Reliability of Diffusion-Tensor Imaging Measurement Parameters in Cervical Spondylotic Myelopathy

Abstract

Introduction: The goal of this study was to assess inter and intra-observer reliability in cervical spinal cord diffusion-tensor imaging (DTI) in cases of cervical spondylotic myelopathy (CSM), as well as agreement among three different ROI measurement methods. Methods: Institutional Review Board approval and informed consent were obtained for this study. From July of 2013 until December of 2013, 34 patients (12 male, 22 female

mean age 58.7 years, range 45-79 years) who underwent surgical decompression due to CSM with pre-operative DTIs were retrospectively enrolled. Four observers independently measured the fractional anisotropy (FA) values twice, using 3 different measurement methods, and a spine segment was selected for each disc level, from C1/2 to C7/T1. Inter and intra-observer reliability were assessed by interclass correlation coefficients (ICCs) and, for the test-retest reliability, the Cronbachs alpha was calculated for each observer. The differences in the mean FA values from C1/2 through C7/T1 of all of the study subjects were assessed using a paired sample T-test. Results: The overall agreement among the 4 observers varied according to the spinal cord level and measurement methods used, ranging from poor to excellent agreement (ICC = 0.374-0.821), with relatively lower agreements in the sagittal ROI method. The radiology resident and neuro-radiologist (observers 3 and 4) showed excellent intra-observer agreement in almost every spinal cord level (ICC = 0.887-0.997). However, the inter-observer agreements varied, with fair to good agreement (ICC = 0.404-0.747). Furthermore, there was a statistically significant difference in the mean FA values from C1/2 through C7/T1 in all three measurement methods. Conclusions: As a diagnostic tool, measured data from the DTI should have high reliability and reproducibility. In spite of the excellent intra-observer reliability of the ROI measurements, the FA values in the patients with CSM had a wide range of variation in inter-observer reliability in our study. Therefore, we should pay more attention to the interpretation of the DTI parameter for clinical usage.