Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Mar 1.
Published in final edited form as: J Orthop Res. 2020 Mar 20;39(3):671–679. doi: 10.1002/jor.24663

The Effects of Age, Pathology, and Fusion on Cervical Neural Foramen Area

Clarissa M LeVasseur 1, Samuel Pitcairn 1, Jeremy Shaw 1, William F Donaldson 1, Joon Y Lee 1, William J Anderst 1
PMCID: PMC7487021  NIHMSID: NIHMS1577767  PMID: 32167190

Abstract

Cervical radiculopathy is a relatively common neurological disorder, often resulting from mechanical compression of the nerve root within the neural foramen. Anterior cervical discectomy and fusion (ACDF) is a common treatment for radicular symptoms that do not resolve after conservative treatment. One mechanism by which ACDF is believed to resolve symptoms is by replacing degenerated disc tissue with bone graft to increase the neural foramen area, however in vivo evidence demonstrating this is lacking. The aim of this study was to evaluate the effects of age, pathology and fusion on bony neural foramen area. Participants included 30 young adult controls (< 35 years old), 23 middle-aged controls (36 to 60 years old), and 36 cervical arthrodesis patients tested before and after ACDF surgery. Participants’ cervical spines were imaged in the neutral, full flexion, and full extension positions while seated within a biplane radiography system. A validated model-based tracking technique determined 3D vertebral position and orientation and automated software identified the neural foramen area in each head position. The neural foramen area decreased throughout the entire sub-axial cervical spine with age and pathology, however, no changes in neural foramen area were observed due solely to replacing degenerated disc tissue with bone graft. Neural foramen area was not associated with disc height in young adult controls, but moderate to strong associations were observed in middle-aged controls. The results provide evidence to inform the debate regarding localized versus systemic spinal degeneration and provide novel insight into the mechanism of pain relief after ACDF.

Keywords: Neural foramen, kinematics, arthrokinematics, cervical spine, arthrodesis, ACDF

Introduction

Cervical radiculopathy affects approximately 1 in every 1,000 men and 6 in every 10,000 women each year in the US1. Radiculopathy is believed to be caused by mechanical compression of the nerve root due to a reduction in the neural foraminal opening2, which leads to paraesthesia, motor loss, and sensory deficits2. Symptoms associated with radiculopathy may be intensified with movements that compress the nerve root3, such as extension and combined extension-rotation. Radicular symptoms that do not resolve after conservative treatment are a common indication for anterior cervical discectomy and fusion (ACDF)4. One of the mechanisms by which ACDF surgery is believed to resolve symptoms is by increasing the neural foramen area at the affected site5, which is believed to occur after replacing degenerated disc tissue with bone graft. However, in vivo evidence demonstrating bone graft alone increases neural foramen area post-ACDF remains sparse.

Cross-sectional area of the neural foramen is not routinely evaluated clinically due to the limitations of standard clinical radiographs and other 2D imaging modalities. Previous studies have used 3D imaging modalities such as MRI or CT6–8 to measure neural foramen area. However, those images were acquired in a relaxed supine position, which eliminated the effects due to gravity and muscular activation. Additionally, previous research on neural foraminal area has been limited to small sample sizes (ranging from 7 to 23 subjects)6–11, a narrow healthy age range (25 to 42 years)6; 8, cadaveric specimens6, or measurements acquired in one static position6. The effects of age, pathology, cervical level, and head position on cervical neural foramen area remain unclear.

The objective of this study was to explore the relationship between age, pathology, and head position on in vivo cervical bony neural foramen openings. We hypothesized that the neural foramen opening would be larger in younger participants compared to middle-aged, and in healthy participants compared to those with symptomatic pathology. Additionally, we hypothesized that the bony neural foraminal opening would be larger in flexion compared to extension. Finally, we hypothesized that the increase of disc height associated with ACDF surgery alone would increase the neural foramen area at the operated level and have no effect on adjacent segment neural foramen area.

