The current study demonstrated that the occurrence of lumbar stenosis complicated with RNRs as assessed by MR, affected approximately 36.2% (42/116 )of patients, This result is consistent with prior literature, which indicated that the incidence of lumbar stenosis complicated with RNRs was ~33.8%–42.3% [9, 14]. Mechanical compression of patients with lumbar spinal stenosis is considered to be the basic mechanism involved in RNRs formation. Suzuki et al performed autopsies on patients with severe lumbar spinal stenosis, and the results demonstrated that cauda equina RNRs formed strands in the stenosis. In the spinal canal, these compressed nerve roots are significantly elongated as they stretch with the legs and trunk, eventually leading to their over-elongation . In addition, other scholars believe that RNRs may be related to an abnormality of the cauda equina microvascular function, and that ischemia of the cauda equina could lead to tortuous never root deformation .
Formerly, the diagnosis of RNRs was mainly assessed by myelography. At present, MRI has provided an important basis for the diagnosis of RNRs, although a deeper understanding and research of RNRs needed. T2-weighted MRI scans usually indicate RNRs as tortuous and entangled, with an overall narrowing of the spinal canal ,The MRI imaging characteristics of RNRs are characterized by high-signal images of spinal canal relaxation, tortuosity, and entanglement on T2-weighted images. In the current study, we found that RNRs were more likely to occur in stenotic segments of L4- 5 and above. We consider that there are more cauda equina nerves in the spinal canal at higher segments, which are more likely to form RNRs after compression. Hur et al demonstrated that during the development of lumbar spinal stenosis the level of L4-5,which is the area most associated with spondylolisthesis,multi-segment stenosis and spondylolisthesis may be important factors for the formation of RNRs in central lumbar spinal stenosis.Yukawa et al found that patients with two-segment or multiple-segment stenosis walked much shorter distances than patients with perioperative single-segment stenosis. In the current study, there was no statistically significant difference between the RNRs group and the NRNRs group regard to preoperative back pain and leg pain VAS scores, and preoperative ODI scores . These results are consistent with the research by Min and Hur[4, 14]. However, our study indicated tha there was a significant difference in the duration of back pain symptoms between the two groupst , In previous studies, it has also been demonstrated that the older patient and longer the duration of symptoms, the greater was the possibility of developing RNRs [9-10, 14]. However,
However, the results of this study indicate that age is not associated with RNRs.Savarese et al.  showed that the probability of RNRs in patients with lumbar spondylolisthesis is 55.56%, whereas the probability of RNRs in patients without lumbar spondylolisthesis is 23.61%, demonstrating that patients with lumbar spondylolisthesis are more prone to RNRs. Our study is consistent with the findings of Savarese et al. Many studies have shown that ligamentum flavum hypertrophy is the main cause of lumbar spinal stenosis. At the same time, hypertrophy of the ligamentum flavum is closely related to the mechanical stress of the cauda equina [20-24]. In the current study, the probability of hypertrophy of the ligamentum flavum was significantly higher in the RNRs group than in the NRNRs group. This result is in agreement with previous literature[13-14]. Hur et al.  indicated that patients with hypertrophy of the ligamentum flavum in the spinal canal had a stronger relationship with the occurrence of RNRs than with other structures in the spinal canal.
The appearance of RNRs is closely related to lumbar spinal stenosis . In the current study, we used DIW-MSD and DV-MSD were used to evaluate the degree of lumbar spinal stenosis. There are many methods of CT to evaluate lumbar spinal stenosis. The simplest and most commonly used method is to measure the median sagittal diameter. There are many methods of CT measurement to evaluate lumbar spinal stenosis. The simplest and most commonly used method is MSD.In previous studies, the MSD of the spinal canal was mostly taken at the vertebral level as it was inconsistent at the intervertebral space level and the vertebral level. We measured DIW-MSD and DV-MSD and found statistically significant differences between the two groups when DIW-MSD was used; whereas none were found when DV-MSD was used. These results indicate that the formation of RNRs in patients with lumbar spinal stenosis is caused by spinal stenosis at the intervertebral space level. We believe that the occurrence of disc herniation, hypertrophy of the ligamentum flavum, and fat behind the dural sac results in spinal stenosis of the MSD at the intervertebral space level; a hypothesis that is consistent with the conclusions of Savarese et al.  and Hur et al. .
Intervertebral height is measured by either taking the height of the anterior, middle, or posterior edges. Here, we measured the height of the posterior edge, as the posterior edge and the height of the vertebral body composed the length of the spinal canal. To date, there have been no reports that the intervertebral height, intervertebral foramen height, and the intervertebral height + vertebral height are associated with formation of RNRs in patients with lumbar spinal stenosis. In addition, in order to avoid individual differences, we used the ratio of intervertebral height and vertebral body height. Because the ratio is relatively constant, we used three values to evaluate the spinal canal length index. Results demonstrated that length indices, including intervertebral height, intervertebral foramen height, and intervertebral height + vertebral height in the RNRs group were less than those in the NRNRs group. We conclude that the length of the spinal canal is shortened due to the decrease in intervertebral and intervertebral foramen height, which eventually the spinal canal length shortens. However, The length of the cauda equina actually not changed, RNRs are more likely to occur.
The curvature is an important feature in the structure of the spine and maintenance of spine curvature is closely related to the prevention of spinal diseases . Ono et al.  found that the incidence of RNRs was higher in the neutral position of the lumbar spine than in the flexed position, when performing myelography. Mendelsohn et al.  observed that RNRs are aggravated during back extension and reduced during flexion. These results indicate that lumbar lordosis and lumbar spine mobility may be related to the occurrence of RNRs. In the current study, we demonstrated that, although not significant, ROM of the stenotic segment of patients in the RNRs group was larger than that in the NRNRs group.