INTRODUCTION
Cervical spondylotic radiculopathy is a common degenerative spinal disorder with symptoms of numbness and radicular pain in the neck, shoulder, and upper extremities [1]. Decreased disk height or degenerative changes of the uncovertebral joints anteriorly or facet joints posteriorly are common contributors [2]. Leveraging the advancements in microchannel technology, endoscopic techniques, and minimally invasive principles, posterior endoscopic cervical discectomy (PECD) has gained increasing prominence in the last decade for treating cervical degenerative disorders [3,4]. The mainstream of PECD is the resection of dorsal bone and removal of free nucleus pulposus without addressing neuroventral osteophytes or hyperplastic ligaments. Recently, whether and how to perform direct neuroventral decompression has attracted the attention and controversy of surgeons [3,5,6]. However, there is no relevant reports on the direct neuroventral decompression approach selection in PECD. Additionally, there is scant literature addressing the anatomical relationship between nerve roots and the intervertebral space, with a notable absence of relevant radiologic studies in living specimens.
Although routine magnetic resonance imaging (MRI) and computed tomography (CT) examinations serve as essential tools for identifying cervical degenerative disorders, MRI suffers from insufficient spatial and bone tissue resolution, while routine CT faces limitations in nerve roots and intradural rootlets resolution. However, CT myelography (CTM), an ancient clinical technique combining CT scanning and myelography, can not only effectively display bone structure, but also make the dural sac and nerve rootlets visible [7-9]. Consequently, the purpose of this study is to clarify the anatomical relation among the nerve roots, intervertebral space, pedicles, and intradural rootlets of the cervical spine in living specimens by combining CTM and multiplane reconstruction (MPR) techniques, and to offer an anatomical basis for precise determination of the intervertebral space location and the selection of the neuroventral decompression approach.
MATERIALS AND METHODS
1. Inclusion and Exclusion Criteria
This study was approved by the Institutional Review Board (IRB) of the Second Affiliated Hospital of Chongqing Medical University (IRB No. 76).
The inclusion criteria encompassed the cervical CTM imaging data of patients from January 2021 to May 2023. The exclusion criteria were as follows: (1) cervical kyphosis deformity (defined as a sagittal C2–7 Cobb angle exceeding 5° on lateral radiographs), cervical scoliosis deformity (defined as a C2–7 Cobb angle exceeding 10° on anteroposterior radiographs), or congenital anomalies (such as block vertebra); (2) cervical bony structure destruction; (3) history of previous cervical surgery; (4) poor myelography; (5) ossification of the cervical posterior longitudinal ligament (OPLL); and (6) cervical spondylolisthesis (defined as a vertebral displacement exceeding 3 mm).
2. Myelography
After confirming the absence of contraindications for myelography, patients assumed a lateral recumbent position with hands folded around their knees, followed by a standard lumbar subarachnoid puncture. Subsequently, an equivalent volume of Iohexol (Omnipaque 300 mg/mL) was injected after releasing 8–10 mL of cerebrospinal fluid [7]. Finally, patients underwent a cervical CT scan using the SIEMENS SOMATOM Force device, with 1.25-mm slices, following a head-down position for 60 to 90 minutes.
3. Standardized Observation Plane
After importing the original data in DICOM format into a personal computer (PC), the RadiAnt DICOM Viewer Software (ver. 2021.1, Medixant, Poznan, Poland) was utilized for MPR, with adjustments to the bone windows. Subsequently, the standardized observation plane is established through the following steps (Supplementary video clip 1): Step 1 –After locating the target intervertebral space, orient the transverse axis to intersect the center of the pedicles on both sides of the lower vertebral body on the coronal plane, adjust the sagittal axis to the median sagittal position on the transverse plane, and ensure the coronal axis is parallel to the posterior wall of the lower vertebral body on the sagittal plane (Fig. 1A-C); Step 2 –On the transverse plane, relocate the intersection of the sagittal axis and the coronal axis to the center of one side's pedicle, and ensure the coronal axis is approximately perpendicular to the bone cortex of the pedicle. Simultaneously, observe on the coronal plane to confirm that the intersection of the sagittal and transverse axes is positioned at the center of the pedicle (Fig. 1D-F); Step 3 –On the coronal plane, shift the intersection of the transverse axis and the sagittal axis to the highest point of the inferior edge of the nerve root. Simultaneously, define the intersection of the sagittal axis and the superior edge of the nerve root as the highest point of the superior edge of the nerve root (Fig. 1G-I); Step 4 –On the sagittal plane, align the coronal axis with the bone cortex of the posterior wall of the lower vertebral body. Concurrently, on the coronal plane, designate the intersection of the transverse axis and the sagittal axis as the projection point of the highest point of the inferior margin of the nerve root (Fig. 1J-L); In addition, building upon step 3, the coronal axis was adjusted to make it consistent with the course of the ventral or dorsal rootlets on the transverse plane. Then the descending course of ventral or dorsal rootlets were clearly visible on the coronal plane (Fig. 2A and B).