Methods

Level of Evidence: III, Retrospective, analytic study

Participants included 30 healthy, asymptomatic, young adults12 (15 M, 15 F; average age: 26.7±4.2 years; age range: 20 to 35 years), 23 healthy, asymptomatic middle-aged adults (8 M, 15 F; average age: 47.3±5.9 years; age range: 36 to 58 years), and 36 patients with symptomatic cervical spondylosis who received single or double-level anterior arthrodesis at the C4/C5, C5/C6 and/or C6/C7 discs (14 M, 22 F; average age: 47.8±7.2 years; age range 24 to 60 years; tested prior to surgery and 12 months after surgery). All participants provided written, informed consent to participate in these IRB-approved studies. Arthrodesis patients completed Neck Disability Index (NDI)13 surveys before and 12 months after surgery.

Participants sat upright within a custom designed biplane radiography system14 and were directed to hold still and look straight ahead. Biplane radiographs of the cervical spine were then collected for a static neutral trial. Next, participants were instructed and encouraged to flex their head and neck as far as possible while biplane radiographs were again collected for a full-flexion trial, followed by instruction and encouragement to extend their head and neck as far as possible while biplane radiographs were collected again for a full-extension trial. Radiographic imaging parameters were 70kV, 160mA, 2.5ms pulse duration, 180cm SID with 55° between cameras to obtain anterior oblique and posterior oblique views for all trials15. The maximum radiation exposure during static trials within the biplane radiography system was estimated to be less than 0.002 mSv using PCXMC software (STUK, Helsinki, Finland) (Figure 1), while the maximum radiation dose for a visit, including additional static and dynamic trials was estimated to be 1.3mSv.

Figure 1. Data Workflow.

Figure 1.

Open in a new tab

(A) Participants performed static flexion/extension while (B) synchronized biplane radiographs were collected. (C) C1 to C7 CT scans were collected and (D) used to create 3D bone models. (E) 3D vertebral kinematics were determined using a validated CT model-based tracking process. (F) 6 DOF kinematics were calculated for each position.

CT scans of the cervical spine (C1–C7) were acquired from each participant (GE Lightspeed, Waukesha, WI) with a resolution of (0.35 × 0.35 × 1.0 mm voxels). CT scans were resliced in the sagittal direction in Mimics software (Materialize NV, Leuven, Belgium) to generate 0.35 × 0.35 × 0.35mm cubic voxels. The average effective dose of the CT scans was 2.6±0.7mSv. Bone tissue was segmented from the resliced CT volume using automated thresholding and region growing algorithms in addition to manual segmentation (Simpleware, Exeter, UK). A three-dimensional (3D) model of each vertebra was generated from the segmented bone tissue16. Markers were interactively placed on the 3D bone models to define bone-specific anatomic coordinate systems as previously described17 (Figure 2). Eight additional markers (four on each side) were placed on the 3D bone models to guide the automatic identification of the neural foramen: one point at the upper edge of the lateral mass, one point on the lowest point on the superior surface of the pedicle, one point on the edge of the vertebral body, and one point on the uppermost point on the inferior surface of the pedicle3 (Figure 2).

Figure 2. Bone-specific anatomic coordinate systems and foramen identification locations.

Figure 2.

Open in a new tab

(A) Eight markers on the anterior, posterior, left and right edges of the superior and inferior vertebral body (red spheres) defined the medial-lateral (red arrow), superior-inferior (green arrow), and anterior-posterior (blue arrow) axes of the anatomic coordinate system for each vertebra. (B) Neural foramen markers (red spheres) were located at the (1) superior-medial edge, (2) middle superior pedicle surface, (3) superior-lateral edge, and (4) middle inferior pedicle surface on the left foramen. Similar points can be seen on the right foramen.

In vivo bone position and orientation was tracked using a volumetric model-based tracking technique that matches digitally reconstructed radiographs generated from the subject-specific bone models to the biplane radiographs, as previously described15 (Figure 1). This model-based tracking technique has been validated in vivo to have a precision of 0.19mm or better for tracking individual bone position15. Measurement reliability (i.e. having different operators track the bone motion) is 0.02mm in translation and 0.06° in rotation due to the computerized optimization that is used to perform the matching process15. Range of motion (ROM) was calculated as the difference in flexion angle between the static flexion and the static extension trials. To determine the disc height, the center of each endplate was calculated by first averaging the location of the four markers placed on each endplate (anterior, posterior, left and right, Figure 2A) then projecting that point onto the vertebral body surface. The distance between the centers of adjacent endplates with the head in the neutral position was then defined as the disc height.

