Clinical Usefulness of Adding 3D Cartilage Imaging Sequences to a Routine Knee MR Protocol

Abstract

The objective of our study was to determine whether 3D sequences can improve the diagnostic performance of a routine MR protocol for detecting cartilage lesions within the knee joint at 3 T. An iterative decomposition of water and fat with echo asymmetry and least-squares estimation (IDEAL) spoiled gradient-recalled echo (SPGR) sequence (n = 75 patients) or an IDEAL gradient-recalled acquisition in a steady state (GRASS) sequence (n = 75 patients) was added to routine 3-T knee MR protocol to examine 150 patients who subsequently underwent arthroscopic knee surgery. Each articular surface of the knee joint was graded at arthroscopy. All MR examinations were independently reviewed twice by two musculoskeletal radiologists. During the first review, the routine MR protocol was used alone to grade each articular surface of the knee joint; during the second review, the routine MR protocol was used with IDEAL-SPGR or IDEAL-GRASS. Using arthroscopy as the reference standard, the sensitivity and specificity for detecting cartilage lesions and the proportion of correctly graded cartilage lesions were determined for the routine MR protocol alone and for the routine MR protocol with IDEAL-SPGR or IDEAL-GRASS. The McNemar test was used to compare sensitivity, specificity, and accuracy values and proportions of correctly graded cartilage lesions. There was a statistically significant improvement in the specificity (p < 0.05) but not the sensitivity (p = 0.08-0.32) for detecting cartilage lesions when using the routine MR protocol with IDEAL-SPGR or IDEAL-GRASS. There was also a statistically significant (p < 0.05) improvement in the proportion of correctly graded cartilage lesions when using the routine MR protocol with IDEAL-SPGR or with IDEAL-GRASS. Adding 3D sequences to a routine MR protocol improves the diagnostic performance for detecting cartilage lesions within the knee joint at 3 T.