|Home | About | Journals | Submit | Contact Us | Français|
T1ρ relaxation was quantified and correlated with intervertebral disc degeneration and proteoglycan content in cadaveric human lumbar spine tissue.
To show the use of T1ρ-weighted magnetic resonance imaging (MRI) for the assessment of degeneration and proteoglycan content in the human intervertebral disc.
Loss of proteoglycan in the nucleus pulposus occurs during early degeneration. Conventional MRI techniques cannot detect these early changes in the extracellular matrix content of the disc. T1ρ MRI is sensitive to changes in proteoglycan content of articular cartilage and may, therefore, be sensitive to proteoglycan content in the intervertebral disc.
Intact human cadaveric lumbar spines were imaged on a clinical MR scanner. Average T1ρ in the nucleus pulposus was calculated from quantitative T1ρ maps. After MRI, the spines were dissected, and proteoglycan content of the nucleus pulposus was measured. Finally, the stage of degeneration was graded using conventional T2 images.
T1ρ decreased linearly with increasing degeneration (r=−0.76, P < 0.01) and age (r=−0.76, P < 0.01). Biochemical analysis revealed a strong linear correlation between T1ρ and sulfated-glycosaminoglycan content. T1ρ was moderately correlated with water content.
Results from this study suggest that T1ρ may provide a tool for the diagnosis of early degenerative changes in the disc. T1ρ-weighted MRI is a noninvasive technique that may provide higher dynamic range than T2 and does not require a high static field or exogenous contrast agents.
Degenerative disc disease afflicts nearly 12 million people in the United States. Although a single initiating cause of degeneration has not been identified, early degenerative changes occur in the nucleus pulposus.1,2 Breakdown of the large aggregating proteoglycans reduces the capacity of the nucleus pulposus to attract and bind water, leading to a loss of disc hydration and decreased hydrostatic pressure.3,4 Ultimately, degeneration progresses to decreased disc height, structural changes in the lamellar architecture of the anulus fibrosus, anular tears and rim lesions, and the formation of osteophytes.5-7 The success of treatment strategies aimed at halting the progression of disc degeneration will require detection of the early stages of the disease, particularly changes in the extracellular matrix content of the nucleus pulposus.
Conventional magnetic resonance imaging (MRI) techniques provide excellent detection of late-stage degenerative changes (i.e., changes in disc morphology, height, hydration, bulge, and herniation).8,9 However, these methods are not sensitive to early degenerative changes in the matrix content of the disc.10 Delayed gadolinium-enhanced MRI has been used to quantify proteoglycan in articular cartilage.11,12 However, the contrast agent must be administered intravenously, and diffusion into the cartilage requires a long time.13 This is a significant limitation in the avascular intervertebral disc, in which the negative fixed charged density of the nucleus pulposus hinders diffusion of ionic contrast agents.14,15 Sodium MRI has also been used to measure proteoglycan in articular cartilage.16-18 However, its clinical use is somewhat limited by a low spatial resolution and the need for instrumentation modifications for use on a clinical scanner.
Spin-lock MRI techniques have been used to provide noninvasive measures of degeneration in articular cartilage and may potentially be used to assess degeneration in the intervertebral disc. Spin-lock pulses are low power rf pulses applied directly on-resonance with the Larmor precession frequency, locking the magnetization vector into a rotated frame. The relaxation that occurs after the application of a spin-lock pulse is referred to as spin-lattice relaxation in the rotating frame, or T1ρ relaxation. Spin-lock allows the coupling of spins to frequencies that are generally lower than the Larmor frequency. Therefore, slow motion regimes can be studied, such as low frequency physicochemical interactions between water and extracellular matrix molecules. Thus, matrix changes, such as loss of proteoglycan, will be reflected in the T1ρ parameter. T1ρ-weighting provides T2-like images with the advantage of increased dynamic range to degenerative changes compared to conventional T2-weighting.19
In articular cartilage, T1ρ is strongly correlated with proteoglycan content and, thus, has been shown to detect early osteoarthritic changes.20-22 Recently, T1ρ-weighted images of bovine intervertebral disc tissue have been acquired.23 However, a relationship between T1ρ and intervertebral disc degeneration has not been established, nor has it been shown that T1ρ is sensitive to proteoglycan content in the disc. Thus, the objective of the present study was to demonstrate the use of T1ρ MRI for the assessment of degeneration and proteoglycan content in the human intervertebral disc. Quantitative T1ρ measurements were obtained from cadaveric human intervertebral disc tissue. Degenerative grade was assessed from standard T2-weighted images, and tissue was subsequently analyzed for total sulfated-glycosaminoglycan content, a measure of proteoglycan content.