Our previously developed algorithm for identifying neural foramen cross-sectional area3 was modified for this analysis by establishing a best-fit plane fit through the three landmarks on the superior part of the pedicle of the inferior bone and the landmark on the inferior part of the pedicle on the superior bone. Intersection of the best-fit plane with the 3D bone surface models defined the boundary of the neural foramen on each bone (Figure 3). The boundary of the neural foramen was then used to calculate the cross sectional area. This process was repeated on both sides of each motion segment from C2/C3 through C6/C7. Bony neural foramen openings were calculated during all three static trials. Absolute and right-to-left side-to-side differences (SSD), were calculated at each motion segment for each individual. Left and right neural foramen areas were averaged at each motion segment for each individual for all statistical analysis. Additionally, SSD between asymptomatic and symptomatic sides were calculated at the operated motion segment(s) for the arthrodesis patients.

Figure 3. Bony neural foramen opening identification.

Figure 3.

Open in a new tab

(A) Neural foramen markers on the superior surface of C4 and inferior surface of C3 were used to establish a (B) best-fit plane through the neural foramen. (C) The intersection of this plane with the superior vertebra (red line) and inferior vertebra (green line) was identified, then the intersection points that were closest to the medial and lateral markers were connected (black lines). (D) Anterior view of C3 and C4 showing left and right foramen openings (blue) that were then used to calculate area.

Reliability of the neural foramen area measurement was determined using nine randomly chosen participants (3 young controls, 3 middle aged controls, and 3 arthrodesis patients). We assessed the reliability of the participant positioning their head and neck, and the reliability of placing markers on the foramen to guide the foramen outline calculation. To assess reliability in participant positioning, foramen area was calculated from a second static neutral trial for all 9 participants, providing a total of 72 foramen areas from C3–C4 to C6–C7 included in the analysis. Second, to assess the effect of different users placing markers to initiate the automated algorithm for defining foramen area, a second operator placed the eight landmarks used to establish the best-fit plane, and the foramen area was calculated from those new landmarks for a total of 90 foramen areas. Intraclass correlation of neural foramen areas was used to determine the reliability of multiple operators placing landmarks along the foramen.

The between-group differences of neural foramen area across all head positions were evaluated at each intervertebral level using a Generalized Linear Mixed model (GLM). PRE to 1YR-POST changes and head position related changes in neural foramen openings at the operated and adjacent levels were evaluated for the arthrodesis group using a GLM with pre and post-surgery values and head position treated as repeated measures. Between group differences in ROM at each motion level were determined using a Student’s t-test. Pearson’s correlation was used to analyze the relationship between disc height and neural foramen area at each motion segment as well as between ROM and the change in neural foramen area from full flexion to full extension. A paired Students t-test was used to identify SSD in neural foramen area between asymptomatic and symptomatic sides at the surgical levels for the arthrodesis patients as well as right to left differences at each motion segment for each group. Significance was set at p < 0.05 for all statistical tests.

Results

Qualitative differences in neural foramen shape were observed among the young controls, middle-aged controls, and arthrodesis patients (Figure 4). Neural foramen openings in the young controls were typically muffin-shaped and consistent in terms of size and shape across motion segments within an individual as well as between corresponding motion segments across subjects. Openings in middle-aged controls were typically narrower or shorter than in the young controls and retained their muffin shape or demonstrated an hourglass shape. Openings in the arthrodesis patients were inconsistently shaped across motion segments within an individual as well as between corresponding motion segments across individuals. Additionally, the openings in the arthrodesis patients were typically much narrower and shorter compared to young and middle-aged controls. C2/C3 appeared to be the exception to these observed differences between groups, as C2/C3 opening shape and size were the most consistent across all subjects. In the arthrodesis patients, the neural foramen openings were generally similar in shape on the symptomatic and asymptomatic sides (Figure 5).

Figure 4. Shape changes in the neural foramen from flexion to neutral to extension in 3 young controls, 3 middle-aged controls, and 3 arthrodesis patients.

