Reliability of quantitative magnetic resonance imaging methods in the assessment of spinal canal stenosis and cord compression in cervical myelopathy

Abstract

STUDY DESIGN.: Prospective, blinded reliability study of quantitative magnetic resonance imaging (MRI) measures in patients with cervical myelopathy. OBJECTIVE.: To assess the intra- and interobserver reliability of commonly used quantitative MRI measures such as transverse area (TA) of spinal cord, compression ratio (CR), maximum canal compromise (MCC), and maximum spinal cord compression (MSCC). SUMMARY OF BACKGROUND DATA.: There is no consensus on an optimal quantitative MRI method(s) in assessing canal stenosis and cord compression. METHODS.: Seven surgeons performed measurements on 17 digital MR images, on 4 separate occasions. The degree of stenosis was evaluated by measuring TA and CR on axial T2, MCC, and MSCC on midsagittal T1- and T2-weighted MRI sequences, respectively. Statistical analyses included repeated-measures analysis of variance and intraclass correlation coefficients (ICCs). RESULTS.: The mean ± SD for intraobserver ICC was 0.88 ± 0.1 for MCC, 0.76 ± 0.08 for MSCC, 0.92 ± 0.07 for TA, and 0.82 ± 0.13 for CR. In addition, the interobserver ICC was 0.75 ± 0.04 for MCC, 0.79 ± 0.09 for MSCC, 0.80 ± 0.05 for CR, and 0.86 ± 0.03 for TA. Higher degree of canal compromise (MCC) was associated with lower modified version of Japanese Orthopaedic Association Scale score (P = 0.05). Also, a strong association was found between MSCC and lower modified version of Japanese Orthopaedic Association Scale score, greater number of steps, and longer walking time (P < 0.05). CONCLUSION.: All 4 measurement techniques demonstrated a good to moderately high degree of intra- and interobserver reliability. Highest reliability was noted in the assessment of T2-weighted sequences and axial MRI. Our results show that the measurements of MCC, MSCC, and CR are sufficiently reliable and correlate well with clinical severity of cervical myelopathy. © 2013, Lippincott Williams & Wilkins.