Introduction

Traumatic cervical spinal cord injury (TSCI) without fracture and dislocation is a special type of cervical spinal injury that usually occurs in high-energy collisions, such as car accidents and falling from a height. When subjected to traumatic hyperflexion or hyperextension movements, cervical vertebrae are often dislocated instantaneously, resulting in SCI. Such injuries do not often show any sign of fracture or dislocation of the spine on plain radiographs or computerized tomography (CT) scan, and only magnetic resonance imaging (MRI) can show the high signal of SCI [1]. In Hasler et al. study, 4489 patients suffered a SCI with or without spinal fractures/ dislocations, and 416 (9.27%) of them were diagnosed with SCI without radiographic (plain radiographs and CT scan) abnormality [2]. Patients with TSCI without fractures and dislocations usually complicated with disc rupture. Due to ruptured disc decreases the stability of the cervical spine, it will aggravate the injury to the spinal cord. Thus, surgical fixation and fusion for the injured cervical segment is necessary for spinal cord recovery.

Taylor AR et al. [3] first reported cervical disc rupture in an autopsy of cervical SCI in 1948. Macnab et al. and Davis et al. elaborated that cervical disc rupture caused by hyperextension injury was difficult to heal and required further timely surgical treatment [4, 5]. Thus, if SCI patients had intervertebral disc injuries, cervical fixation and fusion would be performed for them. Compared with other surgical procedures, anterior cervical surgery can give direct decompression of the spinal cord, and easy restore the physiological curvature of the spine, with a higher fusion rate and lower incidence of postoperative neck pain. Therefore, it was particularly recommended for patients of TSCI combined with disc rupture. Thus, a timely and accurate diagnosis of cervical disc rupture is of great importance in decision-making on fusion or non-fusion for the injury segment responsible for TSCI.

To date, there is no uniform diagnostic method for cervical disc rupture in patients with TSCI without fracture or dislocation. In previous studies, the high signal of the disc and the signal of the anterior longitudinal ligament (ALL) rupture on preoperative MRI were considered to be the signs of ruptured discs [6, 7]. However, internal rupture of a disc sometimes does not have these typical MRI features mentioned above. Usually, there are only some MRI findings such as prevertebral hematoma, high-signal SCI, and high-signal posterior ligamentous complex (PLC). It is still unclear whether these atypical MRI features have a diagnostic value for disc rupture.

In this study, we retrospectively reviewed the clinical data of patients with TSCI without radiographic abnormalities, investigated the diagnostic value of MRI features for disc rupture, and also determined which MRI feature was the best one for locating the level of a ruptured disc.

Materials and Methods

Patients

Patients with cervical SCI who underwent anterior cervical surgery in our hospital between June 2016 and December 2021 were included in this study. All patients’ clinical data, especially the intraoperative findings, were retrospectively reviewed. Written informed consent for publication of the patient’s information and images was obtained from all patients. Institutional ethics committee approval was obtained from our hospital.

The inclusion criteria were as follows: (1) patients had a history of hyperextension injury of the cervical spine within 72 h that resulted in cervical SCI and neurological dysfunction; (2) absence of signs of fractures or dislocation/instability on preoperative cervical spine X-ray and CT images; (3) patients underwent anterior cervical surgery due to pre-existing compression (such as disc herniation, vertebral osteophyte and local ossification of posterior longitudinal ligament) to the spinal cord and aggravating neurological symptoms; (4) availability of complete medical records and imaging data. Exclusion criteria were as follows: (1) associated cervical tracheal, esophageal, or arterial injury; (2) previous history of cervical spine surgery or cervical spine trauma; (3) presence of cervical vertebral abnormality, tumor, or infectious disease as a complication.

For patients with disc material and posterior osteophytes compressing the spinal cord or nerve roots, anterior cervical discectomy and fusion (ACDF) were performed. In cases where the compressive pathology is not confined to the disc space region, and extends behind the vertebral body, anterior cervical corpectomy and fusion (ACCF) was conducted for them [8].