Seven fresh-frozen cadaveric human lumbar spine sections (mean age 51.6 years, range 15–81) were imaged on a 1.5 T whole-body clinical MR scanner (Sonata; Siemens Medical Solutions). A series of T1ρ-weighted images was acquired using a self-compensating turbo spin-echo sequence with parameters: FOV = 28 × 28 cm; slice thickness = 4 mm; acquisition matrix = 512 × 512; and TE/TR = 3000 milliseconds/12 milliseconds. There were 5 evenly distributed spin-lock pulse durations from 15 to 75 milliseconds. The spin-lock pulse amplitude was set to 500 Hz.
T1ρ values were calculated on a pixel-by-pixel basis by a linear regression of intensity data to an exponential decay function: S(TSL) = S0e−TSL/T1ρ. Values were used to create spatial maps of T1ρ. A circular region (5-mm diameter) was manually segmented from the center of the nucleus pulposus, and mean T1ρ was computed within that region.
Diagnostic clinical T2-weighted images were acquired and used for assessment of degenerative grade for each disc (n = 35). Two orthopedic surgeons and a radiologist independently performed the grading according to the classification scale described by Pfirrmann et al.24 Briefly, discs were graded on a 1–5 integer scale, with grade 1 corresponding to nondegenerate, healthy tissue and grade 5 corresponding to severely degenerate tissue (Table 1).
Discs were isolated via sharp dissection, and 1.5-mm punches were harvested from the center of the nucleus pulposus for biochemical analysis. Water content was determined by weighing samples before and after 5 days of incubation at 65°C. Dried samples were then digested in proteinase-K solution, and determination of sulfated-glycosaminoglycan content was performed using 1,9-dimethylmethylene blue in a micro-plate reader assay.25
Linear regressions among degenerative grade, age, T1ρ,water content, and sulfated-glycosaminoglycan content were performed using GraphPad Prism software (GraphPad Software, San Diego CA). Significance was set at P < 0.05, and correlations were considered strong for r > 0.7, moderate for 0.5 < r ≤ 0.7, and weak for r ≤ 0.5.26
Representative T1ρ-weighted images and corresponding quantitative T1ρ maps are shown for 2 spine sections (donor ages 25 and 51 years) (Figure 1). The T1ρ-weighted image is a standard MRI acquired from a single spin-lock pulse. In contrast, the T1ρ map is a graphical representation of the quantitative T1ρ parameter calculated at each pixel location. In general, T1ρ values were higher in the younger, nondegenerate discs. Based on assessment of T2 images, degenerative grades ranged from 1 to 5. Severely degenerate discs (i.e., grades 4 and 5) were excluded from the study because late-stage degenerative changes can be reliably detected using conventional MRI methods, and insufficient tissue was available for biochemical analysis at this late stage.
Therefore, T1ρ measurements were only calculated from discs with grades ≤3.5 (n = 17). T1ρ values ranged from 45 to 173 milliseconds. There was a strong correlation between T1ρ and degenerative grade (r = −0.76, P < 0.01) (Figure 2). T1ρ was strongly correlated with sulfated-glycosaminoglycan per wet weight (r = 0.70, P < 0.01), and was moderately correlated with sulfated-glycosaminoglycan per dry weight (r = 0.67, P < 0.01) and water content (r = 0.58, P < 0.05) (Figure 3). There was a strong correlation between T1ρ and age (r = −0.76, P < 0.01).
