Correlation of magnetic resonance imaging biomarkers (tissue bridges) with neurological recovery following traumatic spinal cord injury

Abstract

BACKGROUND

Spinal cord injury (SCI) imposes enduring physical impairments and substantial socio-economic burdens. These injuries are either traumatic incidents or ischemic but exhibit comparable clinical recoveries. This suggests shared underlying neurodegenerative mechanisms, such as neuronal cell death, demyelination, and axonal degeneration, regardless of aetiology.

AIM

To investigate the relationship between the magnetic resonance imaging (MRI) biomarkers (tissue bridges) and clinical outcome in acute traumatic SCI.

METHODS

In this prospective study adult patients with acute SCI who were examined clinically and radiologically within first 48 hours and subsequently at intervals were included. Clinical assessment included sensory score, motor score and zone of partial preservation. Radiological scores included measurement kyphotic deformities - sagittal index, regional kyphosis, gardener segmental kyphotic deformity. MRI on 3 Tesla machine was done to evaluate quantitative & qualitative parameters, and tissue bridges at one and 6 months.

RESULTS

There were 47 patients with a mean age of 40.43 ± 10.73 years and male/female ratio of 34:13. There was a significant (P < 0.05) improvement in clinical, radiological, and MRI parameters at 6 months. Maximum spinal cord compression (MSCC), maximal canal compression, lesion length, width, and area; dorsal tissue bridges; ventral tissue bridges, total width of tissue bridges; and midsagittal tissue bridge ratios at one month significantly (P < 0.05) correlated with the 6-month total motor score and total sensory score. Further, the regression analysis demonstrated clinical improvement to dimensions of tissue bridges at 6 months.

CONCLUSION

The MRI imaging biomarkers in SCI patients demonstrated substantial improvement over time. There was a negative correlation between neurological recovery, MSCC, maximum canal compromise, and lesion dimensions (lesion length, lesion width, and lesion area). Higher canal compromise and lesion dimensions were associated with a poorer outcome. The evaluation of the midsagittal tissue bridge (including the ventral tissue bridge, dorsal tissue bridge, total width of the tissue bridge, and tissue bridge ratios) at 1 and 6 months showed a positive correlation with the neurological recovery.

Key Words: Spinal cord injury; Magnetic resonance imaging; Tissue bridges; ASIA impairment scale

Core Tip: This prospective study highlights the prognostic value of magnetic resonance imaging (MRI) biomarkers, particularly tissue bridges, in acute spinal cord injury (SCI). Quantitative MRI parameters, especially midsagittal tissue bridge dimensions assessed at 1 and 6 months, positively correlated with motor and sensory recovery. Conversely, greater spinal canal compromise and lesion size predicted poorer outcomes. These findings underscore the utility of early MRI-based assessment in predicting neurological recovery and guiding management in traumatic SCI.

INTRODUCTION

Spinal cord injury (SCI) imposes enduring physical impairments and substantial socio-economic burdens. These injuries can stem from either traumatic incident like falls or non-traumatic causes such as ischemia. Interestingly, regardless of the injury's origin, patients with traumatic and ischemic thoracic and cervical SCI exhibit comparable clinical recoveries. This suggests shared underlying neurodegenerative mechanisms, such as neuronal cell death, demyelination, and axonal degeneration, regardless of etiology. However, while the neurodegenerative processes in cervical SCI have been extensively studied, the structural changes and prognostic significance of these changes in thoracic SCI remain relatively unexplored; and the literature is silent about the combine (pooled data) about cervical, thoracic, and lumbar[1,2].

Although conventional magnetic resonance imaging (MRI) establishes prognostic indicators, it falls short as a neuroimaging biomarker due to its limited insights into nerve fibre tract conditions[3]. The functional recovery after incomplete SCI involves central nervous system plastic adaptations[4].

MRI serves as a valuable tool in clinical diagnostics and prognosis, particularly in SCI cases. T2-weighted scans obtained from the lesion epicenter offer critical insights into the spatiotemporal evolution of the injury, facilitating the quantification of various intramedullary processes such as edema, hemorrhage, and spinal cord compression, particularly in tetraplegic patients[5,6].

Conventional MRI captures macrostructural changes, but quantitative neuroimaging (qMRI) techniques offer precise microstructural insights. qMRI measures myelin, axonal density, iron deposition, and metabolic profiling, demonstrating evolving markers of neurodegeneration in the spinal cord and brain[7]. MRI is commonly used post-traumatic SCI to assess intramedullary damage. Lesion quantification reveals correlations between severity and clinical impairment. Weeks post-injury, preserved neuronal tissue forms midsagittal tissue bridges, crucial for electrophysiologic communication. These bridges, both dorsal and ventral to posttraumatic cysts, predict clinical recovery.