Figure 4.

Open in a new tab

The outline of the neural foramen is shown for 3 representative subjects from each group; blue in flexion, red in neutral, and green in extension.

Figure 5. Pre-surgery symptomatic versus asymptomatic neural foramen openings in 28 C5/C6 arthrodesis patients in the neutral head position.

Figure 5.

Open in a new tab

Solid foramen outlines represent the asymptomatic side while dotted lines represent the symptomatic side.

Neural foramen area decreased with age at all intervertebral levels except C2/C3 (all p < 0.001, Figure 6). In the neutral head position, the neural foramen openings in young controls were, on average, 9mm2, 10mm2, 15mm2, and 18mm2 larger than in the middle-aged controls at C3/C4, C4/C5, C5/C6, and C6/C7, respectively (Figure 6). In both control groups, neural foramen area decreased when moving from flexion to neutral to extension at all levels except C2/C3 (all p < 0.001, Figure 6).

Figure 6. Bony neural foramen areas for all sub-axial cervical motion segments in young and middle-aged controls.

Figure 6.

Open in a new tab

Blue bars correspond to flexion, red correspond to neutral, and green correspond to extension. The solid bars correspond to the middle-aged controls and the dashed bars correspond to the younger controls. Error bars represent ±1 SD.

Neural foramen area decreased with pathology at all intervertebral levels compared to the middle-aged controls (all p < 0.005, Figure 7). Neural foramen area in arthrodesis patients significantly decreased when moving from flexion to neutral to extension at all levels except C2/C3 and C3/C4 (all p < 0.001, Figure 7). We were unable to detect any effect of surgery on neural foramen area at either the superior or inferior adjacent levels (all p > 0.223, Figure 8). We were also unable to detect any effect of surgery on neural foramen area at the arthrodesis segment (p = 0.529, Figure 8). Although disc height at the arthrodesis segment increased from 2.8±1.1mm before surgery to 3.3±1.4mm after surgery (p = 0.027), this was offset by a change in static alignment at the arthrodesis site from 2.7±7.6° (flexion) before surgery to −0.2±7.1° (extension) after surgery (p <0.01). NDI scores improved 10±9 points, from an average of 24±9 before to 14±10 after surgery (p < 0.01).

Figure 7. Bony neural foramen areas for all sub-axial cervical motion segments in middle-aged controls and arthrodesis patients before surgery.

Figure 7.

Open in a new tab

Blue bars correspond to flexion, red correspond to neutral, and green correspond to extension. The solid bars correspond to the middle-aged controls and the dashed bars correspond to the arthrodesis patients. Error bars represent ±1 SD.

Figure 8. Neural foramen areas in flexion, neutral, and extension at the superior adjacent motion segment, the arthrodesis motion segment, and the inferior adjacent motion segment before and after surgery.

Figure 8.

Open in a new tab

Dashed bars correspond to foramen before surgery while vertically striped correspond to after surgery. Blue bars correspond to flexion, red correspond to neutral, and green correspond to extension. Error bars represent ±1 SD.

The neural foramen area in the middle-aged controls was moderately to strongly correlated to disc height at the C4/C5 (R = 0.596), C5/C6 (R = 0.793), and C6/C7 (R = 0.906) levels, while the foramen area in the younger controls was not correlated to disc height at any of those levels (all R < 0.145, Figure 9). Foraminal area was weakly to moderately correlated to disc height in the arthrodesis patients before surgery at C2/C3 (R = 0.34), C3/C4 (R = 0.40), C4/C5 (R = 0.58), C5/C6 (R = 0.55), and C6/C7 (R = 0.38).

Figure 9. Correlation between disc height and neural foramen area in neutral head position.

Figure 9.

Open in a new tab

Red triangles represent young controls, blue circles represent middle-aged controls, and each groups’ correlation between disc height and foramen area is represented with a line matching that color.