Radiological assessment

All patients received cervical spine X-ray, CT scan, and MRI preoperatively. X-ray and CT images were used to assess cervical fracture and dislocation. MRI, especially fat-suppressed T2-weighted images (FS T2WI) in the sagittal plane, was used to observe signal change in the disc, spinal cord and soft tissue after trauma. All MRI scans were obtained on a Siemens MAGNETOM AVANTO 1.5 T MRI scanner (Siemens AG, Berlin, Germany) with 8-channel spinal phased-array coils. The scanning protocol included sagittal turbo spin-echo (TSE) T2WI (repetition time [TR] 3000–4000 ms, echo time [TE] 90–108 ms), sagittal TSE T1WI (TR 400–600 ms, TE 10–20 ms), sagittal FS T2WI (TR3500–4200ms, TE 30–50 ms), and axial TSE sequence T2WI (TR 3000–4000 ms, TE 90–110 ms) with a matrix of 320 × 256, field of 320 mm × 320 mm, thickness of layer of 3 mm, interval gap of 1 mm, and number of excitation of 2–3. The presence and anatomic level of the following MRI findings were recorded: high-signal cervical disc, ruptured ALL signal, prevertebral hematoma, high-signal SCI and high-signal PLC. The levels of the prevertebral hematoma, high-signal SCI and PLC were defined as the segment of the disc or cervical vertebral body corresponding to the maximum diameter of the prevertebral hematoma or high signal SCI/PLC. (Fig. 1).

Fig. 1: Three different MRI features related to ruptured disc in patients with traumatic cervical SCI.
figure 1

A Preoperative MRI showing prevertebral hematoma post-trauma, and C3/4 disc rupture was seen intraoperatively. B High-signal SCI on MRI and C3/4 disc rupture was identified intraoperatively. C MRI image showing only high-signal PLC preoperatively, and C5/6 and C6/7 disc rupture was found intraoperatively.

Full size image

All MRI findings were assessed retrospectively by two experienced radiologists. Also, discrepancies in the findings of the two radiologists were resolved via discussions with a third senior radiologist.

Surgical treatment

All patients underwent anterior cervical surgery due to pre-existing anterior spinal cord compression. Some had local cervical spinal stenosis, which were often combined with ossification of posterior longitudinal ligament or osteophyte formation at the posterior edge of the vertebral body, and some had cervical disc herniation. Operative approach included ACDF and ACCF. The surgical incision was located at the anterior neck slightly to the right of the midline. After the separation of subcutaneous tissues, the platysma was opened up longitudinally. The gap between the trachea, esophagus, and the sternocleidomastoid muscle was identified and the anterior cervical vertebral body and disc were exposed. C-arm fluoroscopy was performed to determine the surgical level. The anterior longitudinal ligaments and suspicious disc were explored. The presence and segment of the cervical ruptured disc were recorded during the surgery.

Disc rupture after TSCI is defined as anterior annulus fibrosus rupture, or anterior longitudinal ligament broken, or both of them presenting, intraoperatively (Fig. 2).

Fig. 2: Intraoperative image and corresponding preoperative MRI sequences of a patient with traumatic cervical SCI.
figure 2

A Both anterior annulus fibrosus rupture and anterior longitudinal ligament broken were seen intraoperatively in a patient with a disc rupture (C3/4). B The corresponding MRI showed ALL rupture, prevertebral hematoma, high-signal SCI, and high-signal PLC on FS T2WI of MRI. C The corresponding MRI on T1WI. D The corresponding MRI on T2WI.

Full size image

Data collection

The correlation between preoperative MRI features and intraoperative findings was analyzed. For the diagnostic value of the different MRI features, the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), false-negative rate (FNR), and false-positive rate (FPR) of different MRI features were calculated. Also, the combinations of these features were analyzed to determine the presence of cervical disc rupture.

The levels of the different MRI features were recorded and compared with the segments of ruptured disc based on the intraoperative findings. Then, the localization method for cervical disc rupture via preoperative MRI features was employed.

