Introduction

Foraminal stenosis often causes nerve root entrapment in patients with lumbar disc herniation, which results in low back pain [1]. Ultrashort echo time magnetization transfer (UTE-MT) imaging can quantify macromolecular content changes in tissues [2]. This study aims to investigate the feasibility of UTE-MT ratio (UTE-MTR) in assessment of lumbar nerve root changes due to disc herniation. T₂ mapping and diffusion tensor imaging (DTI) were also employed for comparison.

Methods

A total of 75 patients with disc herniation (age 43 ± 16years, age range 35-78) were recruited and scanned with UTE-MT, DTI and CPMG sequences in examination of L4/5 and L5/S1 lumbar discs on a 3T MRI scanner (Signa, Pioneer, GE Healthcare). Informed consent was obtained from all participants in accordance with the Institutional Review Board. A Fermi pulse was employed to generate the MT contrast in UTE-MT sequence with duration of 8ms and bandwidth of 160Hz. The frequency offset of this MT pulse was 1500 Hz. The UTE-MT sequence was scanned twice with flip angle of 750° for MT-On and 0° for MT-Off. Other UTE-MT sequence parameters were as follows: TR = 100 ms, TE = 0.032 ms, excitation flip angle = 5°, number of spokes per-TR = 5, FOV = 28cm × 28cm, matrix = 140 × 140, slice thickness = 3.6mm, slice number = 16, bandwidth = 125 kHz, oversampling factor = 1.2, and scan time = 3min. The UTE-MTR is calculated by the signal ratio of the difference between UTE-MT-OFF and UTE-MT-ON to the UTE-MT-OFF. The DTI sequence parameters were as follows: TR = 6000ms, TE = 60ms, FOV = 42cm × 42cm, matrix = 140 × 140, slice thickness = 4mm, slice number = 18, bandwidth = 250 kHz, b-values of 0 and 800 in each of six diffusion gradient encoding directions, and scan time = 3 min. 2D CPMG sequence was used for T₂ mapping with parameters as follows: TEs = 7, 14, 21, 28, 35, 42, 49, and 56 ms, TR = 1400 ms, FOV = 38cm × 38cm, matrix = 140 × 140, slice thickness = 4mm, slice number = 20, bandwidth = 250 kHz, and scan time = 5 min.
Disc herniation was identified in consensus by two musculoskeletal radiologists with 6 and 8 years of experience. 82 herniated discs were found from a total of 150 discs. The region of interest (ROI) of the nerve roots was manually drawn by a radiologist with 11 years of experience. The student’s t-test was used to compare UTE-MTR, T₂ and fractional anisotropy (FA) values in nerve roots between patients with and without disc herniation. Receiver operating characteristic (ROC) analysis and computed the area under the curve (AUC) with 95% CI was to evaluate the performances of the UTE-MT, T₂ and DTI in discriminating between patients with and without disc herniation. A value of P < 0.05 was considered as a statistically significant.

Results

Figure 1 shows the representative axial L4/5 lumbar disc UTE-MT images acquired from a 45-year-old male subject. The UTE-MTR values were significantly increased while FA and T₂ values were significantly decreased in patients with lumbar disc herniation compared to those without disc herniation (Figure 2). The ROC curves showed that the UTE-MTR performed better than FA and T₂ in discrimination of disc herniation and non-disc herniation (with AUC of UTE-MTR = 0.948, AUC of FA = 0.682 and AUC of T₂ = 0.798) (Figure 3).

Discussion and Conclusion

To our best knowledge, this is the first study to apply the UTE-MT imaging to assess nerve root changes due to disc herniation in lumbar. Our findings suggested that patients with disc herniation have significantly higher UTE-MTR values in lumbar nerve roots. These UTE-MTR value increases in the nerve roots may be related to the decrease of proteoglycan content. In addition, compression of nerve roots induces blood flow reduction and ischemia, leading to nerve fiber demyelination, Wallerian degeneration and endoneural cracking [3], which may explain the decrease of FA and T₂ map. Higher AUC values of UTE-MTR demonstrates that UTE-MT imaging may be a better biomarker than DTI and T₂ mapping in assessment of nerve root changes due to disc herniation in lumbar.

Acknowledgements

No acknowledgement found.