ROM was 3.2°, 2.3°, 4.5°, 4.8°, and 3.3° larger in the young controls compared to the middle-aged controls and was 2.2°, 3.4°, 3.6°, 4.5°, and 3.5° larger in the middle-aged controls compared to the arthrodesis patients at C2/C3, C3/C4, C4/C5, C5/C6, and C6/C7 respectively (all p < 0.02, Table 1). Changes in neural foramen area from flexion to extension were moderately to strongly associated with ROM in both young and middle-aged controls at C4/C5 (R = 0.52, R = 0.40, respectively), C5/C6 (R = 0.70, R = 0.44, respectively), and C6/C7 (R = 0.68, R = 0.92, respectively), while there was a weak or no association at C2/C3 or C3/C4 (all R < 0.39). Changes in neural foramen area from flexion to extension were moderately associated with ROM in the arthrodesis patients at C4/C5 (R = 0.43), C5/C6 (R = 0.68), and C6/C7 (R = 0.60), while there was a weak or no association at C2/C3 or C3/C4 (all R < 0.39).

Table 1.

ROM in young controls, middle-aged controls, and arthrodesis patients.

Motion Segment Young Controls Middle-aged Controls Pre-Surgery Arthrodesis
C2/C3 12.2 ± 2.0° 9.0 ± 2.4° 6.8 ± 2.3°
C3/C4 15.9 ± 3.6° 13.6 ± 2.5° 10.2 ± 3.5°
C4/C5 18.5 ± 3.6° 14.0 ± 3.5° 10.4 ± 3.3°
C5/C6 18.7 ± 3.7° 13.9 ± 5.0° 9.4 ± 4.3°
C6/C7 15.6 ± 4.5° 12.3 ± 4.6° 8.8 ± 4.7°
Open in a new tab

Average absolute SSD in neural foramen area ranged from 6.9 ± 5.7mm2 to 11.1 ± 8.3mm2 over all groups (Table 2). Right-to left differences averaged from −2.5 ± 8.9 mm2 to 4.8 ± 10.6 mm2. Right-to-left differences in area were significant only at the C6/C7 motion segment for the middle-aged controls (p = 0.041, Table 1). Within the arthrodesis patient group, the symptomatic neural foramen was 4.6 ± 8.9 mm2 (p < 0.01) smaller than the contralateral asymptomatic neural foramen.

Table 2. Absolute and right-to-left SSD in neural foramen area for young controls, middle-aged controls, and arthrodesis groups.

Bolded values indicate significant right-left differences

Young Controls (mm2) Middle-Aged Controls (mm2) Arthrodesis (mm2)
C2/C3 Absolute 9.8 ± 5.9 10.0 ± 9.5 8.2 ± 5.7
Right-Left −1.6 ± 9.9 0.5 ± 11.6 −0.2 ± 13.9
C3/C4 Absolute 7.9 ± 5.7 7.9 ± 5.5 7.9 ± 7.9
Right-Left 0.9 ± 9.8 0.9 ± 9.7 2.2 ± 11.0
C4/C5 Absolute 6.9 ± 6.0 8.9 ± 7.8 9.9 ± 7.1
Right-Left −2.5 ± 8.9 4.3 ± 11.2 −1.0 ± 12.3
C5/C6 Absolute 8.8 ± 6.7 8.4 ± 8.3 7.9 ± 6.3
Right-Left 3.3 ± 10.4 1.7 ± 11.8 1.6 ± 10.0
C6/C7 Absolute 11.1 ± 8.3 9.0 ± 7.1 8.5 ± 6.5
Right-Left −0.0 ± 14.0 4.8 ± 10.6 3.5 ± 10.2
Open in a new tab

Reliability in neural foramen area due to head repositioning was excellent (ICC3,1 = 0.971, average difference 0.0±3.6 mm2). Reliability in neural foramen area due to different operators placing markers was excellent (ICC3,1 = 0.973, average difference = 0.6±3.8 mm2).

Discussion

The objective of this study was to evaluate the in vivo effects of age, pathology, and surgery on neural foramen area in the cervical spine. The results indicate that sub-axial neural foraminal area decreases with age and pathology. However, there was no evidence to indicate that an increase in disc height associated with ACDF surgery affects neural foramen area at either the arthrodesis or adjacent motion segments.

