Cervical transforaminal ligaments on MRI and its clinical significance

Background This study examined cervical transforaminal ligament(TFL) displays in cadavers and living bodies using magnetic resonance imaging (MRI) and evaluated the correlation between nerve entrapment in the brachial plexus by the TFL and cervical radiculopathy(CR). Methods First, 6 normal intact adult cervical specimens were used to calculate the relevant capacity in displaying the cervical TFLs by the three-dimensional fast imaging employing steady-state acquisition (3D-FIESTA). Second, 10 patients with CR and 10 healthy subjects were selected to perform the 3D-FIESTA sequence scan at the C4-T1 intervertebral foramina. The TFL display rate was calculated, and its correlation with CR was analysed. The microscopic anatomical results showed that the cervical TFL incidence was 39.6%. The relative capacity of the 3D-FIESTA sequence in displaying cervical TFLs showed a 96.6% specificity and a 73.7% sensitivity. In the 10 patients with CR, cervical TFLs were present in 17 intervertebral foramina, of which, 10 cases showed hypertrophy of the TFLs causing nerve entrapment, and corresponding symptoms of CR were found in 8 cases of cervical TFLs. The correlation between nerve root entrapment by the cervical TFL and CR showed a 96.8% specificity and an 80% sensitivity. In the 10 healthy subjects, cervical TFLs were present in 13 intervertebral foramina.


Abstract
Background This study examined cervical transforaminal ligament(TFL) displays in cadavers and living bodies using magnetic resonance imaging (MRI) and evaluated the correlation between nerve entrapment in the brachial plexus by the TFL and cervical radiculopathy(CR).

Methods
First, 6 normal intact adult cervical specimens were used to calculate the relevant capacity in displaying the cervical TFLs by the three-dimensional fast imaging employing steady-state acquisition (3D-FIESTA). Second, 10 patients with CR and 10 healthy subjects were selected to perform the 3D-FIESTA sequence scan at the C4-T1 intervertebral foramina. The TFL display rate was calculated, and its correlation with CR was analysed.

Results
The microscopic anatomical results showed that the cervical TFL incidence was 39.6%. The relative capacity of the 3D-FIESTA sequence in displaying cervical TFLs showed a 96.6% specificity and a 73.7% sensitivity. In the 10 patients with CR, cervical TFLs were present in 17 intervertebral foramina, of which, 10 cases showed hypertrophy of the TFLs causing nerve entrapment, and corresponding symptoms of CR were found in 8 cases of cervical TFLs. The correlation between nerve root entrapment by the cervical TFL and CR showed a 96.8% specificity and an 80% sensitivity. In the 10 healthy subjects, cervical TFLs were present in 13 intervertebral foramina.

Conclusions
The MR 3D-FIESTA sequence has high clinical value in displaying cervical TFLs in both cadavers and living bodies. If 3D-FIESTA sequencing shows nerve entrapment by the TFL, the possibility of CR caused by this TFL is approximately 80.0%. Conversely, the possibility of CR remains at 3.2%.

Background
Research on cervical TFLs is relatively rare. To date, few studies on cervical TFL anatomy have been reported [1][2], and imaging studies of the TFL are also rare. Many researchers believe that the presence of lumbar TFLs may cause lumbar and leg pain due to entrapped nerve roots [3,4]. Shi et al.'s latest study [2] found that cervical TFL hypertrophy likely causes entrapment of the anterior branch of the corresponding cervical nerve, resulting in corresponding clinical symptoms and causing CR. Therefore, imaging studies of cervical TFLs should provide new approaches to diagnosing CR, as well as new surgical strategies for treating it, which have important clinical application value. In current research on the aetiology of CR, the TFL may be a new factor. The 3D-FIESTA sequence clearly displays the anatomic details because it has high soft tissue resolution, with increasing applications in clinical practice due to its short scanning time [5,6]. This study applied 3D-FIESTA sequencing to scan the intervertebral foramina of cervical spine specimens and confirmed its high clinical value in displaying the cervical TFL. The 3D-FIESTA sequence was then applied to a living cervical scan, and the results showed that TFL hypertrophy can cause nerve entrapment in the brachial plexus, demonstrating that TFL hypertrophy is correlated with CR.

Cadavers:
Six adult intact cervical specimens (3 males and 3 females, aged 40-72 years, with an 4 average age of 58 years) were provided by the Department of Human Anatomy, Southern Medical University and were selected for the study. Specimens with cervical trauma, deformity, or surgery were excluded.

Living study objects:
Patients with clinically diagnosed CR were selected. Cases were excluded if they had disc herniation or unobvious protrusion that was inconsistent with the clinical symptoms of resulting CR, cervical trauma or tumours, or hyperplasia and sclerosis of facet or luschka's joint that may lead to CR. Finally, 10 cases (5 males and 5 females, aged 41-65 years, with an average age of 54.7) were included. Ten healthy subjects were also selected (5 males and 5 females, aged 46-56 years, with an average age of 51.2). All met the inclusion criteria: no clinical symptoms of CR, no traumatic history, and no history of neck surgery.