Preserved tissue bridges adjacent to the lesion cavity in incomplete SCI patients facilitate electrophysiological information flow and predict functional recovery. This explores the width of ventral and dorsal tissue bridges, identified through T2-weighted scans, to predict the onset and persistence of SCI-related neuropathic pain[2,8,9]. This study aimed to investigate the correlation between MRI biomarkers—specifically tissue bridges observed one month after traumatic SCI—and neurological recovery across a cohort with varying injury levels.

MATERIALS AND METHODS

This is a prospective study conducted at tertiary care centre in Northern India with approval from the institute ethics committee (IEC- BREC/22/TH/Ortho-08). The study period was extended from November 2022 to March 2024. Consecutive patients reporting to institute, aged between 18-65 years of either sex, with acute SCI and who consented to participate were included in the study. Patients with non-traumatic cause for SCI, patients with head injury/medically unstable condition, patients with previous implanted metallic devices, patients with claustrophobia, pacemakers and cochlear implants, gunshot wounds were excluded from the study.

A sample size of 45 patients was considered necessary to detect statistical significances with an effect size of 0.67 at alpha 0.05 and power of 90% based on previous study by Huber et al[6]. Considering the total loss of patients as 10%, 50 consecutive patients were finally enrolled in the study. Three patients were lost to follow-up; and finally, 47 patients who completed minimum of six-month follow-up were evaluated.

Each patient was thoroughly examined clinically and was subjected to investigations like X-ray spine, computed tomography (CT), and MRI within 48 hours of injury. MRI was done in a 3.0 T (Discovery 750w of GE make, Milwaukee, WI, United States) superconductive MRI machine was used with gradient field 40 mT/m and gradient switching rate 150 mT/millisecond. Axial and sagittal images were acquired in order to accurately identify the midsagittal slice. The midsagittal slice was identified using multiplanar reconstruction on T2-weighted images, selecting the slice that best visualized the cerebral aqueduct, corpus callosum, and brainstem midline structures in true sagittal alignment. Measurements of lesion dimensions and tissue bridges were made on this slice. The measurement angles were standardized to be perpendicular (90°) to the rostro-caudal axis, defined by a reference line drawn from the superior surface of the corpus callosum to the inferior margin of the brainstem on the midsagittal plane. For the qualitative analysis, all sequences were considered. The sagittal T2-weighted images were used for quantitative analysis of lesion size and midsagittal tissue bridges. Table 1 shows the details of acquisition of MRI images.

Table 1 Details of acquisition of magnetic resonance images.

		Axial TR	Slice gap	Slice thickness (mm)	FOV (cm)	NEX	TE
T1	Plan	650	1	4	1	2	10
T2	Sagittal	3841	1	4	0.80	2	102
T1	Sagittal	2671	1	4	1	2	24
T2	Axial	6847	1	4	0.80	3	112

TR: Repetition time; FOV: Field of view; NEX: Number of excitations; TE: Echo time.

One rater (radiologist), blinded to patient identity and scan time point, assessed the anterior-posterior width, rostro-caudal length, and total area of the lesion, and the width of tissue bridges on the midsagittal slice.

Clinical assessment (sensory score, motor score and zone of partial preservation) were done at the time of admission, 3^rd day, 7^th day, one month, 3 months, and 6 months as per international guidelines (ASIA impairment scale)[10].

Spinal stability for thoracolumbar injuries was evaluated according to thoracolumbar injury classification and severity score classification. Unstable spine was taken up for spinal surgery (stabilisation alone or stabilisation with decompression). Type of surgery was decided based on personality of fracture[11,12].

Patients were followed up at one, 3 and 6 months. Clinical evaluation and plain radiography were done at each follow up. MRI was done at one month, 3-months and 6-months follow up. Neurological recovery was documented as per ASIA impairment scale (AIS)[10] and following outcome were measured and assessed.

Radiological outcome

The following radiological measurements was done in subsequent follow ups.

Kyphotic deformities (on sagittal view X-ray): (1) Sagittal index (SI): SI is segmental kyphosis at the vertebrae level adjusted for the baseline sagittal contour in the normal spine. Kyphotic deformity (KD) at the fracture motion segment level minus normal contour [NC; SI = KD - NC] was calculated. The baseline sagittal curve/Level was estimated using a 5° angle in thoracic segments, 0 degree at thoracolumbar junction, and 10° lumbar segments; (2) Regional kyphosis: The measurement of an angle created between the lines drawn on the superior end plate of the upper normal vertebra and the inferior end plate of the lower normal vertebra[13]; and (3) Gardener segmental KD: It was measured as the angle formed from lines drawn on the lower end plate of the adjacent normal vertebrae.

