Morphology of the cartilaginous endplates in human intervertebral disks with ultrashort echo time MR imaging

Abstract

Purpose: To image human disk-bone specimens by using conventional spin-echo (SE) and ultrashort echo time (TE) techniques, to describe the morphology at magnetic resonance (MR) imaging, and to identify tissue components contributing to high signal intensity near the cartilaginous endplates (CEPs).

Materials and methods: This study was exempt from institutional review board approval, and informed consent was not required. Five cadaveric lumbar spines (mean age, 61 years ± 11) were prepared into six sample types containing different combinations of disk, uncalcified CEP, calcified CEP, and subchondral bone components and were imaged with proton density-weighted SE (repetition time msec/TE msec, 2000/15) and ultrashort TE (300/0.008, 6.6, echo-subtraction) sequences. Images were evaluated to determine the presence of intermediate-to-high signal intensity in regions excluding the bone marrow. Logistic regression was used to determine which tissue components were significant predictors of the presence of signal intensity for each MR technique.

Results: On ultrashort TE MR images, intact disk/uncalcified CEP/calcified CEP/bone samples exhibited bilaminar intermediate-to-high signal intensity in the region near the CEP, consistent with the histologic appearance of uncalcified and calcified CEPs. Conversely, proton density-weighted SE images exhibited low signal intensity in this region. Results of logistic regression suggested that the presence of uncalcified CEP (P = .023) and calcified CEP (P = .007) in the sample were strong predictors of the presence of signal intensity on ultrashort TE images, whereas the disk was the only predictor (P < .001) of signal intensity on proton density-weighted SE images.

Conclusion: Ultrashort TE imaging, unlike proton density-weighted SE imaging, enabled direct visualization of the uncalcified and calcified CEP. Evaluation of the morphology and identification of sources of signal intensity at ultrashort TE MR imaging provides opportunities to potentially aid in the understanding of degenerative disk disease.

Figures

Figure 1:

MR images of cadaveric spine (L2 through L4) from 62-year-old man. A, proton density–weighted MR image (TE, 10 msec) and, B, T2-weighted MR image (TE, 80 msec) obtained with conventional SE sequence (repetition time, 2000 msec; field of view, 16 cm [cropped to 6 × 7.5 cm]; matrix, 320 × 256; section thickness, 2 mm; bandwidth, ±25 kHz). Although T2-weighted image was useful for evaluating intervertebral disks, which were mildly degenerate in this sample (Pfirrmann grade 2–3), as well as bone marrow changes in vertebral body and morphology of spinal cord, SE images generally exhibited low signal intensity in region near CEP (arrows). C, Conversely, ultrashort TE echo-subtraction MR image (300/0.008, 6.6) shows linear high signal intensity in region near CEP (arrows) and longitudinal ligaments (curved arrows). Bar = 1 cm. ♦ = intervertebral disks, ▲ = spinal cord, ● = vertebral body.

Figure 2:

Sample preparation. A, Lumbar spines were resected to create rectangular piece (∼1 cm3) containing disk, CEP, and bone. These were further resected to prepare samples containing, B, disk only, C, disk/uncalcified CEP, and, D, uncalcified CEP/calcified CEP/bone. E, To create calcified CEP/bone samples, both resection and chemical treatment (ie, removal of uncalcified CEP with 3% sodium hypochlorite and scrubbing) was used. F, To create bone-only samples, serial resection down to subchondral bone was performed. Sample preparation was verified with histologic examination at end of experiment.. cCEP = calcified CEP, uCEP = uncalcified CEP.

Figure 3:

Images of disk/uncalcified CEP/calcified CEP/bone sample. A, Proton density–weighted SE image shows that region near CEP (arrows) has low and heterogeneous signal intensity indistinct from that of some parts of bone marrow (●). Disk region (♦) has high signal intensity. B, Ultrashort TE MR image, conversely, shows that region near CEP (arrows) has intermediate-to-high signal intensity, with a bilaminar appearance: a lower region with a thin, linear, and high signal intensity (up arrows) and an upper region with variable thickness and intermediate signal intensity (down arrows). Disk region (♦) has low signal intensity. ● = bone marrow. C–E, Photomicrographs (hematoxylin-eosin stain; original magnification, ×40) confirm presence of layers of disk, uncalcified CEP with variable thickness, thin calcified CEP layer, and subchondral bone and demonstrate that the morphology of CEP region on ultrashort TE MR image is consistent with histologic appearance. Bars = 200 μm.

Figure 4:

A–E, Conventional proton density–weighted SE MR images, F–J, ultrashort TE MR images, and K–O, photomicrographs (hematoxylin-eosin stain; original magnification, ×40) of prepared samples including tissue components of disk only (A, F, K), disk/uncalcified CEP (B, G, L), uncalcified CEP/calcified CEP/bone (C, H, M), calcified CEP/bone (D, I, N), and bone only (E, J, O). Sample preparations were validated with histologic examination, and presence or absence of intermediate-to-high signal intensity in regions of tissue other than bone marrow (●) was evaluated on MR images. On conventional proton density–weighted SE MR images, disk only (A) and disk/uncalcified CEP (B) samples demonstrate high signal intensity in disk region (♦). Conversely, ultrashort TE MR images of disk/uncalcified CEP (G), uncalcified CEP/calcified CEP/bone (H), and calcified CEP/bone (I) samples show linear intermediate-to-high signal intensity in region of uncalcified and calcified CEP. The region of subchondral bone (▲ in E and J) did not have linear signal intensity with either imaging technique. Bar = 2 mm. Arrows = calcified and/or uncalcified CEP. cCEP = calcified CEP, uCEP = uncalcified CEP.

Figure 5:

Three samples containing disk/calcified CEP/bone were found incidentally during the study. A, Conventional proton density–weighted SE MR image of disk/calcified CEP/bone sample demonstrates a thin and linear signal void near region of calcified CEP (arrows), which was indistinct from subchondral bone. B, Ultrashort TE MR image, however, shows a thin, linear area of high signal intensity in that region (arrows), which is different from bilaminar morphology found on ultrashort TE MR images of disk/uncalcified CEP/calcified CEP/bone samples. C, D, Photomicrographs (hematoxylin-eosin stain; original magnification, ×40) show three layers of tissues, including disk (which also contained round cells typically seen in uncalcified CEP), calcified CEP (cCEP), and bone. Bars = 200 μm. ♦ = region of disk, ● = region of bone marrow.

Figure 6:

Graph shows proportion of samples exhibiting intermediate-to-high signal intensity in regions other than bone marrow for each sample type and pulse sequence. Bars represent upper and lower 95% binomial confidence intervals. In general, conventional proton density–weighted SE images (SE) exhibited signal intensity when samples contained disk tissue (ie, disk/uncalcified CEP/calcified CEP/bone, disk-only, and disk/uncalcified CEP samples), whereas ultrashort TE MR images (UTE) exhibited signal intensity when samples contained either uncalcified or calcified CEP. Many proportions were significantly different from null hypothesis value of 0.5. cCEP = calcified CEP, uCEP = calcified CEP.