Correlation between magnetic resonance imaging and radiographic measurement of cervical spine in cervical myelopathic patients

Journal of Orthopaedic Surgery, Dec 2004 by Wong, T M, Leung, H B, Wong, W C

ABSTRACT

Purpose. To correlate the radiographic measurement, cord diameter shown on magnetic resonance imaging (MRI), and clinical hand sign of cervical myelopathic patients.

Methods. Patients with clinical cervical myelopathy who had had MRI in Kwong Wah Hospital between January 2001 and December 2002 were enlisted. Their cervical spine radiographs and clinical records were reviewed.

Results. Of 36 patients with a complete set of MRI films, cervical spine radiographs, and clinical notes; 18% did not have Hoffman's sign, 47% had normal supinator reflex, 39% had unimpaired 10-second test, and 45% showed no finger escape sign. The presence of myelopathic hand signs was not correlated to any radiological assessment, cord diameter, or presence of myelomalacia at any level.

Conclusion. Cervical spine radiography cannot predict the level and degree of cervical spinal cord compression. Myelopathic hand signs are not diagnostically fail-safe and cannot predict the level and degree of cord compression.

Key words: magnetic resonance imaging; spinal cord diseases; technology, radiologic

INTRODUCTION

Underreported in the western literature, cervical myelopathy is a common disease within the Asian populations. Pathogenesis includes congenital cervical spinal canal stenosis,1 prolapsed inter-vertebral disc, impinging osteophyte, ossified posterior longitudinal ligament, hypertrophic ligamentum flavum, and dynamic instability.

Although soft tissue is often the culprit of the compression, only the bony boundary can be depicted by standard cervical spine radiography. Since Pavlov et al.2 popularised measurement of cervical spine radiograph, a number of studies correlating the clinical condition to the radiographic measurement have been performed. Nonetheless, questions still arise concerning whether lateral cervical spine radiographs offer adequate sensitivity and specificity to evaluate the extent and location of cervical canal encroachment.3

MATERIALS AND METHODS

From January 2001 to December 2002, patients with clinical cervical myelopathy who had been referred to have magnetic resonance imaging (MRI) in Kwong Wah Hospital were enlisted. Their cervical spine radiographs and clinical records were retrieved. The cervical spine radiographs were taken with the patients in a eye-level position. There was no predetermined focus-to-film or object-to-film distance. Therefore, the magnification ratio cannot be determined. All the MRI films had a scale printed on the margin to allow translating the measurement into actual size.

A transparent film with concentric circles in 2 mm increments was used to measure the radiograph. The spinal canal size was measured from C-3 to C-7. Three readings were taken on each vertebral level, namely, upper-end plate, the cephalocaudal midpoint of the vertebra, and lower-end plate. When the ossified posterior longitudinal ligament was visualised, it was taken as the anterior boundary of the spinal canal instead of the posterior cortex or osteophyte. The nearest point of the spinolaminar line was then read with the transparent film. This distance was referred as the sagittal spinal canal diameter (Fig. 1).

The anteroposterior diameter of the spinal cord was recorded from MRI films at the cephalocaudal midpoint of the vertebrae (C-3 to C-7) and at the disc level (C3/4 to C6/7). Clinical information was recorded including demographic data, presence of Hoffman's sign, 10-second test, finger escape sign, and reverse supinator test.

Statistical analysis was performed using Statistical Package for the Social Sciences (Windows version 11.0; SPSS Inc., Chicago [IL], US). Pearson correlation tests, Chi squared tests, and independent-samples t tests were used. All tests were 2-tailed with a significance level set at 0.05.

RESULTS

Of 62 patients who had their cervical spine examined using MRI, only 36 had the complete set of MRI films, cervical spine radiographs and, clinical notes that can be traced. Their mean age at the time MRI was taken was 63.9 years (range, 34-85 years). 20 of the 36 patients were males. Females had a higher mean age than males (68.2 years for females vs 61.2 years for males; p=0.036).

The descriptive statistics on the radiological and MRI measurement are shown in the Table. Of the 36 myelopathic patients, 18% did not have Hoffman's sign, 47% had normal supinator reflex, 39% had unimpaired 10-second test, and 45% showed no finger escape sign. The presence of myelopathic hand signs was not correlated to any radiological assessment, cord diameter, or presence of myelomalacia at any level of the spine cord.

Age and sex were not found to relate to any radiological measurement of spinal canal size nor cord diameter measured by MRI.

At any level, cord diameter was not correlated with sagittal spinal canal diameter. Using the Pavlov ratio did not enhance the correlation. When the sagittal diameter at the upper-end and lower-end plate level was divided by the size of vertebral body using an analogue of the Pavlov ratio, the ratio still did not offer additional association. Using 0.5 cm as a cut-off value for spinal cord diameter between the affected and unaffected groups, the Pavlov ratio did not offer improved predictive value compared to simple measurement. Receiver operating characteristic curve shown in Fig. 2 exemplified the situation.