Evaluation of Traumatic Spine by Magnetic Resonance Imaging and Its Correlation with Cliniconeurological Outcome.

BACKGROUND: Spinal trauma is associated with long-term disability. Early detection can lead to prompt and accurate diagnosis, expeditious management, and avoidance of unnecessary procedures. Magnetic resonance imaging (MRI) helps to accurately depict the presence and extent of spinal cord injury (SCI) in these patients.
PURPOSE: This study was aimed to look for various MRI findings which are predictive of initial neurological deficit in patients with spinal trauma and to correlate the findings with resultant neurological outcome.
MATERIALS AND METHODS: The present study was conducted over a period of 18 months from January 2016 to June 2017 in 57 patients with spinal trauma who underwent MRI spine. Neurological status of patients was assessed at the time of admission and discharge in accordance with the American Spine Injury Association (ASIA) impairment scale. Various MRI parameters were evaluated for correlation with the severity of the spinal injury.
RESULTS: Patients with cord transection, cord hemorrhage, and epidural hematoma had initial high-grade ASIA impairment scale. Patients with cord transection and cord hemorrhage did not show any improvement in their neurological status during their hospital stay. Patients with only cord edema and epidural hematoma showed favorable neurological outcome. Cord contusion showed lesser neurological recovery, as compared with cord edema and normal cord.
CONCLUSION: MRI findings in acute SCI correlated well with the initial neurological deficits on admission and at the time of discharge. MRI should be recommended in all patients with suspected spinal trauma both as a diagnostic and prognostic indicator.

INTRODUCTION

Acute traumatic spinal cord injury (SCI) is a common cause resulting in debilitating injuries. SCI has high prevalence in the younger population. Diagnostic imaging, particularly magnetic resonance imaging (MRI), plays a crucial role in evaluating and detecting spinal trauma. Bone marrow, soft-tissue, and spinal cord abnormalities, which may not be apparent on other imaging modalities, can be readily detected on MRI. Early detection often leads to a prompt and accurate diagnosis, expeditious management, and avoidance of unnecessary procedures.[1]

MRI has high-contrast resolution, multiplanar capabilities, and with various pulse sequences makes it possible to diagnose spinal trauma more accurately. Adequate information about neural and extraneural injuries requiring surgical interventions, for example, significant disc herniations and epidural hematomas can be obtained. In cases of spinal cord edema, contusion, hemorrhage, and ischemia, MRI findings may serve as prognostic indicators.[2]

Most of the diagnostic information in spinal trauma is derived from the sagittal images. Sagittal T1-weighted images offer an excellent anatomic overview. Axial images serve as a supplement. Disc herniations, epidural fluid collections, subluxation, vertebral body fractures, cord swelling, and cord compression are also visualized. Sagittal T2-weighted images depict most of the abnormalities, including spinal cord edema, ligamentous injury, disc herniations, and epidural fluid collections. Axial and sagittal gradient echo (GRE) images to aid in the identification of acute spinal cord hemorrhage.

The depiction of SCI on MRI not only correlates well with the degree of neurologic deficit but it also bears significant implications in regard to prognosis and potential for neurological recovery.[3]

As MRI is an excellent diagnostic modality for the evaluation of spinal trauma, it is possible that the MRI findings correlate directly with the degree of neurological deficit according to the American Spinal Injury Association (ASIA) impairment scale. The purpose of this study was to evaluate this correlation.

Aims and objectives

The aims and objectives of the study were to document morphological changes in the spinal cord, vertebra and adjacent soft tissues in patients with trauma and to correlate the level and severity of injury on MRI with resultant neurological outcome.

MATERIALS AND METHODS

Source of data

This prospective observational study was conducted over a period of 18 months from January 2016 to June 2017 on 57 patients with spinal trauma who underwent MRI of the spine. Prior written informed consent was obtained. The patients were included if they fulfilled the inclusion and exclusion criteria. Inclusion criteria: All the patients of acute spinal trauma undergoing MRI formed the study group. Exclusion criteria: Patients with associated head injury or patients in whom MRI is contraindicated (including but not limited to noncooperation, in situ metallic implants, cochlear implants, pacemakers, previous spine surgeries, and claustrophobia).

