Prognostic MRI parameters in acute traumatic cervical spinal cord injury

The aim of this study is to estimate the prognostic value of some features documented on preoperative MRI study in patients with acute cervical spinal cord injury. The study was conducted in patients operated for cervical spinal cord injury (cSCI) from April 2014 to October 2020. The quantitative analysis on preoperative MRI scans included: length of the spinal cord intramedullary lesion (IMLL the canal diameter at the level of maximal spinal cord compression (MSCC) and the presence of intramedullary hemorrhage. The canal diameter at the MSCC was measured on the middle sagittal FSE-T2W images at the maximum level of injury. The America Spinal Injury Association (ASIA) motor score was used for neurological assessment at hospital admission. At 12-month follow-up all patients were examined with the SCIM questionnaire. At linear regression analysis, the length of the spinal cord lesion [β coefficient −10.35, 95% confidence interval (CI)−13.71 to−6.99; p < 0.001], the diameter of the canal at the level of the MSCC (β coefficient 6.99, 95% CI 0.65 to 13.33; p = 0.032), and the intramedullary hemorrhage (β coefficient  − 20.76, 95% CI − 38.70 to − 2.82; p = 0.025), were significantly associated with the score at the SCIM questionnaire at one year follow-up: shorter spinal cord lesion, greater diameter of the canal at the level of the MSCC, and absence of intramedullary hemorrhage were predictors of better outcome. According to the findings of our study, the spinal length lesion, canal diameter at the level of spinal cord compression and intramedullary hematoma documented by the preoperative MRI study were associated with the prognosis of patients with cSCI.


Introduction
Cervical spinal cord injury (cSCI) is a devastating condition and a common cause of disability and death in young people [1,2]. Epidemiological studies documented that cSCI has an estimated incidence of 750 cases per million annually with a great impact upon families and society [3].
Fortunately, in recent years the improvement in the surgical, anesthesiologic and neuroradiological procedures allowed to obtain good results in terms of neurological recovery with reduction of permanent disability [4]. Anyway, the preoperative prognosis is uncertain because generally based on the neurological preoperative status, sometimes rapidly changing especially in the early hours after cSCI due to the spinal shock [5].
In the context of the cSCI, spinal MRI is the gold standard to assess any damage to the discal-ligaments complex, and to the spinal cord. By applying the T1 and T2-weighted sequences, these images reveal the extent of myelopathy, presence of edema or hematoma, degree of spinal cord compression, and level of ligamentous instability. All these macroscopic elements, together with the clinical status evaluation, constitute crucial information for the surgeon in a urgence decision-making context, inspiring a surgical or conservative treatment [6][7][8].
More recently, the literature has focused on the use of advanced quantitative MRI techniques (DTI, relaxation mapping, magnetisation transfer) to investigate microstructural and functional changes in acute cSCI and their predictive role in functional recovery [9][10][11]. Nevertheless, conventional MRI still provide insights concerning the primary pathological features of the cord injury, while quantitative MRI studies reveal information on the secondary pathological changes affecting the entire neuraxis, thus most applicable in a post-surgical follow-up and neuro-rehabilitation context [6].
The aim of our study is to investigate the association between qualitative MRI features in patients with cSCI and long-term neurological and functional recovery, because we believe that this imaging information could represent a useful and prompt tool to the neurosurgeon in the surgical decision moment.

