Reproducibility and reliability of MR measurements in white matter: clinical implications.

The purpose of this study was to determine the reproducibility and reliability of five MRI-derived measurements, namely, total water content (WC), myelin water content (MWC), mean T2 relaxation time (GMT2), T1 relaxation time (T1) and magnetization transfer ratio (MTR). Five controls were scanned 5 times over 1 year. The five MR measurements were made for 5 white matter regions. All measurements were found to be highly reproducible. MTR had a low reliability coefficient because all individual values were similar. Therefore, MTR would be most sensitive in detecting changes from normal. WC had a high reliability coefficient in all regions. For MWC, GMT2 and T1, the overall reliability coefficients were high but for some individual regions were low. The high coefficients suggest that these measurements, although different between normal subjects, are consistent over time. They could be used to explore natural differences in the normal population, but due to the large spread in normal values, larger sample sizes are needed to detect pathological changes.