Cartilaginous end plates: Quantitative MR imaging with very short echo times-orientation dependence and correlation with biochemical composition

Abstract

Purpose: To measure the T2* of the human cartilaginous end plate by using magnetic resonance (MR) imaging with very short echo times and to determine the effect of the orientation of the end plate on T2* and on relationships between T2* and biochemical composition.

Materials and methods: This study was exempt from institutional review board approval, and informed consent was not required. Thirty-four samples of three cadaveric lumbar spines (from subjects who died at ages 51, 57, and 66 years) containing cartilaginous end plates and subchondral bone were prepared. Samples were imaged with a 3-T imager for T2* quantification by using a three-dimensional very short echo time sequence (repetition time msec/echo times msec, 30/0.075, 2, 5, 12, 18). Samples were imaged with the end plate at three orientations with respect to the constant magnetic induction field: 0°, 54.7°, and 90°. After imaging, the cartilage was assayed for its water, glycosaminoglycan, and collagen content. Pearson correlations were used to investigate the effect of orientation on the relationships between T2* and biochemical composition.

Results: T2* was significantly longer when measured at an orientation of 54.7° (21.8 msec ± 2.8 [± standard error of the mean]) than at 0° (10.0 msec ± 0.7, P < .001) or 90° (9.9 msec ± 0.4, P < .001). At 54.7°, T2* was highly correlated with glycosaminoglycan content (r = 0.85, P < .001), the collagen-to-glycosaminoglycan ratio (r = -0.79, P < .001), and water content (r = 0.62, P = .02); at 0° and 90°, there were no significant differences in these relationships, with a minimum P value of .19.

Conclusion: T2* evaluation can allow noninvasive estimation of the degeneration of the cartilaginous end plate; however, the accuracy of T2*-based estimates of biochemical composition depends on the orientation of the end plate.

Figure 1:

UTE images (top) with echo time of 75 μs and T2* maps (bottom) of a typical cartilaginous end plate sample imaged at 0° and 54.7° orientation show greater signal intensity and T2* at 54.7°. ms = Milliseconds.

Figure 2:

Scatterplots show relationships between mean T2* and biochemical composition for cartilaginous end plate samples imaged at 0° (left), 54.7° (middle), and 90° (right). T2* variation at 54.7° was significantly associated with GAG content, the collagen-to-GAG ratio, and water content. T2* variation at 0° and 90° (approximate in vivo orientation in the middle-to-upper lumbar spine) was not associated with biochemical composition. All end plate samples were harvested from the same spines; however, samples imaged at 0° and 54.7° (n = 14) were different from those imaged at 90° (n = 23). ms = Milliseconds.

Figure 3:

Bar graphs show comparison of T2* for surface, middle, and deep regions of cartilaginous end plate for samples imaged at different orientations. At 0° and 90°, all comparisons showed significant differences (at 0°: surface vs middle, P < .001; surface vs deep, P < .001; middle vs deep P = .05; at 90°: surface vs middle, P = .04; surface vs deep, P = .002; middle vs deep, P = .002). At 54.7°, T2* in the deep region was significantly shorter than in the other regions. (deep vs surface, P = .02; deep vs middle, P = .008). Data are mean ± standard error of the mean for 14 samples per group (0° and 54.7°) or 23 samples per group (90°). ms = Milliseconds.

Figure 4:

Midsagittal UTE image (left) with echo time of 292 μs from a 37-year-old male donor showed higher end plate signal intensity in lower lumbar region than in middle-to-upper lumbar region. For reference, cartilaginous end plate caudal to L5 disc was approximately 50°, the cartilaginous end plates cranial to L3 disc and caudal to L4 disc were approximately 90°. T2* map (right) of L4 and L5 discs in lower lumbar region showed T2* of 10–40 msec, which appeared to decrease with depth. ms = Milliseconds.