4. Measurement Index
Based on the above standardized observation plane, the axes were adjusted to be perpendicular to the PC screen on the coronal plane. Subsequently, measurements were independently conducted at each level by 3 observers, and the results were averaged after excluding the data with only one side. To assess reproducibility, 3 observers repeated measurements of C4–5 segment for twenty randomly selected patients. The following parameters were defined and recorded (Fig. 2): (1) width of nerve root (WN); (2) the nerve root-superior pedicle distance (NSPD) (the shortest distance between the highest point of the superior edge of the nerve root and the superior pedicle); (3) the nerve root-inferior pedicle distance (NIPD) (the minimum distance between the highest point of the inferior edge of the nerve root and the inferior pedicle); and (4) the descending angles between the intradural rootlets and the longitudinal axis of the spinal cord, including the uppermost ventral and dorsal rootlets angle (UVRA and UDRA), the middle ventral and dorsal rootlets angle (MVRA and MDRA), and the lowest ventral and dorsal rootlets angle (LVRA and LDRA). Additionally, the distance between the projection point of the highest point of the inferior margin of the nerve root and the upper and lower margins of the intervertebral space was measured respectively. Hence, the position relationship of the intervertebral space relative to the corresponding nerve root can be categorized into 3 types: shoulder, anterior, and axillary (Fig. 3). If the lower edge of the intervertebral space is caudal to the projection point, it means that the intervertebral space is located in the axilla of the nerve root. Alternatively, if the lower edge is cranial to projection point, and the distance between the projection point and the upper edge of the intervertebral space is less than the WN, it indicates that the intervertebral space is positioned in front of the nerve root; otherwise, it is on the shoulder.
5. Statistical Analysis
Continuous variables were presented as the mean±standard deviation and categorical variables were expressed as frequency or percentages. Statistical differences were analyzed using the Mann-Whitney U-test, the Wilcoxon test, and the chi-square test. Intraclass correlation coefficient (ICC) was used to test the consistency of the interobservers. All statistical tests were performed using IBM SPSS Statistics ver. 27.0 (IBM Co., Armonk, NY, USA). The statistical significance level was set at p<0.05.
RESULTS
1. Patients
The CTM imaging date of 87 patients (40 males and 47 females) were retrospectively analyzed according to the inclusion and exclusion criteria. The mean age for males was 50.1±10.2 years (range, 32–75 years), and for females, it was 53.1±9.8 years (range, 35–76 years). The age and sex distribution of patients included in each segment exhibited remarkable consistency for each index (Supplementary Tables 1–3). ICCs indicated excellent agreement for all measurements (Supplementary Table 4).
2. Width of Nerve Root
The mean WN exhibited a progressive increase from the C4 (5.29±0.66 mm) to the C7 (6.67±0.83 mm) nerve roots, with a higher measurement observed in males across all segments (p<0.05). Moreover, the measurements were larger in ≤ 50 years age group compared to the > 50 years age group across the C4 to C8 nerve roots, with significant differences observed at the C6 and C8 nerve roots (p<0.05) (Table 1).