Age has previously been shown to be a primary factor in morphological changes to the cervical spine18; 19. However, we believe this is the first in vivo study to establish the age-related decrease in neural foramen cross-sectional area at all sub-axial cervical motion segments. Our results can be compared to a pair of previous studies that measured neural foramen area in asymptomatic adults. Lentell et al. previously reported foramen areas in the neutral position as 48.6±12.2mm2, 47.7±10.8mm2, 46.3±9.9mm2, and 48.1±11.2mm2 in C3/C4, C4/C5, C5/C6, and C6/C7 motion segments, respectively, in 20 young controls aged 22 to 25 years8, while Mao et al. previously reported average foramen areas in the neutral position of 50.1±7.5mm2, 51.4±12.3mm2, 48.6±10.7mm2, and 53.1±7.7mm2 in C3/C4, C4/C5, C5/C6, and C6/C7 motion segments, respectively, in a group of middle-aged adults with an average age of 40±10 years11. The average neural foraminal areas we measured in the young controls at the C3/C4 and C4/C5 motion segments (45.2±10.2mm2 and 48.9±11.4mm2, respectively) were similar to those previous reports, however, the average neural foramen areas we measured at C5/C6 and C6/C7 (54.9±11.4mm2 and 61.6±16.2mm2, respectively) were larger than previously reported. Meanwhile, our middle-aged controls had smaller neural foraminal areas than previously reported: 35.8±9.3mm2, 38.7±9.2mm2, 39.7±10.3mm2, and 44.2±14.1mm2 at the C3/C4, C4/C5, C5/C6, and C6/C7 motion segments, respectively. The discrepancy between our findings and previous reports could be due to differences in average age and differences in the techniques used to measure neural foramen area. Our finding that neural foramen area decreased with age may have been anticipated given previous research that suggests cervical disc height decreases gradually with increasing age19. However, our unexpected finding that neural foramen area at the lower cervical levels appears to be associated with disc height in middle-aged, but not younger adults, suggests that differences in neural foramen area cannot always be attributed solely to differences in disc height.

Patients with single or two- level symptomatic pathology had smaller foraminal areas at all sub-axial motion segments in comparison to asymptomatic controls of similar age. This result provides evidence to support the theory of systemic, rather than localized spinal degeneration. This is highlighted by the C2/C3 neural foramen where no differences were observed between the younger and middle-aged controls, but we did observe a decrease in area in the pathological population compared to the middle-aged controls.

Contrary to our hypothesis, an increase in interbody height associated with ACDF alone did not affect foraminal area at the fused or adjacent motion segment 1 year after surgery. This finding conflicts with previously reported data from Albert et al. who found an increase in foraminal area from 37.53±12.32mm2 preoperatively to 49.04±14.27mm2 postoperatively10. However, it is important to note their post-operative measurements were taken two days after surgery in a supine CT scan. The supine CT scan would have allowed them to measure changes due to a foraminotomy during surgery, but it also prevented them from assessing the effects of graft subsidence that may have occurred in our patients over the year following surgery. Our results are more similar to Brenke et al. who measured cervical neural foramen area using supine CT before and one year after ACDF and therefore could account for changes due to foraminotomy and subsidence. They found neural foramen area decreased an average of 1.5mm2 to 3.0mm2 one year after ACDF20. Our upright flexion and extension data is in agreement with the static supine data of Brenke et al. in contradicting the commonly held belief that long-term symptom resolution is related to increased foraminal area due to replacing degenerated disc tissue with bone graft of greater height5. Rather, symptoms may resolve due to removal of soft disc, extraforaminal bony decompression, or stabilization of the motion segment as evidenced by a larger area at the arthrodesis site in extension after surgery compared to before.

