Sun Jung Rhee MD, Presenter: Nothing to Disclose

Jang Gyu Cha MD, Abstract Co-Author: Nothing to Disclose

Hyun-Joo Kim MD, Abstract Co-Author: Nothing to Disclose

Jae Ho Yoo, Abstract Co-Author: Nothing to Disclose

Jai Soung Park, Abstract Co-Author: Nothing to Disclose

Hae-Kyung Lee MD, Abstract Co-Author: Nothing to Disclose

MoonJung Hwang, Abstract Co-Author: Research support, General Electric Company

To compare iterative decomposition of water and fat with echo asymmetry and least-squares estimation(IDEAL) with conventional fat saturation in detecting artificial cartilage lesions in terms of sensitivity, specificity and accuracy of lesion depiction, signal-to-noise ratio (SNR), and contrast-to-noise (CNR) and determine optimal sequence.

46 artificial cartilage lesions were created in 11 porcine knees according to a modified Noyes classification. All imaging were performed on a 1.5T and 3.0T with sagittal imaging and were divided into two groups according to type of fat saturation method. The conventional group was composed of T2-fast spin echo(FSE), proton density(PD)-FSE, 3D spoiled gradient echo(SPGR) with conventional fat-saturation. The IDEAL group was corresponding sequences using IDEAL algorithm. Images were assessed by two musculoskeletal radiologists. Diagnostic performance of each sequence for detecting cartilage lesions was determined by using macroscopic findings as a reference standard. The ROC compare test and chi-square test were used to compare diagnostic performance between two groups and between 1.5T and 3.0T. Cartilage-, fluid SNR and cartilage-fluid CNR of IDEAL group were compared with those of conventional group by using one-way analysis of variance with a Scheffe post hoc test.

The diagnostic performance of IDEAL–SPGR was significantly higher than that of SPGR(P: <.05). For reader 1 and 2, the sensitivity in IDEAL-SPGR for detecting cartilage lesions was 82.1% and 77.8%, specificity was 97.7% and 94.1%, and area under the curve(AUC) was 0.90 and 0.86. Diagnostic performance of MR imaging at 3.0T was not significantly higher than at 1.5T. Cartilage and fluid SNR of IDEAL group were significantly higher than those of conventional group in SPGR and PD-FSE. Cartilage-fluid CNR of IDEAL group were significantly higher than those of conventional group in all sequences. The SNRs and CNR were highest in IDEAL-SPGR at 3.0 T.

MR imaging with IDEAL increased the accuracy for detecting cartilage lesion and the SNR and CNR, when compared with those of conventional fat-saturation method. IDEAL-SPGR was the most accurate sequence for detecting cartilage lesions.

With IDEAL, cartilage assessment with MR imaging can be more accurate and IDEAL-SPGR can be a new clinical imaging protocol in patients with osteoarthritis or post traumatic cartilage defects.

Rhee, S, Cha, J, Kim, H, Yoo, J, Park, J, Lee, H, Hwang, M, Articular Cartilage of the Knee: Comparison of IDEAL and Conventional Fat Saturation MR Imaging with 3D SPGR and 2D FSE in a Porcine Model at 1.5 T and 3.0 T.  Radiological Society of North America 2009 Scientific Assembly and Annual Meeting, November 29 - December 4, 2009 ,Chicago IL. http://archive.rsna.org/2009/8014631.html