Quantitative analysis of the difference between an intact complete discoid lateral meniscus and a torn complete discoid meniscus on MR imaging: a feasibility study for a new classification

To determine the quantitative difference between an intact complete discoid lateral meniscus (CDLM) and a torn CDLM on MR imaging. Between May 2005 to November 2009, 137 patients with a CDLM (107 intact CDLM and 30 torn CDLM) and 92 patients with a normal meniscus were included in this study. The evaluated parameters were the height of the posterior horn of the lateral and medial menisci on the sagittal images and their ratio as assessed by two observers twice at an interval of 1 month. Each parameter was analyzed based on the Kruskal Wallis test, and the analysis using the mixed model. Intraclass correlation coefficient (ICC) was used to determine the interobserver reliabilities at session 2. The mean heights of the posterior horn of the lateral and medial menisci on the sagittal images for an intact CDLM, a torn CLDM, and a normal meniscus were 6.5, 7.3, 5.7 and 6.6, 6.4, 6.7 mm at session 1, respectively. The mean heights of the posterior horn of the lateral and medial menisci on the sagittal images for an intact CDLM, a torn CDLM, and a normal meniscus for both observers were 6.5, 7.2, 5.7 and 6.6, 6.3, 6.8 mm at session 2, respectively. The ratio of the height of the lateral to the height of the medial meniscus for an intact CDLM at both sessions for both observers was 1.0. The ratios were 1.2 and 0.8 for a torn CDLM and for a normal meniscus, respectively, at both sessions for observer 1. The ratios were 1.2 and 0.9 for a torn CDLM and for a normal meniscus, respectively, at session 2 for observer 2. The heights of the posterior horn of the lateral meniscus on the sagittal images and the ratios of the heights of the lateral to the medial menisci in all three groups were statistically significantly different for both sessions (p < 0.0001). The interobserver ICCs for each parameter of both an intact CDLM and a torn CDLM at session 2 showed high correlations (p < 0.0001). The height of the lateral meniscus and the ratio of the height of the lateral to the height of the medial meniscus for a torn CDLM were significantly higher than those for an intact CDLM.