Statistical analysis

SPSS (version 21.0) was used for statistical analyses. The results were presented as patient-based and segment-based findings. Comparisons between groups of disc rupture and no rupture were performed using the chi-square test. Correlation analyses between the levels of MRI findings and the segments of cervical disc rupture were performed via the Spearman rank correlation coefficient. The threshold for statistical significance was set at α = 0.05 on the basis of two-sided tests.

Results

Patients

A total of 140 consecutive patients, 120 males and 20 females with a mean age of 53 years (age range: 27–80 years), were included in the present study. Cervical disc rupture was observed in 98 (70%) of these patients, with 135 ruptured discs in total. Among patients with disc rupture, 64 presented with a single ruptured disc, 31 with two ruptured discs, and 3 with three ruptured discs. As for the level of ruptured disc, the C5/6 segment was found to be the most common one (accounted for approximately 30% of cases). Motor vehicle accidents were found to be the major cause of injury among these patients, accounting for approximately 48% of cases (Table 1).

Table 1 Demographic data of patients with traumatic cervical spinal cord injury.
Full size table

Correlation between MRI features and intraoperative findings

According to the intraoperative findings, prevertebral hematoma, high-signal SCI, and high-signal PLC were highly correlated with the presence of cervical disc rupture. Among the 98 patients with cervical disc rupture, 26 had high-signal disc and 14 had ALL rupture signal. Of the remaining 58 patients without high-signal disc or ALL rupture signal, 54 had prevertebral hematoma, 55 presented with high-signal SCI, and 56 complicated with high-signal PLC. The differences for these MRI features between disc rupture and non-rupture groups were statistically significant (P < 0.05, Table 2).

Table 2 The differences for MRI findings between disc rupture and non-disc rupture groups.
Full size table

Diagnostic efficiency of different MRI features for disc rupture

Different MRI features have different diagnostic efficiency for traumatic disc rupture in this study (Fig. 1). All these three preoperative MRI features have high sensitivity (93%–97%) for the diagnosis of disc rupture. The diagnostic value of high-signal PLC was the highest among these three features, with a sensitivity of 97%, specificity of 72%, PPV of 84%, and NPV of 93%, also had the lowest FNR (3%) and FPR (28%) (Table 3).

Table 3 Diagnostic efficiency of the different MRI features for traumatic cervical disc rupture.
Full size table

When using the combination of two different MRI features (Fig. 3) to diagnose the disc rupture, we found that combined high-signal SCI with high-signal PLC had a higher specificity (97%) and PPV (98%), and a lower FPR (3%) and FNR (9%) for the diagnosis of cervical disc rupture (Table 3).

Fig. 3: Combination of two different MRI features for the diagnosis of disc rupture in SCI patients.
figure 3

A Prevertebral hematoma and high-signal SCI were seen on preoperative MRI, and C3/4 disc rupture was confirmed intraoperatively. B MRI showed prevertebral hematoma and high-signal PLC, and C3/4 disc rupture was confirmed intraoperatively. C Preoperative MRI demonstrated high-signal SCI and high-signal PLC, and C3/4 disc rupture was confirmed intraoperatively.

Full size image

The combination of three different MRI features had the highest accuracy in diagnosing traumatic cervical disc rupture (Fig. 4), with both specificity and PPV of 100% and zero percent misdiagnosis rate. However, if these three MRI features were used as diagnostic criteria, the missed diagnosis rate of cervical disc rupture would increase (14%) and the sensitivity would decrease (86%) correspondingly (Table 3).

Fig. 4: The combination of three different MRI features for diagnosing disc rupture in patients with traumatic SCI.
figure 4

Preoperative MRI showing prevertebral hematoma, high-signal SCI, and high-signal PLC, and C4/5 disc rupture was confirmed intraoperatively.

Full size image

Correlation between the level of MRI features and the segment of ruptured disc

The correlation between the levels of preoperative MRI features and the segment of intraoperative cervical disc rupture was analyzed in this study. We found that the segment of cervical disc rupture was strongly correlated with the level of high-signal SCI (r = 0.845, P < 0.001) and high-signal PLC (r = 0.792, P < 0.001), whereas it was weakly correlated with the presence of prevertebral hematoma (r = 0.348, P = 0.003). The prevertebral hematoma was mostly located between C3 and C4 [54 cases, (75%)]. Most of their locations were above the injured disc segment. The level of high-signal SCI usually located at the same segment of ruptured disc, with a high consistency rate of 64%. The high-signal PLC was mostly located at the adjacent level of the ruptured disc (Fig. 5).