MRI parameters: These were measured on MRI as per following criteria: Three quantitative measures were used: Maximum canal compromise (MCC), maximum spinal cord compression (MSCC) and length of lesion. Mid-sagittal T1- and T2-weighted imaging were used to determine the MCC and MSCC respectively as described by Fehlings et al[14]. The values were ascertained by measuring the distance of the canal or spinal cord one segment above and below the lesion to compute the average distance. The distance was subsequently measured at the lesion site and expressed as a percentage of the average. The length, width and area of the lesion were determined on T2-weighted images. The length was defined as the distance between the most superior and most inferior extent of the lesion. The midsagittal T2-weighted image was utilised to measure the width of the ventral and dorsal tissue bridges, defined as the shortest distance from the cystic cavity to the ventral and dorsal edges of the spinal canal, respectively, at a right angle to the rostro-caudal orientation of the spinal cord, with the total width of the tissue bridges being the sum of both measurements. The ratios of midsagittal tissue bridges were determined by dividing the total tissue bridges by the diameter of the cord.

Neurologic outcome

A neurological examination was conducted at each follow-up assessment. An enhancement in motor power, restoration of sensation, and recovery of bladder and bowel function were observed, and graded to the ASIA score[10].

RESULTS

There were 47 patients with mean age of 40.43 ± 10.73 years and male is female ration of 34:13. The detail of demography are given in Table 2. The details of symptomatology among the subjects in tabulated in Table 3. The comparison motor index score (MIS) and sensory index score (SIS) at the 1-month, 3-month and 6-month intervals is outlined in Table 4. Table 5 shows distribution of subjects according to neurological parameters on clinical examination. The radiological indices at various time intervals are outlined in Table 6. Table 7 gives the correlation coefficients the radiological indices in relation to tissue bridges at 1 and 6 months. Table 8 compares the length, width, and area of the lesion at different time intervals. Finally, the regression analysis between various clinical measures at 6 months and MR parameters at 1 month is given in Table 9.

Table 2 Demographic data of the cohort.

Variable		Frequency	Percentage (%)
Age (year)	< 30	10	21.3
	31-40	9	19.1
	41-50	22	46.8
	> 60	6	12.8
Gender	Female	13	27.7
	Male	34	72.3
Socio economic status	Lower class	17	36.2
	Lower middle class	21	44.7
	Upper middle class	9	19.1
Marital status	Married	43	91.5
	Unmarried	4	8.5
Educational status	Higher sec certificate	7	14.9
	Secondary	3	6.4
	Middle school	4	8.5
	Primary school	12	25.5
	Illiterate	21	44.7
Level of injury	Cervical	8	17.02
	Upper dorsal (up to D5)	8	17.02
	Lower dorsal (D6-D11)	16	34.04
	Thoracolumbar area (D12-L2)	15	31.91
Mode of injury	Diving injury	4	8.5
	Fall	7	14.9
	Fall from height	14	29.8
	Motorcycle injury	9	19.1
	RSA	13	27.7

RSA: Road side accident.

Table 3 Distribution of subjects according to their symptomatology.

Pain		Initial		1 month		3 months		6 months		P value1
		No. of cases	Percentage (%)	No. of cases	Percentage (%)	No. of cases	Percentage (%)	No. of cases	Percentage (%)
Pain score (VAS)	No	0	0.0	20	42.6%	47	100.0	47	100.0	< 0.001
	Mild	0	0.0	22	46.8%	0	0.0	0	0.0
	Moderate	0	0.0	5	10.6%	0	0.0	0	0.0
	Severe	47	100.0	0	0.0%	0	0.0	0	0.0
Swelling		47	100.00	5	10	4	8.50	3	6.40	< 0.001
Deformity		5	10.60	5	10.60	5	10.60	5	10.60	< 0.001
Weakness-lower limb		47	100.00	47	100.00	10	21.30	10	21.30	0.008
Weakness-upper limb		5	10.60	0	0.00	3	6.40	3	6.40	< 0.001
Incontinence		47	100.00	3	6.40	0	0.00	0	0.00	< 0.001
Retention		47	100.00	3	6.40	0	0.00	0	0.00

¹Cochran’s Q test.

VAS: Visual analog scale.

Table 4 Comparison motor index score and sensory index score at the 1-month, 3-month and 6-month intervals.

MIS		Mean	SD	t value	P value1	Difference			95% confidence interval of the difference
						Mean	SD	SE mean	Lower	Upper
MIS upper limb (month)	1	44.32	6.29
	3	44.85	6.31	-5.080	0.001	-0.53	0.72	0.10	-0.74	-0.32
	6	45.68	5.75	-6.643	0.001	-1.36	1.41	0.20	-1.77	-0.95
MIS lower limb	1	31.21	14.57
	3	34.57	13.74	-5.601	0.001	-3.36	4.11	0.60	-4.57	-2.15
	6	38.23	13.63	-6.466	0.001	-7.02	7.44	1.09	-9.21	-4.84
Total (MIS)	1	75.53	14.99
	3	79.43	15.13	-6.699	0.001	-3.89	3.98	0.58	-5.06	-2.72
	6	83.91	15.54	-8.358	0.001	-8.38	6.88	1.00	-10.40	-6.36
SIS light touch (R + L)	1	92.60	17.41
	3	94.74	17.24	-12.121	0.001	-2.15	1.22	0.18	-2.51	-1.79
	6	99.85	17.88	-11.647	0.001	-7.26	4.27	0.62	-8.51	-6.00
SIS pin prick (R + L)	1	93.55	17.32
	3	96.21	17.35	-8.109	0.001	-2.66	2.25	0.33	-3.32	-2.00
	6	101.13	17.58	-11.678	0.001	-7.57	4.45	0.65	-8.88	-6.27
Total SIS	1	186.15	34.50
	3	190.96	34.30	-12.632	0.001	-4.81	2.61	0.38	-5.57	-4.04
	6	200.98	35.16	-14.144	0.001	-14.83	7.19	1.05	-16.94	-12.72