Method of data collection

MRI of the spine was performed with the patient in the supine position using 1.5 Tesla MR Scanner (Magnetom Avanto, Siemens) both in the axial and sagittal planes using a combination of pulse sequences. Sagittal and axial T1- and T2-weighted c, coronal short tau inversion recovery and GRE sequences were performed for the evaluation of vertebra, disc, spinal cord, and soft tissue. Sagittal images were 5.0-mm thick with a 0.5-mm slice gap. The field of view (FOV) of the area of interest was adequate at 24 cm in the cervical spine and at 32 cm in the lumbosacral spine. In the dorsolumbar spine, a large FOV was used (34/36 cm) for accurate labeling of the involved levels.

T2-weighted information was obtained using a single fast spine echo (FSE) acquisition using a split echo train, resulting in an intermediate T2 WI sequences. For the short TE image, an echo train of three with two excitations was used, whereas for the long TE image an echo train of 15–30 with single excitation was used. For each sequence, 256–448 steps were followed in both the frequency and phase axes. Fat suppression was employed on the long TR sequences to improve visualization of edema in the posterior ligamentous complexes. Axial images were obtained using FSE or GRE pulse sequences. Technical parameters included 16° flip angle, minimum Repetition time (TR)/Echo time (TE), 224 × 320 matrix and two excitations in T1 WI and one excitation in T2 WI. The TE used was <15 ms in T1 WI and up to 100 ms in T2 WI to minimize unwanted susceptibility artifacts that might exaggerate bony stenosis.

The following findings were identified after assessing the MR images: Cord transection, cord hemorrhage, cord edema, epidural hemorrhage, and normal cord.

Clinical assessment of spinal cord injury

Patient assessment was performed using standardized physical examination at the time of admission and at the time of discharge according to the International Standards for Neurological and Functional Classification of SCI Patients, also commonly called the ASIA guidelines. A detailed motor and sensory examination of the patient was performed and graded according to the ASIA scale which was as follows:

Complete: No motor or sensory function is preserved in the sacral segments S4–S5

Incomplete: Sensory but not motor function preserved below the neurologic level and includes the sacral segments S4–S5

Motor function is preserved below the neurologic level, and more than half of the key muscles below the neurologic level have a muscle grade <3

Incomplete: Motor function is preserved below the neurologic level, and at least half of key muscles below the neurologic level have a muscle grade of 3 or more

Normal: Motor and sensory function are normal.

Change in ASIA impairment scale (ASI) toward lower grade between admission and discharge was considered neurological recovery, whereas no improvement or worsening was considered as no recovery.

Data analysis

The strength of association between extent of SCI and outcome were described using the Odds ratio. Chi-square test of significance was used to assess the association between MR findings and clinical outcome.

RESULTS

In this study, 57 patients of spinal trauma were observed. Majority of the patients were males (n = 49; 86%). Most common age group in our study was 21–40 years (n = 26; 45.6%), followed by 41–60 years (n = 19; 33.3%). Patients with an age group of 20 years and below were 12.3% (n = 7), and 8.8% of patients (n = 5) were in the age group of 61 years and above.

In 32 patients (56.14%), the cause of injury was fall from height followed by road traffic accident (RTA) (n = 21; 36.84%) and fall of weight (n = 4; 7.02%).

Level of injury

The most common level of injury was cervical level. Of 57 patients, 25 patients (43.86%) had injury at cervical level followed by dorsal and lumbar levels (n = 12 each; 21.05%) and finally, eight patients (14.04%) had dorsolumbar injury.

Magnetic resonance imaging findings in spinal trauma cases

Out of 57 patients, 44 patients (77.19%) had cord abnormalities, whereas rest (n = 13; 22.80%) had no cord changes (normal cord).

The various MRI findings were cord edema (n = 23), cord contusion/edema (n = 10) [Figures 1 and 2], spinal canal stenosis due to retropulsion (n = 7) [Figure 3], cord transection with contusion/edema (n = 4) [Figures 1 and 2], cord hemorrhage [Figure 1], epidural edema (n = 3 each) [Figure 4], and normal findings in 13 patients [Table 1]. Out of 57 patients, 12 patients (21.05%) had cord compression. It was observed that patients with normal cord were associated with significantly better neurological status at admission (P< 0.001).