Materials and methods
The study included a consecutive series of patients with cervical spine trauma who were admitted to the Department of Neurosurgery from April 2014 to October 2020.
We applied the following inclusion criteria: surgically treated patients with a cervical spinal trauma, presenting vertebral fracture, bone fragment in the spinal canal, clinically in ASIA A-B-C-D score. The exclusion criteria were: spondylotic cervical myelopathy, history of infectious or tumor disease, precedent cervical trauma, vascular-based myelopathy, alcohol/drug abuse, ASIA E score. The final sample included 39 patients (29 men, 10 female; mean age 58.2 years; age range 20-86 years) with complete clinicalinstrumental documentation and reliable follow-up data.
At admission in the Emergency department, all patients were administered NASCIS III protocol. No cervical traction was used before surgery.
The America Spinal Injury Association (ASIA) motor score was utilized for the neurological assessment at hospital admission and follow-up. At 12 months follow-up all patients were examined with the SCIM questionnaire ( Fig. 1); higher scores at the SCIM questionnaire correspond to better outcome. All patients were evaluated with Computed Tomography (CT-scan) and MRI study pre and postoperatively at 6 and 12 months and the vertebral fractures classified according to the AO Spine Classification [14].
According to the AOSpine guidelines and surgeon's experience, the surgical approach was selected with the aim to obtain adequate decompression of the neural structures and restore the cervical spine alignment. Three surgical approaches were performed: anterior cervical discectomy or corpectomy and fusion with interbody cage and anterior plate in case of anterior column impairment, posterior cervical screw fixation with lateral trans-mass screws and laminectomy in patients with involvement of posterior elements, a combination of anterior and posterior approach when both elements were injured (Table 1).
After surgery, all patients wore a rigid cervical collar for at least 1 month. All patients with neurological damage followed a specific rehabilitation process, to promote the neurological recovery. All MRI scans were performed with 1, 5 Tesla magnet (Philips Achieva 1, 5 T Philips Healthcare).
Standardized MRI protocols were used for the acute cervical spine injury with axial and sagittal T1, T2 (FSE) and STIR (inversion recovery) plus axial T1, T2 (gradient Echo) sequence. Quantitative analysis on MRI scans included the length of the spinal cord intramedullary lesion (IMLL) and the canal diameter at the level of the maximal (spine) cord compression (MSCC). The length of the spinal cord lesion was defined as the distance between the most cranial and the most caudal point of the spinal cord signal intensity change on middle sagittal T2-weighted images. The canal diameter at the level of the MSCC was measured on the middle sagittal FSE-T2W images at the maximum level of injury. The presence/absence of intramedullary hematoma was also assessed in the mid-sagittal plane on MRI T1-weighted 'Fast Spin Echo FSE' (Fig. 2). Images Software used: PACS WEB (Fig. 3).

Statistical analysis
Values are presented as mean ± standard deviation (SD) or median (interquartile range [IQR]) for continuous variables and as number (percent) of subjects for categorical variables.
The length of spinal cord edema and the canal diameter at the level of the MSCC were considered as continuous variables; the intramedullary hemorrhage was analyzed as a dichotomous variable (presence versus absence). The association between the score obtained at the SCIM questionnaire at 12 months and baseline characteristics was estimated using a linear regression analysis. Age, length of spinal cord edema, the canal diameter at the level of the MSCC and the intramedullary hemorrhage were included in the multivariate model.
Results were considered significant for p values < 0.05 (two sided). Data analysis was performed using STATA/IC 13.1 statistical package (StataCorp LP, Texas, USA).

Results
The most common mechanism of injury in all 39 patients was motor vehicle accident. The most common level involved was C5-C6. At admission, a complete neurological deficit (ASIA A) was documented in 10 patients out of 39 (33%). An incomplete deficit was evaluated in most of the patients (77%), of which 7 with ASIA score B, 4 with ASIA score C and 18 with ASIA score D. All patients underwent surgery within 24 h from the spinal trauma: 17 patients (43.6%) were treated with anterior decompression and fusion with cage and plate, 19 (48.7%) with posterior screw fixation and laminectomy and 3 patients (7.7%) with a combination of both (Table 1) Table 1.
At linear regression analysis, the length of the spinal cord lesion [β coefficient − 10.35, 95% confidence interval (CI) − 13.71 to − 6.99; p < 0.001], the diameter of the canal at the level of the MSCC (β coefficient 6.99, 95% CI 0.65 to 13.33; p = 0.032), and the intramedullary hemorrhage (β coefficient − 20.76, 95% CI − 38.70 to − 2.82; p = 0.025), were significantly associated with the score at the SCIM questionnaire at one year follow-up: shorter spinal cord lesion, greater diameter of the canal at the level of the MSCC, and absence of intramedullary hemorrhage were predictors of better outcome ( Table 2).

