A Dynamic MRI Diagnostic Biomarker for Cervical Myelopathy

In This Article

Abstract and Introduction

Abstract

Study Design: Multicenter prospective observational study.

Objective: Diffusion tensor imaging in flexion extension improves the diagnosis of degenerative cervical myelopathy (DCM). We aimed to provide an imaging biomarker for the detection of DCM.

Summary of Background Data: DCM is the most common form of spinal cord dysfunction in adults; however, imaging surveillance for myelopathy remains poorly characterized.

Patients and Methods: Symptomatic DCM patients were examined in maximum neck flexion-extension and neutral positions in a 3T-magnetic resonance imaging scanner and allocated to 2 groups: (1) Patients with visible intramedullary hyperintensity (IHIS) on T2-weighted imaging (IHIS+, n = 10); and (2) Patients without IHIS (IHIS–, n = 11). Range of motion, space available for the spinal cord, apparent diffusion coefficient (ADC), axial diffusivity (AD), radial diffusivity, and fractional anisotropy were measured and compared between the neck positions and between the groups as well as between control (C2/3) and pathologic segments.

Results: Significant differences between the control level (C2/3) and pathologic segments were appreciated for the IHIS+ group at neutral neck position in AD; at flexion in ADC and AD; and at neck extension in ADC, AD, and fractional anisotropy values. For the IHIS– group, significant differences between the control level (C2/3) and pathologic segments were found only for ADC values in neck extension. When comparing diffusion parameters between groups, radial diffusivity was significantly different in all 3 neck positions.

Conclusion: Significant increases in ADC values between the control and pathologic segments were found for both groups in neck extension only. This may serve as a diagnostic tool to identify early changes in the spinal cord related to myelopathy to indicate potentially reversible spinal cord injury and support the indication for surgery in select circumstances.

Introduction

Degenerative cervical myelopathy (DCM) is the most common form of spinal cord dysfunction in adults; however, imaging surveillance for myelopathy remains poorly characterized.^[1,2] Intramedullary hyperintensity (IHIS) on T2-weighted imaging reflects cord edema or myelomalacia, however, it only has a sensitivity of 60% for the detection of clinical myelopathy. The sensitivity can be increased to 80% when combined with the analysis of apparent diffusion coefficient (ADC) values.^[3] The sensitivity of T2-weighted imaging can be further improved by performing dynamic flexion-extension T2-weighted imaging.^[4]

The use of diffusion tensor imaging (DTI) allows for the evaluation of microstructural changes in the spinal cord not otherwise detected on routine conventional magnetic resonance imaging (MRI).^[5] Schatlo et al^[6] combined T2-weighted imaging with dynamic DTI and showed that increased ADC values in extension were the most sensible marker for suspected cervical spine stenosis based on clinical examination. This additional information can be of substantial value in cases where surgical intervention is being considered.

DTI enables estimates of diffusivity within each voxel. Fractional anisotropy (FA), ADC, radial diffusivity (RD), and axial diffusivity (AD) are calculated from the diffusion tensor. ADC represents the average diffusivity, regardless of direction, whereas FA refers to the fraction of diffusion that is directionally dependent. AD, also called longitudinal diffusion, applies to the diffusivity that corresponds with the direction of the fastest diffusion in each voxel. In the spinal cord, it aligns with the rostrocaudal orientation along the white matter tracts. RD, or transverse diffusivity, applies to the average diffusivity perpendicular to AD.^[5,7]

The use of DTI in neck flexion-extension allows the degree of stenosis and the space available for the cord to be assessed between flexion extensions. This provides information regarding possible cord compression, which reflects the diffusion parameters. Although DTI in neck flexion extension has been performed in healthy volunteers and DCM patients,^[6] we hypothesized that DTI is more sensitive for the detection of DCM. Here, we examined patients with clinical myelopathy symptoms and visible IHIS sign (IHIS+) versus patients with clinical myelopathy symptoms without IHIS sign (IHIS–), to see whether the MRI in neck extension-flexion combined with T2-weighted imaging and extended DTI parameters (ADC, AD, RD, and FA) can be used as an imaging biomarker for clinical myelopathy.

Table 1. Baseline Characteristics and Range of Motion
	IHIS+ (n = 10)	IHIS– (n = 11)
Age (yr), range	55.1 ± 11.2 (38–66)	58.7 ± 14.3 (37–81)
Sex (F); n, %	4 (40)	6 (55)
Weight (kg)	83.9 ± 14.9	79.2 ± 18.1
Height (cm)	169.6 ± 11.1	170.5 ± 10.8
BMI (kg/m²)	27.5± 5.3	27.4±5.1
Range of motion (°)	35.9 ± 5.9	37.8 ± 12.2
Angle normal position (°)	16.2 ± 4.2	14.8 ± 9.2
Angle flexion (°)	9.5 ± 3.5	10.1 ± 6.3
Angle extension (°)	25.3 ± 5.5	29.1 ± 7.3
mJOA score	15.5 ± 1.5	14.7±3.0
Muhle classification (0, 1, 2); n (%)*	0, 1 (43), 2 (57)	0 (27), 1 (55), 2 (18)

Values are presented as mean ± SD.
No statistically significant differences in the baseline characteristics were detected between the groups.
*Classification score: 0 = cord free; 1 = cord touched; 2 = cord compressed.
BMI indicates body mass index; IHIS, intramedullary hyperintensity; mJAO, modified Japanese Orthopedic Association.

Table 2. Diffusion Parameters of IHIS+ and IHIS– Patients
Neck Position	IHIS+ (n = 10)			IHIS– (n = 11)
Neck Position	Flexion	Neutral	Extension	Flexion	Neutral	Extension
AD (10^–6 mm²/s); C/23, control	1762 ± 195	2036 ± 211	1969 ± 372	1822 ± 310	2206 ± 547	1966 ± 376
AD (10^–6 mm²/s); pathologic level	2076 ± 418†	2433 ± 527*	2433 ± 712*	2004 ± 615	2427 ± 562	1996 ± 421
RD (10^–6 mm²/s); C/23, control	644 ± 136	727 ± 155§	782 ± 366	546 ± 167	542 ± 93§	642 ± 217
RD (10^–6 mm²/s); pathologic level	744 ± 143‡	714 ± 129‡	850 ± 176‡	607 ± 146‡	608 ± 100^‡	710 ± 147^‡
ADC (10^–6 mm²/s); C/23, control	1011 ± 139	1127 ± 80	124 ± 210	1054 ± 156	1090 ± 221	1066 ± 221
ADC (10^–6 mm²/s); pathologic level	1137 ± 195*	1303 ± 367	1353 ± 351*	1008 ± 316	1239 ± 306	1183 ± 129*
FA (10^–6 mm²/s); C/23, control	0.679 ± 0.101	0.596 ± 0.120	0.447 ± 0.052	0.682 ± 0.115	0.666 ± 0.085	0.707 ± 0.057
FA (10^–6 mm²/s); pathologic level	0.668 ± 0.095	0.648 ± 0.219	0.595 ± 0.073*	0.666 ± 0.113	0.661 ± 0.095	0.652 ± 0.089

*Statistically significant difference (P < 0.05) between affected and unaffected levels.
†Statistically significant difference (P < 0.01) between affected and unaffected levels.
‡Statistically significant difference (P < 0.05) between IHIS+ and IHIS– groups.
§Statistically significant difference (P < 0.01) between IHIS+ and IHIS– groups.
AD indicates axial diffusivity; ADC, apparent diffusion coefficient; FA, fractional anisotropy; IHIS, intramedullary hyperintensity; RD, radial diffusivity.