Evaluation
3.1. Evaluating the capacity of the 3D-FIESTA sequence to display the TFL using the anatomical observation results A TFL dissected from a cervical corpse was used as a control standard to assess the 3D-FIESTA sequence's capacity to display the ligament. The calculation included the 5 specificity, which meant the ability of the 3D-FIESTA sequence to determine when the specimen had no ligament, and the sensitivity, which meant the ability of the 3D-FIESTA sequence to identify when the ligament was present.

The level of nerve entrapment in the patients with clinically diagnosed CR was used
as the standard, and the nerve entrapment of the intervertebral foramen at different levels was judged via the MR 3D-FIESTA sequence, thereby evaluating the correlation between the TFL and CR.
Clinically, the level of nerve entrapment is determined based on its symptoms. A 3D-FIESTA sequence scan was performed on this object to display the TFL and its relationship with the nerve root. The calculation included the specificity, or the 3D-FIESTA sequence's ability to show no ligament if no entrapment was clinically determined on the nerve root, and sensitivity, or the 3D-FIESTA sequence's ability to display the ligament causing the nerve entrapment if nerve root entrapment was clinically determined. The calculation also included the positive predictive value, or the possibility that the 3D-FIESTA sequence would show the CR caused by the ligament entrapment when ligament entrapment was clinically determined, the negative predictive value, or the possibility that the 3D-FIESTA sequence would show no ligament entrapment when no ligament entrapment was clinically determined, and the accuracy, or the sequence's ability to determine that the CR was caused by ligament entrapment.

3D-FIESTA sequence scanning results for the cadaver specimens
The TFL signal differed from that of the nerve root, showing a low-signal linear structure ( Fig.1-a). Fifteen TFL signals were observed in 48 MR imaging findings of intervertebral foramina. The number of TFLs found in each intervertebral foramen is shown in Table 1.

Anatomical study results for the cervical cadaver specimens
6 Among the 6 cervical cadaver specimens, 19 TFLs ( Fig.1-b) were found in 48 C4-T1 intervertebral foramina (one of the intervertebral foramina had two TFLs). The TFL incidence varied greatly among the individuals. Some cervical specimens had multiple TFLs, while some specimens had none. The TFL incidences in the various C4-T1 segments are shown in Table 2.

1.3.Anatomical observations of the cervical specimens compared with the MR 3D-FIESTA sequence scanning results
The TFL results showed 14 true positives, 1 false positive, 5 false negatives, and 28 true negatives. The 3D-FIESTA sequence results for the cervical TFLs showed a 96.6% specificity and 73.7% sensitivity. Fig.1 shows an anatomical image of a TFL in a cervical specimen with its corresponding MR 3D-FIESTA sequence displaying the ligament.  Table 3.
The clinical data together with the 3D-FIESTA sequence scans resulted in 8 true positives, 2 false positives, 2 false negatives, and 61 true negatives. The healthy subjects' results included only 1 false positive and 79 true negatives. Fig.3 shows the MR 3D-FIESTA sequence of a living cervical TFL. The MR 3D-FIESTA sequence showed that the entrapment by the lateral TFL had a 96.8% specificity for the clinically diagnosed CR, with an 80% sensitivity. The positive predictive value was 80.0%, the negative predictive value was 96.8%, and the accuracy was 94.5%. The healthy subjects' specificity was 98.8%.