Discussion
In the acute setting of patients with cervical spinal cord trauma, MRI study is important to define the extent of soft tissue and bone injuries and associated ligament and disc injuries, as well as to define the extent of intramedullary hematomas and cord compression. Preoperative MRI study in cervical cSCI is the gold standard and some studies pointed out that it is useful also for the decision making in terms of surgery/no surgery, type of approach, instrumentation need, levels of decompression and reoperation decision when the decompression is insufficient [12].
Recently, some parameters on conventional MRI study have been described such as intramedullary lesion length (IMLL) and maximal spinal canal compression (MSCC). The author's hypothesis is that these parameters could be used for prognostic purpose in addition to the preoperative ASIA score evaluation [13]. In parallel, the scientific community is increasingly turning its attention to more sophisticated imaging techniques for studying spinal cord injury and its clinical implications. While conventional MRI images show relevance in decision-making process and clinical management, functional and quantitative MRI, detecting the presence of intramedullary biomarkers (iron deposit, demyelination, etc.) provide more information about the evolving neurodegenerative and reorganizational changes during recovery from the primary event [6,9]. These advanced imaging techniques appear to be more useful during the rehabilitative phase, potentially revealing the interplay between the microstructural cerebral and spinal neuro-plasticity and the clinical evolution, even if not always a standardized and widespread procedure [14].
In this study we evaluated some basilar MRI parameters for prognostic purpose. In this series, 10 patients out of 39, after early decompression surgery, at 12 months follow-up had a SCIM score below 20, which in clinical terms implies  that they are quadriplegic and totally dependent on third parties for self-care. These patients are those who had a degree of IMLL greater than 7.36 mm at admission, (p value = 0.001, see Table 2). Raj Kamal et al. analyzed the role of IMLL and outcome in cSCI patients and, in line with our results, he concluded that the preoperative IMLL was a significant predictor of long-term outcome [2].
Ahamad M. et al. obtained the same result pointing out the important role of MRI in the preoperative setting for diagnosis and prognosis [15].
All our patients were operated within 24 h, according to the wide-known guidelines of spinal cord trauma, where the delay of the surgical intervention is recognized to play a crucial role in the following neurological recovery [9,[16][17][18].
The maximal spinal cord compression (MSCC) is frequent in cervical cSCI and some studies reported a presence on preoperative MRI about 89% with T1w sagittal sequence and 92% with T2w sagittal sequence in patients with neurological impairment [12]. These data confirm the ability to use this parameter in all patients with cSCI and to verify the possible correlation with the long-term outcome.
The maximal spinal cord compression (MSCC) was studied extensively from many authors and in all cases, a close correlation was demonstrated with the neurological outcome [19][20][21]. In our study we documented that maximal spinal cord compression (MSCC) varies from 4,18 cm to 9,84 cm and a relatively larger canal diameter was associated with better neurological recovery with statistical significance (p = 0.032). This may be due to the protective role of the perimedullary space during the normal flexion/extension of the cervical spine but, in case of cervical trauma, the movement is exaggerated and uncontrolled with compression and contusion of the spinal cord within the vertebral canal, especially when the space is restricted.
As documented by many Authors, the presence of intramedullary hemorrhage correlates with a worse prognosis [22,23]. Hemorrhage indicates spinal cord damage with poor outcome and neurological recovery. In our study, 9 patients out of 39 presented, at the preoperative MRI study, an intramedullary hemorrhage and the SCIM score was below 32 at 1-year follow-up, significantly lower than the SCIM score of patients who did not, with statistically significance (p value = 0.025).
Only one patient had the infection of the wound but the treatment with antibiotics and negative wound pressure allowed a complete recovery [24].
In our series, the age of patients was not significative for the outcome, even if other Authors confirmed the correlation between age and the neurological recovery [21]. In our study, this may be due to the small numerosity of the sample.
The results of the statistical study show how the extension of intramedullary cranial-caudal lesion, the degree of spinal cord compression and the presence of intramedullary hematoma have a significant influence on the final 1-year neurological status of patients, as stated in literature [11,21].
Our study documents the important role of the cervical MRI study in patients with cervical spinal cord trauma, which is confirmed to be the gold standard for diagnosis, decision making process and prognosis. Moreover, the MRI study may be of value in cases with clinical and radiological mismatch [1]. In our experience patients with longer lesion length, relevant spinal cord compression and intramedullary hematoma documented by MRI study, could be surgically treated as soon as possible to achieve the best possible recovery [25].

Conclusions
While technology improvement furnishes more accurate and detailed elements about the cellular and biomolecular damage in cSCI, conventional MRI study represent an immediate available and reliable radiological mean, essential in an emergency context.
According to the results of this study, the spinal length lesion, the degree of spinal cord compression and the intramedullary hematoma documented by the MRI study in cSCI patients have a direct correlation with the prognosis and play a crucial role in the pre-surgical decision-making process. permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.