In all groups, head position consistently influenced neural foramen area such that with the head in extension, the neural foramen was its smallest, and with the head in flexion the neural foramen area was its largest. Kitagawa et al. previously reported an average 12mm2 increase in neural foramen area from 44mm2 in neutral to flexion, and an average 7mm2 decrease from neutral to extension in 7 men aged 25 – 41 at C3/C4 through C6/C7 motion segments. The changes with head position we saw in young and middle-aged controls fell within these ranges. The Kitagawa et al. study, however, failed to differentiate among motion segments, and instead grouped all foramen at C3/C4, C4/C5, C5/C6 and C6/C7. As our data demonstrates, head position has a greater effect on neural foramen area in lower motion segments than on upper motion segments as evidenced by stronger associations between foramen changes from extension to flexion and ROM in all three groups. For example, in the young controls, the neural foramen area decreased an average of 30mm2 from flexion to extension at C6/C7 whereas at C3/C4 there was only an average decrease of 8mm2 from flexion to extension. This new finding is contrary to what was previously reported by Mao et al.11, who found more consistent changes across motion segments, with the largest changes between flexion and extension at C4/C5 (16mm2), followed by C3/C4 (14mm2), then C6/C7 (10mm2), and lastly C5/C6 (9mm2).

In young and middle-aged control groups, the absolute SSD in neural foramen area ranged from 6.9mm2 to 11.1mm2. Those measures provide a valuable baseline reference for average SSD that can be expected in healthy individuals. We were unable to find a difference between left and right foramen systematically, except at C6/C7 in the middle-aged controls, who were on average 9.4% larger on the right side compared to the left. This is similar to what Lentel et al. found at C6/C7 where they reported an 11.6% increase in foramen area on the right side compared to the left. Similarly, they found differences of 8.8%, 9.7%, −4.4% and −2.8% in right foramen area compared to left at C2/C3, C3/C4, C4/C5, and C5/C6, respectively, where we found −1.8%, 2.0%, −6.7%, and −6.1% differences in the young controls and 1.7%, −0.7%, 13.2%, and 4.6% differences in the middle-aged controls. Our finding that the neural foramen on the symptomatic side in the arthrodesis patients was on average 4.6±8.9mm2 smaller than the corresponding asymptomatic side at the levels to be fused suggest that symptoms may be related to decreased bony neural foramen area. However, this difference is less than the absolute SSD found in our control groups, which suggests the magnitude of the side-to-side difference may not reflect pathology.

Strengths of this study include a relatively large sample size, as well as multiple head positions and age groups analyzed. Additionally, the automated method to determine neural foraminal area provided consistent measurements within each subject among all head positions with high repeatability.

It is important to understand the study limitations when interpreting these results. One limitation is that soft disc tissue morphology was not explicitly evaluated. It is plausible that radiculopathy symptoms may have been relieved following ACDF due to removal of the disc tissue that was compressing the nerve root, however, our radiographic technique could not image disc tissue morphology. This is a limitation shared with other studies using similar techniques9; 11. Additionally, the arthrodesis patients only received a CT scan preoperatively, so any bony changes in the foramen due to resection during surgery were not evaluated.

This study provides evidence to inform the debate regarding localized versus systemic spinal degeneration and provides novel insight into the mechanism of pain relief after ACDF. The differences in neural foraminal area with age and pathology suggest the foramen are sensitive to both age-related and systemic degenerative changes. No evidence was found to support the common belief that bony foraminal area increases long-term after ACDF due to replacing degenerated disc tissue with bone graft of greater height. Further exploration of neural foraminal area changes during dynamic movements may provide additional insight into the relationship between global head motion and neural foraminal area during midrange motions that comprise the majority of activities of daily living.

Acknowledgements:

This work was supported by a research grant from Synthes Spine, NIH grants 1R01AR069543-01 and R03-AR056265, and a Cervical Spine Research Society 21st Century Development grant. The authors thank Tony Fabio, PhD for assisting with the statistical analysis.