¹Repeated measures ANOVA.

n = 47 at all-time points for all variables. MIS: Motor index score; SIS: Sensory index score; R: Right; L: Left.

Table 5 Distribution of subjects according to neurological parameters on clinical examination.

Variables		1 month		3 months		6 months		P value1
		No. of cases	Percentage (%)	No. of cases	Percentage (%)	No. of cases	Percentage (%)
ASIA score	A	10	21.3	2	4.3	2	4.3	0.548
	B	0	0	0	0	0	0
	C	10	21.3	12	25.5	9	19.1
	D	22	46.8	28	59.6	21	44.7
	E	5	10.6	5	10.6	15	31.9
Voluntary anal contraction	0	14	29.8	17	36.2	14	29.8	0.857
	1	22	46.8	22	46.8	22	46.8
	2	11	23.4	8	17.0	11	23.4
Temperature		8	17.0	4	8.5	0	0.0	0.345
Deep anal pressure		12	25.5	4	8.5	0	0.0	0.548
Superficial reflexes	Abdominal reflex (absent)	18	40	4	8.0	4	8.0	< 0.001
	Babinski reflex (absent)	47	100	47	100	47	100	-
	Patellar (absent)	18	40.0	8	16.0	6	12.0	< 0.05
	Ankle (absent)	18	40.0	8	16.0	4	8.0	< 0.01
	Clonus (absent)	47	100	47	100	47	100	-

¹Cochran’s Q test.

n = 47 at all-time points for all variables.

Table 6 The radiological indices at various time intervals.

Radiological variables	Time interval in months	Mean	SD	t value	P value1	Difference			95% confidence interval of the difference
						Mean	SD	SE mean	Lower	Upper
Sagittal index	1	10.30	3.20
	3	5.41	2.19	12.468	0.000	4.88	2.68	0.39	4.09	5.67
	6	4.70	1.56	12.140	0.000	5.60	3.16	0.46	4.67	6.52
Regional kyphosis	1	17.40	2.60
	3	16.68	2.12	2.174	0.035	0.72	2.28	0.33	0.05	1.39
	6	16.76	2.08	2.028	0.048	0.65	2.19	0.32	0.00	1.29
Gardener segmental kyphotic deformity	1	16.00	0.42
	3	15.47	1.63	2.477	0.017	0.53	1.47	0.21	0.10	0.96
	6	15.47	1.64	2.452	0.018	0.53	1.49	0.22	0.10	0.97

¹Repeated measures ANOVA.

n = 47 at all-time points for all variables.

Table 7 Correlation coefficients and P-values for the relationship between radiological measurements at 1 and 6 months (sagittal index, regional kyphosis, gardener segmental kyphotic deformity) and magnetic resonance imaging parameters related to tissue bridges at 1 and 6 months.

MRI parameters at 1 month	1 month				6 months
	Spearman’s rho	Sagittal index	Regional kyphosis	Gardener segmental kyphotic deformity	Sagittal index	Regional kyphosis	Gardener segmental kyphotic deformity
Ventral tissue bridge	Correlation coefficient	0.006	0.108	0.226	-0.161	0.134	0.073
	P value	0.969	0.470	0.126	0.280	0.369	0.628
Dorsal tissue bridge	Correlation coefficient	0.024	0.047	-0.181	-0.055	-0.182	-0.094
	P value	0.871	0.753	0.224	0.714	0.221	0.529
Total width of tissue bridges	Correlation coefficient	0.037	0.082	0.152	-0.112	-0.019	0.045
	P value	0.806	0.583	0.307	0.452	0.901	0.766
Midsagittal tissue bridge ratios	Correlation coefficient	0.005	-0.011	-0.095	0.046	-0.116	-0.076
	P value	0.973	0.940	0.525	0.760	0.436	0.612

n = 47 at all-time points for all variables. MRI: Magnetic resonance imaging.

Table 8 Compares the mean and SD of lesion length, width, and area at different time intervals.

