Quantitative MRI in early intervertebral disc degeneration: T1rho correlates better than T2 and ADC with biomechanics, histology and matrix content

Abstract

Introduction: Low-back pain (LBP) has been correlated to the presence of intervertebral disc (IVD) degeneration on T2-weighted (T2w) MRI. It remains challenging, however, to accurately stage degenerative disc disease (DDD) based on T2w MRI and measurements of IVD height, particularly for early DDD. Several quantitative MRI techniques have been introduced to detect changes in matrix composition signifying early DDD. In this study, we correlated quantitative T2, T1rho and Apparent Diffusion Coefficient (ADC) values to disc mechanical behavior and gold standard early DDD markers in a graded degenerated lumbar IVD caprine model, to assess their potential for early DDD detection.

Methods: Lumbar caprine IVDs were injected with either 0.25 U/ml or 0.5 U/ml Chondroïtinase ABC (Cabc) to trigger early DDD-like degeneration. Injection with phosphate-buffered saline (PBS) served as control. IVDs were cultured in a bioreactor for 20 days under axial physiological loading. High-resolution 9.4 T MR images were obtained prior to intervention and after culture. Quantitative MR results were correlated to recovery behavior, histological degeneration grading, and the content of glycosaminoglycans (GAGs) and water.

Results: Cabc-injected IVDs showed aberrancies in biomechanics and loss of GAGs without changes in water-content. All MR sequences detected changes in matrix composition, with T1rho showing largest changes pre-to-post in the nucleus, and significantly more than T2 and ADC. Histologically, degeneration due to Cabc injection was mild. T1rho nucleus values correlated strongest with altered biomechanics, histological degeneration score, and loss of GAGs.

Conclusions: T2- and T1rho quantitative MR-mapping detected early DDD changes. T1rho nucleus values correlated better than T2 and ADC with biomechanical, histological, and GAG changes. Clinical implementation of quantitative MRI, T1rho particularly, could aid in distinguishing DDD more reliably at an earlier stage in the degenerative process.

Fig 1. Process map showing the IVD specimen handling steps after dissection of the lumbar caprine IVDs from the spine.

Fig 2. Scheme of the daily simulated-physiological loading (SPL) regimes.

On the Y-axis the axial load (MPa) as applied on the IVDs. All IVDs started with 8 hours of low dynamic load (LDL) around 0.1 MPa, after which a 16 hour SPL loading regime was applied as indicated in the caption, this diurnal regime is repeated daily.

Fig 3. IVD Region definition.

High resolution T2w MRI of a mid-height transverse slice (left) with a schematic representation of the five ROIs drawn in the image on the right. The nucleus, inner- and outer-annulus region were separated by their signal intensity and morphological appearance on the transverse image. The anterior and posterior outer-annulus were defined as the most anterior and posterior 120o of outer-annulus laminae. The lateral section comprised the 60o sections in between anterior and posterior outer-annulus on each side. The mean signal intensity of the three outer-annulus regions was calculated from the signal intensity measured in two separate areas on opposite sides of the IVD. The posterior notch was omitted for analyzes because of lack of signal homogeneity.

Fig 4. Representative quantitative parameter maps (T2, T1rho, and ADC) of a mid-level transverse slice of 0.5 Cabc-injected IVD pre- and post-experiment.

Color scales are 0-150ms, 0-200ms, and 0–2.5x10-3 mm2/s, for T2, T1rho, and ADC, respectively.

Fig 5. Bar graphs showing mean ± SD post-to-pre differences (mean difference of sample paired post-experiment value to pre-experiment differences) for all three experimental groups.

Results are shown of the T2 (a) and T1rho (b) values in the nucleus and ADC in the outer-annulus; anterior (c), lateral (d) and posterior (e). PBS clearly had the least influence on quantitative MRI values for all three parameters, whereas 0.25 and 0.5 Cabc show larger loss of signal in the nucleus. Asterisk (*) indicates p-values below 0.05.

Fig 6. Correlation of post-experiment T2 (top row) and T1rho (bottom row) nucleus values with day 20 stretched-exponential parameters tau, beta and delta infinite.

Fig 7. Correlation of post-experiment MR T2, T1rho, and ADC nucleus values with Rutges histological degeneration score.

Fig 8. Correlation of post-experiment MR T2, T1rho, and ADC nucleus values to GAG and water content.