Fig. 5: Correlation between the levels of MRI features and the segments of ruptured disc.
figure 5

Spearman rank correlation was performed between the segment of ruptured disc and prevertebral hematoma, high-signal SCI and  high-signal PLC, with Pearson coefficients of 0.348, 0.845, and 0.792, respectively.

Full size image

Discussion

It has been recognized that the soft tissues of the cervical spine, including the ALL, PLL, ligamentum flavum, interspinous ligament, supraspinous ligament, and articular processes and joint capsules, play an important role in maintaining the stability of the cervical spine, in which the stability of the cervical intervertebral disc plays an important part [9]. Cervical disc rupture is difficult to heal, which will result in cervical spine instability. What’s more, cervical disc rupture has an important impact on the choice of surgical methods. If combined with intervertebral disc injury, cervical fixation and fusion is needed to perform. If not diagnosed and properly treated, can lead to severe neurological impairment or chronic neck pain and deformity [10, 11].

Generally, MRI is considered the gold standard in the evaluation and diagnosis of soft tissue injury because of its high soft tissue contrast [12,13,14,15,16,17]. MRI has also been reported to be highly efficient in detecting injuries to the soft-tissue structures of the spine. However, for patients with TSCI without fractures and dislocations, there is still no effective diagnostic method to determine whether they have cervical disc rupture as a complication. When the cervical spine is exposed to mechanical stress, some cervical discs can directly be torn horizontally during cervical hyperextension, which is often accompanied by cervical disc bleeding, edema, and ALL/ PLL rupture. Therefore, in previous studies, typical features such as high signal of the disc and discontinuity of ALL in preoperative MRI were taken as a method for diagnosing cervical disc rupture following acute trauma [6, 7]. However, in our clinical practice, we found that many TSCI patients with disc injury were missed due to lack of typical MRI features (high-signal disc or ALL rupture signal in MRI). Thus, we determined the effect of atypical MRI features (prevertebral hematoma, high-signal SCI and PLC) on the diagnosis of cervical disc rupture in this study. Kim et al. [18] reported that intervertebral disc injury was associated with MRI imaging signals such as abnormally high signals surrounding soft tissues. Vaccaro AR et al. and Maeda T et al. [9, 19] also used the high signals of soft tissues surrounding the cervical vertebral injury to determine the presence of cervical disc injury. In our study, preoperative MRI findings of soft tissue injury, such as prevertebral hematoma, high-signal SCI, and high-signal PLC, were used as atypical MRI features to diagnose cervical disc rupture.

Previous studies have compared MRI features of traumatic cervical discs and ligamentous injury with the intraoperative findings to investigate the properties of MRI in detecting surgically verified disruptions of the ALL, PLL, and intervertebral discs [6, 20, 21]. The authors of these studies concluded that MRI was considerably accurate and sensitive in detecting cervical spine soft tissue injury and is a reliable and important tool for the evaluation of disco-ligamentous injury. However, they did not give a specific diagnostic method for disc rupture based on MRI. In this study, we compared the preoperative MRI features of the cervical spine with the intraoperative findings of disc rupture to determine the diagnostic value of atypical MRI features and found out which MRI feature was the best one for locating the level of ruptured disc.

In earlier studies, the most common causes of hyperextension cervical SCI were found to be motor vehicle accidents (40.4%) and falls (27.9%) [22, 23]. In our study, the main cause of cervical spine injury was car accidents (48%), followed by falls from high places (29%) and tumbling (13%), these results were consistent with previous studies. As for sex, we can easily find that women represent a minimal proportion of all the patients enrolled (14%). This, too, was consistent with previous findings [24]. Intraoperatively, the most frequent disc rupture levels were the C4/5 and C5/6 segments. These results were similar to those of Benjamin et al. [25]. This may be because the C4/C5 vertebral body is located at the top of the physiological lordosis of the cervical spine. So, when acute hyperextension injury occurs, the focal point of mechanical stress is generally centered on the C4/5 or C5/6 segment. The ALL and cervical intervertebral disc near the C4/C5 vertebral body suffered the greatest tension and, therefore, is most likely to cause C4/5 and C5/6 injuries [26, 27].