MRI parameters	Mean	SD	t value	P value1	Difference			95% confidence interval of the difference
					Mean	SD	SE mean	Lower	Upper
Lesion length (mm)
1 month	33.96	27.09
3 months	18.43	14.68	4.808	0.001	15.53	22.14	3.23	9.03	22.03
6 months	14.14	15.59	4.837	0.001	19.82	28.09	4.10	11.57	28.07
Lesion width (mm)
1 month	4.65	1.76
3 months	3.94	1.65	4.900	0.001	0.71	0.99	0.14	0.42	1.00
6 months	3.57	1.73	6.957	0.001	1.07	1.06	0.15	0.76	1.39
Lesion area (mm2)
1 month	110.66	83.41
3 months	80.05	32.29	2.532	0.015	30.61	82.88	12.09	6.27	54.94
6 months	73.35	37.22	2.700	0.010	37.31	94.76	13.82	9.49	65.14

¹Repeated measures ANOVA.

n = 47 at all-time points for all variables. MRI: Magnetic resonance imaging.

Table 9 Regression analysis between various clinical measures at 6 months and magnetic resonance parameters at 1 month.

Clinical measure at 6 months; dependent variable	MR parameter at 1 month; independent variable	R	R square	P value1	Standardized regression coefficients	95% confidence interval
						Lower bound	Upper bound
MIS upper limb	MSCC	0.454	0.206	0.001	-1.442	-2.291	-0.592
	MCC	0.135	0.018	0.365	-0.232	-0.744	0.279
	Lesion length (mm)	0.098	0.010	0.513	0.021	-0.043	0.084
	Lesion width (mm)	0.030	0.001	0.840	0.099	-0.879	1.076
	Lesion area (mm2)	0.217	0.047	0.142	0.015	-0.005	0.035
	Total width of tissue bridges	0.301	0.090	0.040	3.755	0.179	7.332
	Midsagittal tissue bridge ratios	0.342	0.117	0.019	25.095	4.408	45.781
MIS lower limb	MSCC	0.194	0.038	0.192	-1.459	-3.676	0.759
	MCC	0.144	0.021	0.333	-0.588	-1.798	0.622
	Lesion length (mm)	0.242	0.059	0.101	-0.122	-0.268	0.025
	Lesion width (mm)	0.368	0.135	0.011	-2.841	-4.997	-0.685
	Lesion area (mm2)	0.286	0.082	0.051	-0.047	-0.094	0.000
	Total width of tissue bridges	0.209	0.044	0.159	6.182	-2.512	14.876
	Midsagittal tissue bridge ratios	0.146	0.021	0.328	25.355	-26.281	76.991
Total (MIS)	MSCC	0.338	0.114	0.020	-2.900	-5.326	-0.475
	MCC	0.177	0.031	0.235	-0.820	-2.193	0.552
	Lesion length (mm)	0.176	0.031	0.237	-0.101	-0.271	0.069
	Lesion width (mm)	0.311	0.097	0.033	-2.743	-5.255	-0.231
	Lesion area (mm2)	0.171	0.029	0.251	-0.032	-0.087	0.023
	Total width of tissue bridges	0.294	0.087	0.045	9.937	0.250	19.625
	Midsagittal tissue bridge ratios	0.255	0.065	0.084	50.450	-7.101	108.000
SIS light touch	MSCC	0.359	0.129	0.013	-3.543	-6.312	-0.775
	MCC	0.402	0.162	0.005	-2.149	-3.618	-0.679
	Lesion length (mm)	0.145	0.021	0.330	-0.096	-0.292	0.100
	Lesion width (mm)	0.341	0.116	0.019	-3.458	-6.318	-0.598
	Lesion area (mm2)	0.337	0.114	0.021	-0.072	-0.133	-0.012
	Total width of tissue bridges	0.077	0.006	0.608	2.985	-8.646	14.616
	Midsagittal tissue bridge ratios	0.069	0.005	0.645	15.748	-52.577	84.073
SIS pin prick	MSCC	0.353	0.125	0.015	-3.427	-6.155	-0.700
	MCC	0.333	0.111	0.022	-1.748	-3.236	-0.261
	Lesion length (mm)	0.086	0.007	0.565	-0.056	-0.250	0.138
	Lesion width (mm)	0.322	0.104	0.027	-3.206	-6.037	-0.375
	Lesion area (mm2)	0.264	0.070	0.073	-0.056	-0.117	0.005
	Total width of tissue bridges	0.064	0.004	0.670	2.434	-9.008	13.876
	Midsagittal tissue bridge ratios	0.084	0.007	0.576	18.770	-48.307	85.846
Total SIS	MSCC	0.359	0.129	0.013	-6.971	-12.413	-1.528
	MCC	0.371	0.137	0.010	-3.897	-6.828	-0.966
	Lesion length (mm)	0.117	0.014	0.434	-0.152	-0.539	0.235
	Lesion width (mm)	0.334	0.112	0.022	-6.664	-12.301	-1.027
	Lesion area (mm2)	0.303	0.092	0.038	-0.128	-0.248	-0.007
	Total width of tissue bridges	0.071	0.005	0.636	5.419	-17.459	28.297
	Midsagittal tissue bridge ratios	0.077	0.006	0.607	34.518	-99.736	168.773

¹Spearman's rho correlation.

n = 47 at all-time points for all variables. MR: Magnetic resonance; MIS: Motor index score; SIS: Sensory index score; MSCC: Maximum spinal cord compression; MCC: Maximum canal compromise.