Early degenerative changes in the disc occur in the nucleus pulposus, where proteoglycan content decreases. Spin-lock MRI techniques, such as quantitative T1ρ measurements, may provide a noninvasive method to detect proteoglycan content, allowing detection of early stages of degeneration (Figure 4). As more disc treatment options are developed, including biologic treatments,27,28 nucleus pulposus replacements,29 and total disc replacement,30 better and more sensitive imaging methods will be required to direct the surgeon toward the appropriate treatment. The most important finding of this study was that T1ρ is linearly related to sulfated-glycosaminoglycan content, suggesting that T1ρ may be sensitive to proteoglycan content in the intervertebral disc. We also found that T1ρ MRI was correlated to degenerative grade. Current grading schemes are susceptible to observer bias and are limited in their ability to detect subtle changes because they are based on a limited 5-level integer scale. In addition, these schemes lack the ability to localize degenerative changes within the disc substructures. In contrast, T1ρ provides a spatial, quantitative measurement that may be more sensitive to early degenerative changes.
The use of quantitative imaging of the intervertebral disc has been investigated using T1 and T2 relaxation, magnetization transfer, spectroscopy, and diffusion measurements.31-37 Boos et al38 found that asymptomatic disc herniations showed shorter T1 and T2 relaxation times than symptomatic herniations, although the differences were considered small with respect to the reproducibility of measurement in vivo. In the nucleus pulposus, noncollagenous proteins and granular tissue increase with age and degeneration.1 This fibrosus of the tissue may dampen the signal, shortening T2 substantially. Because T1ρ is always higher than T2, the increased dynamic range provided by T1ρ MRI may be particularly beneficial in the disc.
We found that T1ρ in the nucleus pulposus is directly correlated to proteoglycan content. However, in articular cartilage, T1ρ is inversely correlated to proteoglycan content and fixed charge density. Although both disc nucleus pulposus and articular cartilage are primarily comprised of water, proteoglycan, and type II collagen, the relative composition and degenerative processes of these tissues differ.39 Specifically, other matrix constituents (i.e., collagen, water, degree and type of collagen cross-linking, and proteoglycan aggregation) may influence T1ρ values in the disc. Therefore, future work will be directed at identifying the contributions of individual matrix components to T1ρ. This initial study was limited to a relatively small number of cadaveric samples. However, results from this study highlight the potential use of T1ρ-weighted imaging of the intervertebral disc, and we recently completed in vivo imaging studies.39a We found that T1ρ was strongly correlated with age in the present study; future application of this technique to a large patient population of both symptomatic and nonsymptomatic individuals may enable us to separate degeneration from age-related changes in the disc. Ultimately, the content and organization of the extracellular matrix determine the mechanical function of the disc,40-42 thus, T1ρ may potentially be used to assess disc mechanics.
T1ρ is a promising technique for the in vivo diagnosis of intervertebral disc degeneration. It may be particularly useful in detecting the early stages of the disease and directing the appropriate treatment option. Conventional MRI is well suited for detecting the more dramatic changes in disc morphology and hydration that occur in late-stage degeneration. Correlations between T1ρ and sulfated-glycosaminoglycan suggest that T1ρ maybesensitive to proteoglycan changes in the intervertebral disc. Future in vivo imaging studies will seek to determine the diagnostic capabilities of T1ρ for early disc degeneration and subsequent low back pain.
The authors thank Drs. Miltiadis Zgonis and Carol Dolinskas for their assistance with disc grading.
Federal funds were received in support of this work. No benefits in any form have been or will be received from a commercial party related directly or indirectly to the subject of this manuscript.