3.0-T Evaluation of Knee Cartilage by Using Three-Dimensional IDEAL GRASS Imaging: Comparison with Fast Spin-Echo Imaging

Abstract

To compare the diagnostic performance of iterative decomposition of water and fat with echo asymmetry and least-squares estimation (IDEAL) gradient-recalled acquisition in steady-state (GRASS) imaging with a routine magnetic resonance (MR) imaging protocol for evaluating knee cartilage at 3.0 T in patients by using arthroscopy as the reference standard. This prospective Health Insurance Portability and Accountability Act-compliant study was performed with a waiver of informed consent from the institutional review board. IDEAL GRASS was added to routine 3.0-T knee MR protocol performed in 95 symptomatic patients (48 males, mean age, 34.5 years; 47 females, mean age, 35.5 years) who underwent subsequent arthroscopic surgery. Radiologists used the routine MR protocol during the first review and IDEAL GRASS during the second to grade each articular surface and to determine the presence of meniscal tears. By using arthroscopy as the reference standard, the sensitivity, specificity, and accuracy of both imaging methods for detecting cartilage lesions and meniscal tears were determined. By using the z test to compare parameters between methods, the respective sensitivity, specificity, and accuracy for detecting all 192 cartilage lesions were 68.5%, 92.6%, and 84.5% for IDEAL GRASS and 66.1%, 92.9%, and 83.9% for the routine MR protocol. There was no significant difference (P = .34-.83) in parameters between methods for detecting cartilage lesions. The respective parameters for detecting 50 medial meniscal tears were 85.0%, 91.1%, and 87.9% for IDEAL GRASS and 94.0%, 90.0%, and 92.1% for the routine MR protocol. The parameters for detecting 31 lateral meniscal tears were 58.0%, 90.6%, and 80.0% for IDEAL GRASS and 80.1%, 91.4%, and 87.9% for the routine MR protocol. The routine MR protocol had a significantly higher sensitivity for detecting medial meniscal tears (P = .04) and lateral meniscal tears (P = .01) and significantly higher accuracy for detecting lateral meniscal tears (P = .03) than IDEAL GRASS. IDEAL GRASS has similar diagnostic performance as routine MR protocol for evaluating the articular cartilage of the knee in clinical patients at 3.0 T but has significantly lower sensitivity and accuracy for detecting meniscal tears.