Few case illustrations are given in Figures 1, 2, 3, 4 and 5.

Figure 1 Magnetic resonance imaging T2-weighted sagittal images of an 18-years-old patient who sustained trauma to thoracic spine (D10-D11) with ASIA grade C. A: Dorsal tissue bridge (1.09 mm); B: Ventral tissue bridge (0.38 mm); C: The lesion length (5.71 cm); D: Lesion width (6.81 mm); E; Lesion area (1.785 cm²); F: Normal width of spinal cord (5.52 mm).

Figure 2 Magnetic resonance imaging T2-weighted sagittal images of a 30-years old patient who sustained trauma to thoracic spine(D11) with ASIA grade C. A: Dorsal tissue bridge (0.88 mm); B: Lesion area (0.7007 cm²); C: Lesion width (5.29 mm); D: Lesion length (2.15 cm); E: Ventral tissue bridge (0.57 mm); F: Normal width of spinal cord (5.66 mm).

Figure 3 Magnetic resonance imaging T2-weighted sagittal images of a 39-years old patient who sustained trauma of cervical spine (C5-C6) with ASIA grade D. A: Lesion area (1.008 cm²); B: Lesion length (2.61 cm); C: Lesion width (7.38 mm); D: Dorsal tissue bridge (0.62 mm); E: Ventral tissue bridge (0.94 mm); F: Normal width of spinal cord (7.49 mm).

Figure 4 Magnetic resonance imaging T2-weighted sagittal images of a 48-years old patient who sustained trauma of cervical spine (C5-C6) with ASIA grade C. A: Ventral tissue bridge(1.19mm); B: Dorsal tissue bridge (0.86 mm); C: Lesion area (1.546 cm²); D: Lesion width (3.83 mm); E: Lesion length (5.42 cm); F: Normal width of spinal cord (7.50 mm).

Figure 5 Magnetic resonance imaging T2-weighted sagittal image of a 34-years old patient who sustained trauma of thoracolumbar spine (D12-L1) with ASIA grade A. A: Ventral tissue bridge (1.44 mm); B: Dorsal tissue bridge (0.72 mm); C: Lesion area (2.498 cm²); D: Lesion length (5.79 cm); E: Lesion width (1.07 cm); F: Normal width of spinal cord (1.12 cm).

DISCUSSION

Demography

We found the mean age of patients in present study was 40.23 ± 10.73 year. Pfyffer et al[2] had a mean of 51.20 ± 20.07 in 2019, but Singh et al[15] had a mean age of 37.32 ± 13.31 in 2020. Male were more affected in our study similar to previous study by Singh et al[15], but contrary to study by Pfyffer et al[2]. The study population predominantly consisted of individuals with low socioeconomic status and high illiteracy rates, reflective of the regional demographics in Northern India. These factors can influence health-seeking behavior, time to presentation, and access to care, which in turn may affect injury outcomes. The majority (> 90%) were married. Fall from height, accounting for nearly a third of the cases and road traffic accidents (RTAs) represented about 27.7% of cases. These findings were consistent with other studies that have also identified falls and RTA as leading causes of SCIs[16,17]. In developed countries, RTAs tend to be the primary cause, whereas fall from height remain prevalent in developing countries like those represented in this study[14].

We found SCI at various levels, with a notable concentration at levels D4-D5, L1, and D12-L1. These findings align with previous studies that also observed clustering of SCI patients around the D12 and L1 levels. Research by Chadha and Bahadur[18] noted D12-L1 as the most common level in 80% of cases, while studies by Liu et al[19] and Knop et al[20] reported L1 as predominant in 40% and 55% of cases, respectively[19,20]. Additionally, Singh et al[21] found L1 to be the most involved level in 36% of cases, followed by D12 in 24%. The increased incidence can be attributed to transition from fixed to mobile segments and from thoracic kyphosis to lumbar lordosis, making it a zone of higher biomechanical stresses during trauma. Importantly, injuries at levels D12 and L1, and above, often result in poorer neurological outcomes due to the anatomical endpoint of the spinal cord around the lower border of L1 or upper L2 level. The mean time elapsed between injury and imaging was 21.38 ± 6.28 hours. It ranged between 12-45 hours. In a study by Shimada et al[22], all patients were imaged within 48 hours of injury. Most patients were imaged rapidly after injury; as cord lesions evolve during the early phase, it should assist in efforts to correlate neurological findings with those on MRI.