Discussion
Current research on TFLs focuses on the anatomy and mostly the lumbar vertebrae [7][8][9].
Studies on the cervical TFL are uncommon, with fewer imaging studies on TFLs. In this study, anatomic and imaging findings on TFLs in the cervical spine specimen were compared, and they confirmed that the MR 3D-FIESTA sequence clearly displayed the cervical TFLs. Furthermore, cervical spines from living bodies were examined, and for the first time, an imaging study was conducted of TFLs of living cervical spines. Shi et al.'s [2] study showed that the incidence of cervical TFLs at the extraforaminal space of the intervertebral foramen is higher in the lower cervical vertebrae. The lower cervical spine is the location issuing the brachial plexus and is closely related to CR. Therefore, the intervertebral foramina from C4 to T1 were selected for this investigation.
In this study, the average TFL display rate was approximately 35.0% on the cervical cadaver specimens and 22.8% on the living bodies. The TFL display rate in the imaging study with the living cervical spines was lower than that of the cadaver specimens, which may be related to the images' signal-to-noise ratio. Cervical cadaver specimen MR images 8 contain no motion artefacts, while those of living cervical spines include unavoidable involuntary movements such as swallowing and blood vessel pulsation. This results in a reduced signal-to-noise ratio for the live image, which impacts the radiologists' analytical power in assessing the TFLs. However, we believe that with improved imaging equipment and technology, these motion artefacts can be reduced, and their impact can be minimized in future TFL analyses. In addition, cervical TFLs in different individuals differ greatly. Shi et al.'s study [2] showed that up to 7 cervical TFLs were found in some individuals, while no cervical TFLs were found in others. Therefore, the TFL display rates for the living bodies and cadaver specimens in our MR study may differ. At present, relatively few scholars are investigating cervical TFLs, with few cases in these studies. To more accurately understand the incidence of TFLs, the study's sample size must be increased, and multicentre research should be conducted in the future.
In clinically determining CR [10][11], Wainner et al. [11] found that more than 75% of patients were diagnosed with CR based on their clinical history. In this study, the nerve entrapment position was determined by the chief physician in neurology, who had been involved in clinically diagnosing cervical diseases for more than 20 years. When clinically diagnosing CR, subjective opinions may exist in individual cases, and objective and accurate indicators are lacking; thus, the results may be biased. Therefore, we conducted myoelectric evoked potential examinations in some patients [12] to further improve the accuracy in clinically diagnosing CR.
Our study showed that the correlation between nerve entrapment caused by the cervical TFL and CR had a 96.8% specificity, which is relatively high. This may be related to the absence of TFLs in most intervertebral foramina. The TFL display rate on the cervical spine in the 3D-FIESTA sequence was only 22.8%. Of the 22.8% TFL displays, only 58.8% showed entrapment, indicating entrapment in only 13.4% of the TFLs overall. Thus, only 13.4% of 9 the TFLs in the foramina showed entrapment. In other words, when the nerve in the intervertebral foramen is clinically determined as not being entrapped, the radiologist can confirm that no TFL exists in this foramen, or the presence of a TFL did not cause nerve entrapment. The 80.0% sensitivity of the living cervical TFLs suggested a 20.0% possibility that the 3D-FIESTA sequence did not reveal the CR caused by nerve entrapment due to the TFL. This might suggest that the accuracy in clinically determining CR in this study was not 100%, and some cases were misdiagnosed. In the future, we will increase the number of examinations for clinically diagnosing CR by adding additional objective data, such as electromyography (EMG) examinations, to improve the clinical diagnostic capability, thereby better detecting the 3D-FIESTA sequence's display capacity in determining nerve entrapment caused by TFLs.
In the healthy subjects, the average display rate of the TFL by 3D-FIESTA sequencing was approximately 16.3%, while the average incidence of the TFL in the patients with CR was approximately 21.3%. This indicated a large difference in those with TFLs and indirectly reflected that the incidence of TFLs in patients with CR is higher than that in people without CR and the higher this incidence, the higher the probability of nerve entrapment.
We conducted a comparative study of cervical spine specimens and MR 3D-FIESTA sequence scans to confirm that 3D-FIESTA sequencing clearly displays the cervical TFLs.
3D-FIESTA sequencing was applied to studying living cervical spines, confirming that nerve entrapment by the TFL is correlated with CR. Therefore, surgical treatment, such as releasing or removing the ligament, may be an option for treating CR in the future, which provides a good idea for clinical surgery. In the future, it may be possible to reselect the surgical plan.

Conclusion
MR 3D-FIESTA sequencing clearly displays the TFL and its relationship with the nerve root, both in cadaver cervical specimens and in living cervical vertebrae. If the 3D-FIESTA sequence shows that the nerve is compressed by a TFL, the possibility of CR caused by this TFL is approximately 80.0%. Conversely, the possibility of CR remains at 3.2%. MR 3D-FIESTA sequencing can confirm whether entrapment of the nerve root by the TFL is causing the clinical symptoms of CR. It has a high clinical application value and provides a new idea and strategy for diagnosing and treating CR in the future.
Abbreviations CR cervical radiculopathy MRI magnetic resonance imaging TFL transforaminal ligament 3D-FIESTA three-dimensional fast imaging employing steady-state acquisition

Notes And Declarations
Notes Junlin Li and Lina Wang contributed equally to this work.

Declarations
Ethics approval and consent to participate Current study was approved by the Inner Mongolia Autonomous Region People's Hospital and need for signed informed consent was waived.

Availability of data and material
The datasets used and/or analyzed during the current study available from the corresponding author on reasonable request.

Competing Interests
The authors have no conflicts of interest relevant to this work.

Funding
This study was was supported by the Natural Science Foundation of Inner Mongolia No.
2017MS(LH)0850; In the role of collection data.

Authors' contributions
JLL and LNW participated in the design of the study and drafted the manuscript, XHOY collected the patients' data, and processed the figures. XQZ conceived the study and supervised the project. All authors read and approved the final version of the manuscript.