3. Shortest Distance Between the Nerve Root and the Corresponding Pedicle
The mean NSPD decreased as the level descended from C2–3 (3.66±1.62 mm) to C5–6 (2.78±1.45 mm), while the mean NIPD increased gradually from C2–3 (3.06±1.53 mm) to C5–6 (4.07±1.86 mm). In the lower cervical spine, the mean NSPD was smaller than the NIPD at the same level, with a statistical difference observed at the C4–5, C5–6, and C7–T1 levels. The mean NSPD was larger in the ≤ 50 years age group than in the > 50 years age group at most segments, with a statistical difference observed at the C2–3 level (p=0.009). Similarly, the mean NIPD showed the same aging-related change at each level, with a significant difference observed at most segments (p<0.05). Additionally, the mean NIPD was larger in males at all levels, with a statistical difference observed at most segments (p<0.01) (Tables 2, 3).
4. Relationship Between Intervertebral Space and Nerve Root
The anatomic relationship between the intervertebral space and the corresponding nerve root depends on the spinal level. The intervertebral space was mainly positioned anterior to the nerve roots at the C2–3, C3–4, and C6–7 segments. However, the predominant type observed is the axillary type at the C4–5, C5–6, and C7–T1 segments. The incidence of the anterior type gradually decreased from the C2–3 to the C5–6 levels, while the axillary type exhibited an opposite trend. The shoulder type was infrequent in each segment, primarily observed at the C2–3 and C3–4 levels (Table 4).
5. Descending Angles between the Nerve Rootlets and the Spinal Cord
The mean descending angles between the uppermost, middle and lowest edges of the nerve rootlets and the spinal cord increased gradually at the same segment. The mean VRA decreased as the level descended from C3–4 to C7–T1. Similarly, the mean DRA showed the same level-related change from C4–5 to C7–T1. The mean VRA and DRA were larger in females, with a significant difference observed in most segments (p<0.05). However, there was no specific aging-related change pattern observed at most segments (Tables 5, 6).
DISCUSSION
In this study, CTM imaging data from 87 patients were measured following multiplanar reconstruction and then grouped by age and sex. The highest point of the lower edge of the nerve root serves as the reference for establishing the standardized observation plane. It is an important anatomical sign to determine the boundary between the nerve root and the dural sac in PECD. This landmark is located slightly on the medial side of the intervertebral foramen entrance zone, with the uncinate joint area positioned anteriorly [10]. The nerve root compression predominantly occurs in the entrance zone, given the funnel-like shape of the intervertebral foramina, with the entrance zone representing its narrowest part [11].
In the present study, we observed that the mean NSPD was smaller than the corresponding NIPD in the lower cervical spine. This finding suggests that the axillary region of the nerve root may provide a larger operative space compared to the shoulder. However, Uğur et al. [12] and Xu et al. [13] both reported that the pedicle-inferior nerve root distance was larger than the pedicle-superior nerve root distance, which means the nerve root is closer to the inferior pedicle. Their findings seem to be contrary to ours. Indeed, the region we examined is closer to central axis of the spine. This implies that the positional relationship between the intervertebral space and the nerve roots is variable. If the herniated disc is closer to the central axis of the spine, it is more likely to be located in the axilla of the nerve root due to the smiling face-like structure (Fig. 1B) of intervertebral space caused by the uncinate joint. Conversely, if the herniated disc is closer to the intervertebral foramen, it is more likely to be positioned in front of the nerve root or in the shoulder.
Tanaka et al. [11] reported that the intervertebral disc at C4–5 to C7–T1 segment is more likely to be located on the caudal side of the corresponding nerve root as the level descends, which is similar to our results. Keorochana et al. [14] also concluded that the inferior border of cervical nerve root is located more cranial to the inferior border of intervertebral disc, ranging from C3 (0.25±1.07 mm) to C8 (2.97±2.35 mm). So, we hypothesize that performing direct neuroventral decompression via the axillary region of the nerve root is a safer and more effective approach. This approach can minimize nerve and spinal cord retraction, ultimately reducing the risk of iatrogenic injury, as the intervertebral space is predominantly situated anterior or axillary to the nerve root, and the axillary region of the nerve root potentially has a larger operative space compared to the shoulder region in the lower cervical spine. Kim et al. [15] indicated that the lateral dura edge is more laterally located than the V-point (including the inferior margin of the cephalic lamina, the medial junction of the inferior and superior facet joints, and the superior margin of the caudal lamina) at the C4–5, C5–6, and C6–7 levels. Hence, the authors recommended a more extensive bony resection, particularly when performing indirect decompression at the more caudal cervical levels or in elderly patients. Keorochana et al. [14] also recommended removing the inferior part of the upper lamina further than the superior edge of the inferior lamina as the V-point is located inferiorly to the nerve root and intervertebral disc. However, to achieve direct neuroventral decompression, we propose that a more inferior and lateral bone entry point than the V-point warrants discussion, as it may help minimize lesion to the facet joint and reduce retraction of nerve root and spinal cord. The lateral mass approach introduced by Chen et al. [16] has demonstrated significant clinical efficacy in achieving direct neuroventral decompression. Additionally, we found that the incidence of intervertebral space located anterior to the nerve roots is not uncommon. In our study, the considerable disparity between the height of the intervertebral space and the width of the nerve roots may result in an overestimation of the prevalence of the axillary and shoulder types, while simultaneously underestimating the occurrence of the anterior type. The present study also observed a larger mean NIPD in men, possibly attributed to the height difference between genders. Additionally, older individuals seem to have a smaller mean NIPD compared to their younger counterparts, possibly due to a decrease of intervertebral space height caused by age-related disc degeneration.