Symptomatology

The study assessed pain outcomes over a period from initial assessment up to six months among 47 the patients. Initially, all the patients reported severe pain. There was gradual but statistically significant (P < 0.001) improvement during first six months. Weakness in the lower limbs was universally reported initially and persisted in a significant (P = 0.008) proportion of cases up to the 3-month mark, after which a gradual decrease was observed. Weakness in the upper limbs, while less common, was present in a subset of cases with cervical spine injury at the beginning and showed a slight increase in prevalence at the 3-month follow-up (P < 0.001). Incontinence and retention exhibited similar patterns, being prevalent initially and gradually decreasing over time, with a notable decrease observed at the 6-month mark (P < 0.001). Tenderness and deformity were consistently observed in all cases across all time points, indicating their persistent presence throughout the observation period.

Neurological assessment

In present study we had incomplete SCI with preserved motor function below the injury site in majority of the cases. There were improvements noted across most ASIA categories (P = 0.548).

Time interval: Similar findings were noted by Singh et al[21], Narasinga Rao et al[23], and Butt et al[24], varying degrees of neurological recovery were observed among patients with traumatic SCIs. These findings underscore that the extent of initial spinal cord trauma correlates closely with subsequent neurological deficits and recovery outcomes. Patients with incomplete lesions generally showed more substantial neurological improvement compared to those with complete injuries.

Similarly, we had improvement in MIS upper limb scores (44.32 ± 6.29 to 45.68 ± 5.7 at 6 months) and lower limb (31.21 ± 14.57, 38.23 ± 13.63 at 6 months). The study by Steeves et al[25] investigated the pattern of MIS improvement over time in individuals with SCI. Improvements occurred in MIS within the first six-month post-injury. Most people recover similar motor points or levels regardless of initial cervical motor level. We also observed improvement in SIS for light touch (92.60 to 99.85 at 6 months) and pin prick (93.55 to 101.13 at 6 months). On the contrary, Vasquez et al[26] found a disparity (P < 0.001) between light touch (64.5 ± 3.2, mean ± SE) and pin prick (54.7 ± 2.9) AIS sensory scores.

Radiological findings

We had KD correction as evidenced by improvement in radiological indices. The same has been observed in various studies[15,21,27-30]. However, there is a weak correlation between the measurements of SI, regional kyphosis, with the MRI biomarkers and midsagittal tissue bridges at 1 month and 6 months. The weak correlations between sagittal alignment (e.g., segmental kyphosis, spinal index) and tissue bridge dimensions suggest that spinal angulation alone may not strongly influence intramedullary tissue preservation. This may reflect the multifactorial nature of secondary SCI, in which vascular compromise, edema, and microstructural disruption play more prominent roles than gross alignment. Moreover, KD correction may not necessarily reverse internal cord damage already established at the time of injury. Previous studies have similarly shown limited direct association between alignment correction and neurological recovery (e.g., Fehlings et al[14], 1999). These findings underscore the need to evaluate both mechanical stability and neuroanatomical preservation as independent but complementary targets in spinal trauma management.

Unlike study by Hayashi et al[31] who assessed MSCC on MRI, the present study used an objective method to quantify MR images obtained from patients with SCI. We found improvement in MSCC (13.12 to 12.26) MSCC (P = 0.063) was predictive of a poor neurological outcome.

Chandra et al[18] found that cord compression severity affected prognosis and was an important outcome measure.

MRI findings

Compared to study conducted by Pfyffer et al[2] on 12 patients with SCI who underwent longitudinal follow-up scans showed a 5.68 mm² decrease in lesion area per month in 2019. In 2017, Aarabi et al[32] reported ASI grade conversion had statistically significant relationship with admission ASI grade, lesion length and haemorrhage. Intramedullary lesion was the sole and strongest indicator of AIS grade conversion[32]. In present study lesion length, width, and area exhibited mixed correlations with MISs and SISs for light touch and pin prick. Notably, lesion width and area show strong negative correlations with all MISs and SIS scores at the 6-month interval (P < 0.05). Pfyffer et al[8] also reported significant negative correlation of one-month post-SCI lesion area and length with lower extremity motor score, light touch, and pinprick scores at one year of injury.

The widths of ventral tissue bridges over three time points: 1 month, 3 months, and 6 months

Over months, there were minor fluctuations in these widths, with some segments showing slight increases while others remained stable. Ventral tissue bridge at 1 month showed statistically significant positive correlations with MIS at 6-month time interval (r = 0.294, P = 0.004); while statistically non-significant positive correlations with SIS light touch (r = 0.77, P = 0.188), SIS pin prick (r = 0.264, P = 0.384), and Total SIS (r = 0.303, P = 0.184) were observed at 6-month time interval. Ventral tissue bridge at 6-month showed statistically significant positive correlations with MIS at 6-month time interval (r = 0.165, P = 0.006).

Like this study, Vallotton et al[9] found that ventral tissue bridge width predicted better lower extremity motor scores at 12 months (r = 0.41, P = 0.035) independent of baseline clinical score and dorsal tissue bridges. Midsagittal tissue bridge ratios were also associated with six-minute walk distance (r = 0.68, P = 0.03), according to O'Dell et al[33]. Contrary to the findings of the present study, Pfyffer et al[8] documented an increased width of ventral tissue bridges—a surrogate for spinothalamic tract functionality—at one month following SCI, which correlated with the onset and persistence of neuropathic pain and heightened pin-prick sensitivity.