Figure 1

Sagittal T1-weighted (a), T2-weighted (b), and gradient echo (c) images showing comminuted fracture of D12 vertebral body with posterior translation. There is complete transection of distal thoracic cord at this level with blooming on gradient echo sequence, representing cord hemorrhage

Figure 2

Sagittal T1-weighted (a), T2-weighted (b) and gradient echo (c) and Coronal Short tau inversion recovery (d) images of cervical spine. There is linear hypointense signal with blooming on gradient echo sequence involving near complete thickness of cord at C5 mid vertebral body level, likely partial transection (c). Note the central focal T2 hypointense cord signal at this level likely hemorrhagic contusion. There is diffuse surrounding T2 hyperintense signal from C3 vertebral body level to C6-7 disc level likely cord edema (b). Short tau inversion recovery image (d) showing transverse hypointense signal involving near complete thickness of cord at C5 mid vertebral body level likely partial transection

Figure 3

Sagittal T2-weighted (a) and short tau inversion recovery (b) images showing anterior wedge compression fracture of D12 vertebra with mild retropulsed fracture fragment. T2 hyperintensity seen in the cord likely cord contusion versus edema at D11–D12 level

Figure 4

Sagittal T1 weighted (a), T2 weighted (b) and axial T2 weighted (c) images showing anterior wedge compression fracture of D12 vertebral body with mild retropulsion causing cord compression and edema at D11–D12 level. There is T1 isointense and T2 hypointense posterior epidural hematoma at D11–D12 disc level (white arrow)

Table 1

Magnetic resonance imaging findings

MRI findings	Number of patients (%)
Cord transection with contusion/edema	4 (7.0)
Cord hemorrhage	3 (5.3)
Cord contusion/edema	10 (17.5)
Cord edema	23 (40.4)
Spinal canal stenosis due to retro pulsed fracture fragment	7 (12.3)
Normal cord	13 (22.8)
Epidural hematoma	3 (5.3)

MRI: Magnetic resonance imaging

Categorization of patients based on the American spinal injury association impairment scale

In this study, spinal trauma patients were grouped into five categories based on ASIA impairments scale (AIS) at the time of admission.

The most common presentation was ASIA A in 22 patients (39%) followed by ASIA C in 13 patients (23%), ASIA D in 12 patients (21%), ASIA E in 8 patients (14%), and finally ASIA B in two patients (3.5%).

Magnetic resonance imaging findings and neurological status at the time of admission

In all patients with cord transection (n = 4), cord hemorrhage and epidural hematoma (n = 3 each) the initial neurological status was ASIA A. There were 23 patients with cord edema among whom 10 patients had initial neurological status of ASIA A, seven had ASIA C, four had ASIA D and remaining two had ASIA B [Table 2].

Table 2

Magnetic resonance imaging findings and neurological status at the time of admission

MRI findings	ASIA impairment scale					Total
	A	B	C	D	E
Cord transection with contusion/edema	4	0	0	0	0	4
Cord hemorrhage	3	0	0	0	0	3
Cord contusion/edema	8	0	2	0	0	10
Cord edema	10	2	7	4	0	23
Spinal canal stenosis	0	0	3	3	1	7
Normal cord	0	0	1	5	7	13
Epidural hematoma	3	0	0	0	0	3
Total	28	2	13	12	8	63*

*Few patients had multiple findings. ASIA: American Spine Injury Association, MRI: Magnetic resonance imaging

Among 10 patients with cord contusion/edema, eight patients had initial neurological status of ASIA A and rest of two had ASIA C. Spinal canal stenosis was noted in seven patients, among whom three patients had the initial neurological status of ASIA C, three patients had ASIA D and one patient had ASIA E [Table 2].

Out of 57 patients, 30 patients (53%) were managed by surgically and 27 patients (47%) were managed conservatively [Table 2].

Neurological recovery in patients with different magnetic resonance imaging findings

Four patients with cord transection and three patients with cord hemorrhage did not show any improvement in their neurological status over the period of their hospital stay. Out of 10 patients with cord contusion/edema, the neurological improvement was seen in five patients (50%) and remaining five patients (50%) showed no improvement. Out of 23 patients with only cord edema, 13 patients (56.52%) showed improvement and 10 patients (43.47%) showed no improvement. Out of three patients with epidural hematoma, two patients showed improvement, whereas one patient did not improve. Finally, of 13 patients with normal cord, six patients had neurological deficit, all of them improved. In our study, seven patients had spinal canal stenosis out of which six patients had neurological deficit, of them four patients showed improvement and two patients did not improve.