Studies on nerve root width have shown mixed findings. Aydoğmuş and Çavdar [17] reported that the mean width of the cervical nerves at their origin was largest at C6 and smallest at C3 and C8. Kobayashi et al. [18] showed that the width of the spinal nerve entry was largest at C5 and smallest at C8. However, our study revealed that the mean WN increased gradually from C4 to C7, with a higher measurement observed in males across all segments. The potential explanations for the varied outcomes could stem from variations in selected measurement sites, as well as potential disparities in sex, age, and racial demographics among the subjects investigated across different studies.
In this study, the mean descending angles between the uppermost, middle and lowest rootlets and the spinal cord increased gradually, indicating that the shape of nerve rootlets is similar to conical [11]. Many previous studies have reported a decreasing trend in the descending angle between the nerve rootlets and the spinal cord as the level descends [11,19-23]. Alleyne et al. [21] reported that the average cranial angle of the superior dorsal rootlets with the spinal cord was smallest at C5 and largest at T1. Furthermore, the mean cranial angle of the inferior dorsal rootlets with the spinal cord was smallest at C6, followed by C5. Shinomiya et al. [23] also indicated that the C5 ventral rootlet angle was more obtuse than those of other lower rootlets. These findings are similar to the results of our study. C5 nerve root palsy (C5P) is a relatively uncommon complication following the cervical decompression surgery that leads to a variety of debilitating symptoms, including muscle weakness, numbness, and poor quality of life [24]. Previous studies have indicated that C5 nerve root palsy is associated with various risk factors, including C4–5 intervertebral foramina stenosis, preoperative C4–5 spinal cord T2 high signal, combined with OPLL, higher preoperative cervical spine curvature, a larger superior articular process on CT, and spinal cord rotation [25-27]. Takeuchi et al. [28] suggested that the thinner nature of the C5 nerve may render it more susceptible to damage during cervical surgery. In this study, we observed that the descending angle of the C5 rootlets exceeded that of other lower rootlets, suggesting it might be one of the anatomical factors contributing to postoperative C5P.
This study has certain limitations. On one hand, the essence of this study is a two-dimensional anatomical study. Despite MPR and CTM techniques were combined, the measurement results were affected by the observation plane, measurement tools, and the subjective judgments of the researchers. Consequently, we standardized the observation plane and employed the ICC to assess consistency. On the other hand, the patients enrolled in this study presented with cervical spine-related symptoms. Most of the patients exhibited some degenerative lesions, but not severe. Therefore, future research should include a case-control study with a normal anatomical control group, stratified by disease type (the position of disc herniation, osseous foraminal stenosis) and disease severity. Besides, a multicenter study with a better design and larger sample size would provide more reliable insights into the topic.
CONCLUSION
A prediction of the positional relationship between the intervertebral space and the nerve root is essential for the direct neuroventral decompression in PECD to avoid damaging the neural structures. At the junction of the nerve root and the dura mater, the cervical intervertebral space is predominantly situated anterior or axillary to the nerve root, and the axillary region potentially has a larger operative space compared to the shoulder region in the lower cervical spine. Therefore, performing direct neuroventral decompression via the axilla approach might be safer and more effective than the shoulder approach.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Full text via PMC
Download Citation
Supplement
Print