Dorsal tissue bridges at I month did not show any statistically significant correlation with MIS at 6-month time interval (r = 0.294, P = 0.771); while SIS pin prick (r = 0.264, P = 0.039) and total SIS (r = 0.303, P = 0.05) showed statistically significant correlation at 6-month time interval. No such correlations of these parameters were observed with dorsal tissue bridges at 6-month. Vallotton et al[9] also reported greater width of dorsal tissue bridges predicted better light-touch score at 12 months (r = 0.40, P = 0.045) independently of baseline clinical score and ventral tissue bridges. These findings can be explained based on spaciotemporal orientation of sensory fibres (tactile, vibrational, and touch) in the dorsal column of the spinal cord.

Midsagittal tissue bridge ratios exhibited positive correlations with MISs and SIS scores at the 6-month interval, though not all correlations were statistically significant (P < 0.05). Pfyffer et al[2] also reported improved pinprick scores (P = 0.004, n = 21, r = 0.610) at 1-year post-injury was associated with wider midsagittal tissue bridges at 1-month post-SCI. Moreover, there was a trend toward improved light touch scores with wider midsagittal tissue bridges (P = 0.082, n = 21, r = 0.398).

The results of the present study revealed a significant positive correlation between midsagittal tissue bridge ratios and the MIS - upper limb (P = 0.019), suggesting that greater presence or extent of these tissue bridges aligns with improved motor function recovery over time. Similarly, Pfyffer et al[2] also reported improved LEMS (P = 0.022, n = 21, r = 0.508) and pinprick scores (P = 0.004, n = 21, r = 0.610) at 1year post-injury were linked with wider midsagittal tissue bridges at 1-month post-SCI.

Smith et al[34] conducted a study involving 136 participants with cervical SCI, establishing a significant correlation between wider ventral tissue bridges and pinprick scores (r = 0.31, P < 0.001). Moreover, broader dorsal tissue bridges exhibited substantial correlations with light touch scores at discharge from inpatient rehabilitation for the motor complete SCI group, while ventral tissue bridges were significantly associated with discharge pinprick scores (n = 56, r = 0.39, P = 0.003), and dorsal tissue bridges were significantly correlated with discharge light touch scores (n = 56, r = 0.39, P = 0.002). In the motor incomplete SCI group, ventral tissue bridges exhibited a significant correlation with discharge pinprick scores (n = 80, r = 0.26, P = 0.023), while dorsal tissue bridges demonstrated a significant correlation with discharge light touch scores (n = 80, r = 0.26, P = 0.022).

Clinical implications of the study

The findings of the present study suggest that MRI biomarkers, such as tissue bridges, can be used to predict the likely prognosis of recovery following SCI. This information can aid clinicians and physiotherapists in personalizing rehabilitation protocols for individual patients.

Strength and limitations

An important aspect of this study is the thorough evaluation of different MRI parameters in relation to clinical and neurological outcomes. This study incorporates objective measurements, statistical significance, thorough analysis, clinical relevance, predictive values, and recommendations for future research. By leveraging these strengths, such a study can provide valuable insights into the prognosis and management of SCIs. However, this study has several limitations, including a small sample size, single-center design, potential measurement variability (a single blinded rater, and inter- or intra-rater reliability not formally assessed), absence of multivariate adjustments, heterogeneity in injury levels, and a lack of long-term follow-up beyond six months - all of which may impact the robustness and generalizability of the statistical findings. Conducting multi-centre studies with larger cohorts, standardized measurement protocols (possibly automated segmentation), and inclusion of advanced qMRI techniques (e.g., diffusion tensor imaging, magnetization transfer) can further elucidate microstructural predictors of neurological prognosis post-SCI. Furthermore, these MRI biomarkers can be combined with electrophysiological assessments for multimodal prognostication, or longitudinal tracking beyond six months to examine sustained recovery patterns.

Additional factors such as the age of the patient, their overall health, and the presence of other medical conditions can also have an impact on the final clinical and neurological outcome.

CONCLUSION

The MRI imaging biomarkers in SCI patients demonstrated substantial improvement over time. There was a negative correlation between neurological recovery, MSCC, MCC, and lesion dimensions (lesion length, lesion width, and lesion area). Higher canal compromise and lesion dimensions were associated with a poorer outcome. The evaluation of the midsagittal tissue bridge (including the ventral tissue bridge, dorsal tissue bridge, total width of the tissue bridge, and tissue bridge ratios) at 1 and 6 months showed a positive correlation with the neurological recovery. The current study underscores the potential of MRI imaging biomarkers to predict neurological recovery in SCI, providing confidence for their use in clinical practice and research. Further research is necessary to confirm these findings and investigate their clinical implications for improving patient care and rehabilitation strategies in the management of SCIs.