When MRI findings were compared with the neurological outcome, it was observed that there was no neurological improvement in patients with cord transection with contusion/edema and cord hemorrhage. Furthermore, these patients had a significantly worse outcome when compared with other MRI findings (P < 0.05). There was no significant difference in outcome among patients with cord contusion/edema compared with spinal canal stenosis (P = 0.34) and cord contusion/edema compared with cord edema only (P = 0.37). There was statistically no significant difference in terms of neurological improvement among patients with cord contusion/edema and epidural hemorrhage (P = 0.34). Similarly, when patients with cord edema were compared with patients with epidural hemorrhage, there was no significant difference in terms of neurological improvement (P = 0.39). There was no significant difference in neurological outcome among patients with spinal canal stenosis when compared with patients with cord contusion/edema (P = 0.28), cord edema (P = 0.34), and epidural hematoma (P = 0.49). All the patients with normal cord showed neurological improvement at the end of the study [Table 3].

Table 3

Neurological recovery in patients with different magnetic resonance imaging findings

MRI findings	Number of patients	Neurological status on admission		Neurological improvement at discharge		Hospital stay (days), mean±SD
		No deficit	Deficit present	Improved	Not improved
Cord transection with contusion/edema	4	0	4	0	4	21±6.38
Cord hemorrhage	3	0	3	0	3	20±7.64
Cord contusion/edema	10	0	10	5	5	22±9.1
Cord edema	23	0	23	13	10	18±5.24
Spinal canal stenosis*	7	1	6	4	3	16±2.27
Normal cord	13	7	6	6	7	14±2.76
Epidural hematoma	3	0	3	2	1	24±2.62
Total		8	55	30	33	-

*Spinal canal stenosis due to retropulsed fracture fragments. SD: Standard deviation

Neurological outcome at the time of discharge

Change in the American spinal injury association status at discharge (initial American spinal injury association scale a)

Out of 22 patients with complete SCI (ASIA A), nine patients (41%) showed improvement in their neurological status over the period of hospital stay.

Change in the American spinal injury association status at discharge (initial American spinal injury association scale b)

Two patients were in ASIA B, among them one patient (50%) improved to ASIA C and one did not improve.

Change in the American spinal injury association status at discharge (initial American spinal injury association scale c)

Out of 13 patients with ASIA C, eight patients (61.5%) showed improvement and five patients (38.5%) did not improve.

Change in the American spinal injury association status at discharge (initial American spinal injury association scale d)

Out of 12 patients with initial ASIA D, 10 patients (83%) showed improvement, two patients (17%) did not improve.

Neurological outcome in patients of spinal trauma

In this study, out of 57 patients, 27 patients (49%) showed improvement in neurological status, 21 patients (37%) showed no improvement and eight patients (14%) had no neurological deficit at the time of admission.

DISCUSSION

In our prospective study, 57 patients underwent MRI for evaluation of spinal trauma with majority being males (86%) a trend reported elsewhere among the Indian population.[4] Young- and middle-aged men were the most common age group involved. Again a similar trend that has been reported in other studies.[567]

The most common cause of injury was fall from height, followed by RTA. A literature review by Chiu et al. also reported that the most common modes of injuries were fall from height, followed by RTA, in concordance with our study.[8] Other studies among Indian population have reported fall from height as the most common cause of SCI followed by RTA (44.5%–50.9%).[26] Singh et al. also reported that falls were more common in young adults, it partly explains the higher number of patients presenting with a history of fall from height.[6] It is possible that young- and middle-aged men are more active outdoors and this may probably explain their increased numbers in our study.

The cervical injury was the most common injury followed by dorsal and lumbar injuries. Cervical spine is probably the most common site for injury due to its excessive mobility and lack of supporting structures.[4] Cervical spine fractures are also more common among the elderly and uncommon among children.[9] Our study population mostly involved adults with <15% involving children. This may explain the similar incidence of cervical injuries reported elsewhere.[4] Other studies have reported that thoracolumbar spinal injuries are more common compared with cervical spine fractures.[9]

In our study, abnormal cord findings were observed in >75% of patients. Other studies have also reported a large proportion of patients with abnormal cord findings ranging from 70% to 75%.[310] Cord edema was the commonest cord signal abnormality detected, which is in agreement with the studies done by Parashari et al.[3] and Andreoli et al.[11] Other cord abnormalities observed were cord contusion/edema only, spinal canal stenosis due to retropulsion, cord transection with cord contusion/edema, cord hemorrhage, and epidural hematoma.

The categorization of patients into different groups based on ASIA impairments scale (AIS) showed that most common neurological status was ASIA A in 22 patients (39%), followed by ASIA C in 13 patients (23%), ASIA D in 12 patients (21%), ASIA E in eight patients (14%) and the least common being ASIA B in two patients (3.5%). Parashari et al.,[3] Andreoli et al.[11] and Magu et al.[2] all pointed out to ASIA A being the most common neurological status in patients with spinal trauma in agreement with our study.