Cervical disc ruptures are usually caused by high-energy injuries, resulting in the rupture of small blood vessels and hemorrhage presenting as hematomas or edema between the ALL and the prevertebral fascia. As has been observed, abnormal MRI signals of cervical spine soft tissue injury are common in patients with cervical hyperextension injury. In our study, the sensitivities of prevertebral hematoma, high-signal SCI, and high-signal PLC for diagnosing disc rupture were high (93%–97%), with low missed diagnosis rates (3%–7%). Among these MRI features, high-signal PLC has the highest sensitivity, specificity, PPV, and NPV and the lowest missed diagnosis rate in the diagnosis of disc rupture. It is the best indicator of disc rupture. This may be due to the higher strength of the PLC. The strength of the PLC is higher than that of the spinal cord and prevertebral soft tissue. The presence of high-signal PLC indicates greater mechanical stress injury and a higher probability of disc injury.

Concerning the use of a combination of two different MRI features as a criterion to diagnose cervical disc rupture, the diagnostic accuracy will be higher. We found that high-signal SCI combined with high-signal PLC had a better diagnostic accuracy than those two other combinations. Therefore, we believe that the presence of high-signal SCI and high-signal PLC on preoperative MRI should be considered indicative of the presence of cervical disc rupture in patients with traumatic hyperextension injury. However, the misdiagnosis rate was higher when combined prevertebral hematoma with high-signal SCI as diagnostic criteria. This may be cervical SCI patients often comorbid with cervical intervertebral disc herniation or degeneration. Thus, slight mechanical stress will easily cause compression to the spinal cord and results in high-signal SCI, but without disc rupture [19, 28]. Additionally, concerning the use of a combination of three atypical MRI features as a diagnostic criterion, the diagnostic efficiency was the highest one, the PPV and specificity reached 100%, and the misdiagnosis rate was 0%. Therefore, if the patient has three imaging features on preoperative MRI, cervical disc rupture can be diagnosed.

For cervical SCI patients, preoperative localization of disc rupture will help surgeons to choose the right surgical method (fusion or non-fusion). Usually, the three atypical MRI features were not in the same level. So, we performed a correlation analysis between the segment of the three atypical MRI signals and the segment of the ruptured disc, in order to explore which MRI feature was the best one for locating the level of ruptured disc. For the localization of the ruptured disc segment, the level of high-signal SCI and the ruptured disc segment located at the same level had the highest consistency rate (64%). This may be due to direct herniation of the ruptured disc and the compression of the corresponding segment of the spinal cord, resulting in a hematoma or edema at the spinal cord. The maximum diameter level of a prevertebral hematoma signal was usually higher than the segment of the ruptured disc and mostly occurred in front of the C3 and C4 vertebrae [29], which means the level of the prevertebral hematoma was not an accurate indicator of the position of the ruptured disc. The level of high-signal PLC was mostly located on the adjacent segment of the disc rupture, and its accuracy in locating the ruptured cervical disc segment was less than that of high-signal SCI.

However, this study still has some limitations. First, it is a retrospective study with a limited sample size and no control group. Second, a selection bias existed in this study due to the inclusion criteria. In this study, we just included patients undergoing anterior cervical surgery. Patients who underwent posterior cervical surgery and those did not receive surgery were not included, which will affect the analysis results. A prospective study with large sample sizes should be conducted to verify the reliability of our results.

In conclusion, our results demonstrate that preoperative MRI features such as prevertebral hematoma, high-signal SCI, and high-signal PLC were of great value in the diagnosis of traumatic cervical disc rupture, especially high-signal PLC. The concomitant presence of three MRI features could confirm the diagnosis of cervical disc rupture. Additionally, high-signal SCI was the best sign to locate the segment of the ruptured disc.