoral condyles. For FS 3D gradient-echo images grade 3 defects were graded as grade 3 (n=15) and grade 4 (n=2), and all grade 4 defects (n=4) were correctly graded. However, for FS PD-weighted MR images all grade 3 defects were misinterpreted as grade 1 (n=1) and grade 4 (n=16), whereas all grade 4 defects (n=4) were correctly graded. FS 3D gradient-echo MRI could differentiate grade 3 from grade 4 defects (P=0.003), whereas FS PD-weighted imaging could not (P=1.0). Inter-observer agreement was substantial (kappa=0.70) for grading of cartilage using FS PD-weighted imaging, whereas it was moderate (kappa=0.46) using FS 3D gradient-echo imaging. Conclusion: FS 3D gradient-echo MRI is more helpful for differentiating between grade 3 and grade 4 cartilage defects than is FS PD-weighted imaging.