Correlation of MRI findings with neurological outcome revealed that out of 57 patients, 28 patients (49%) showed improvement in their neurological status, 21 patients (37%) showed no improvement, whereas the remaining eight patients had no neurological deficit at the time of admission. Four patients with cord transection and three patients with cord hemorrhage did not show any improvement in their neurological status over the period of their hospital stay. The presence of these findings on MRI indicated dismal prognosis. Various studies have reported that cord hemorrhage was associated with the poor neurological outcome as compared to cord contusion and edema.[1213141516] Shepard and Bracken[17] showed that the presence of cord hemorrhage is associated with worse prognosis. Gupta et al.[4] reported that cord hemorrhage is associated with complete SCI with no recovery on follow-up. Qiu et al.[18] showed that all the cases with cord transection showed complete SCIs without neurological recovery. Cord transection should be the best predictors for complete SCI.

Among patients with cord contusion/edema pattern, the neurological improvement was seen in half of them. Similarly, more than half of patients with cord edema only (56.5%) showed improvement. These findings are similar to other studies done by Gupta et al.,[4] Ramón et al.[14] and Kulkarni et al.[10] Cord edema indicates the incomplete type of SCI as the damage at the cellular level is reversible to some extent. Many of these patients initially show evidence of neurological deficit. However, there is a good chance of neurological recovery and usually has a favorable outcome.

Out of three patients with epidural hematoma, two patients showed improvement, whereas one patient did not improve. As the sample size of patients with epidural hematoma was meager, we could not make a meaningful analysis. Andreoli et al.[11] reported that patients with cord hemorrhage had poor prognosis while those with cord edema had a better prognosis. Flanders et al. showed that patients without spinal cord hemorrhage had significant improvement in their neurological status.[19] Selden et al. reported that severe cord compression by extra-axial hematoma is associated with poor neurological function and also showed that MRI after SCI provides accurate prognostic information regarding neurological function.[20] These studies support our finding that the presence of cord hemorrhage may be a poor prognostic indicator of neurological recovery. All the patients with normal cord findings with initial neurological deficit improved in our study, showing a positive correlation between initial normal MRI and neurological recovery.

Patients with initial high-grade ASIA A have lower chances of recovery (41%), whereas the maximum chances of recovery were associated with low-grade ASIA C and D (61.5% and 83%, respectively). This corresponds with the severity of MRI findings in patients with initial ASIA A score, which included cord transection, cord hemorrhage, and epidural hematoma. Harrop et al., reported 7% improvement in patients with initial ASIA A score, which dramatically increased to 94.3% in patients with initial ASIA D score.[21] Rao et al. also observed that none of the patients with ASIA A showed neurological improvement whereas all patients in ASIA D showed improvement. They also observed that 84% of patients with cord edema showed improvement.[22] Magu et al. reported that 86% of patients with cord edema showed neurological improvement.[2] Our findings among patients with initial ASIA A score have been more optimistic compared with data reported by Harrop et al. and Rao et al.[2122] It could possibly due to >40% of patients with initial ASIA A score had cord edema and/or cord contusion, among which about half of patients showed improvement as discussed before.

Neurological improvement was observed by two-third of patients with spinal canal stenosis in our study. Miyanji et al. reported that the final outcome of neurological recovery has no significant correlation with the presence or absence of canal stenosis.[23] Saifuddin suggested that MRI is particularly useful in unconscious patients who cannot undergo motor and sensory, neurological evaluation.[24]

Our study had some limitations. Some of the MRI findings such as epidural hemorrhage were small and as such a meaningful interpretation could not be performed. We did not perform diffusion tensor imaging (DTI). DTI helps delineate the preserved white matter tracts at the level of injury as well as the extent of injury to distant normal-appearing white matter.

CONCLUSION

We concluded that MRI plays a major role in the diagnosis of SCIs, directing early and prompt management and predicting prognosis of neurological recovery. MRI can depict changes in the injured spine and cord. Cord edema and normal cord were associated with favorable neurological outcome. Cord contusion showed poor neurological recovery, as compared to cord edema and normal cord. Cord transection and cord hemorrhage were associated with complete SCI and with poor neurological recovery. MRI findings in acute SCI correlate well with the initial neurological deficits on admission and at the time of discharge according to ASIA impairment scale. We recommend MRI to be performed in cases with spinal trauma.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.