References

  • 1.Radhakrishnan K, Litchy WJ, O’Fallon WM, et al. 1994. Epidemiology of cervical radiculopathy. A population-based study from Rochester, Minnesota, 1976 through 1990. Brain : a journal of neurology 117 (Pt 2):325–335. [DOI] [PubMed] [Google Scholar]
  • 2.Garfin SR, Rydevik B, Lind B, et al. 1995. Spinal nerve root compression. Spine 20:1810–1820. [DOI] [PubMed] [Google Scholar]
  • 3.Anderst WJ. 2012. Automated measurement of neural foramen cross-sectional area during in vivo functional movement. Computer methods in biomechanics and biomedical engineering 15:1313–1321. [DOI] [PubMed] [Google Scholar]
  • 4.Matz PG, Holly LT, Groff MW, et al. 2009. Indications for anterior cervical decompression for the treatment of cervical degenerative radiculopathy. Journal of neurosurgery Spine 11:174–182. [DOI] [PubMed] [Google Scholar]
  • 5.Bayley JC, Yoo JU, Kruger DM, et al. 1995. The role of distraction in improving the space available for the cord in cervical spondylosis. Spine 20:771–775. [DOI] [PubMed] [Google Scholar]
  • 6.Ebraheim NA, Liu J, Ramineni SK, et al. 2009. Morphological changes in the cervical intervertebral foramen dimensions with unilateral facet joint dislocation. Injury 40:1157–1160. [DOI] [PubMed] [Google Scholar]
  • 7.Kitagawa T, Fujiwara A, Kobayashi N, et al. 2004. Morphologic changes in the cervical neural foramen due to flexion and extension: in vivo imaging study. Spine 29:2821–2825. [DOI] [PubMed] [Google Scholar]
  • 8.Lentell G, Kruse M, Chock B, et al. 2002. Dimensions of the cervical neural foramina in resting and retracted positions using magnetic resonance imaging. The Journal of orthopaedic and sports physical therapy 32:380–390. [DOI] [PubMed] [Google Scholar]
  • 9.Yeni YN, Baumer T, Oravec D, et al. 2018. Dynamic foraminal dimensions during neck extension and rotation in fusion and artificial disc replacement: an observational study. The spine journal : official journal of the North American Spine Society 18:575–583. [DOI] [PubMed] [Google Scholar]
  • 10.Albert TJ, Smith MD, Bressler E, et al. 1997. An in vivo analysis of the dimensional changes of the neuroforamen after anterior cervical diskectomy and fusion: a radiologic investigation. Journal of spinal disorders 10:229–233. [PubMed] [Google Scholar]
  • 11.Mao H, Driscoll SJ, Li JS, et al. 2016. Dimensional changes of the neuroforamina in subaxial cervical spine during in vivo dynamic flexion-extension. The spine journal : official journal of the North American Spine Society 16:540–546. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Anderst WJ, Donaldson WF 3rd, Lee JY, et al. 2015. Three-dimensional intervertebral kinematics in the healthy young adult cervical spine during dynamic functional loading. Journal of biomechanics 48:1286–1293. [DOI] [PubMed] [Google Scholar]
  • 13.Vernon H, Mior S. 1991. The Neck Disability Index: a study of reliability and validity. Journal of manipulative and physiological therapeutics 14:409–415. [PubMed] [Google Scholar]
  • 14.Tashman S, Princehorn J, Pennatto S, et al. 2010. Bi-plane X-ray imaging system. United States. [Google Scholar]
  • 15.Anderst WJ, Baillargeon E, Donaldson WF 3rd, et al. 2011. Validation of a noninvasive technique to precisely measure in vivo three-dimensional cervical spine movement. Spine 36:E393–400. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Treece GM, Prager RW, Gee AH. 1999. Regularised marching tetrahedra: improved iso-surface extraction. Computers & Graphics 23:583–598. [Google Scholar]
  • 17.Anderst W, Donaldson W, Lee J, et al. 2013. Cervical disc deformation during flexion-extension in asymptomatic controls and single-level arthrodesis patients. J Orthop Res 31:1881–1889. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Kettler A, Werner K, Wilke HJ. 2007. Morphological changes of cervical facet joints in elderly individuals. European spine journal : official publication of the European Spine Society, the European Spinal Deformity Society, and the European Section of the Cervical Spine Research Society 16:987–992. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Yukawa Y, Kato F, Suda K, et al. 2012. Age-related changes in osseous anatomy, alignment, and range of motion of the cervical spine. Part I: Radiographic data from over 1,200 asymptomatic subjects. European spine journal : official publication of the European Spine Society, the European Spinal Deformity Society, and the European Section of the Cervical Spine Research Society 21:1492–1498. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Brenke C, Dostal M, Carolus A, et al. 2014. Clinical relevance of neuroforaminal patency after anterior cervical discectomy and fusion. Acta neurochirurgica 156:1197–1203. [DOI] [PubMed] [Google Scholar]

RESOURCES