Within-participants reliability and measurement error of magnetization transfer imaging determinations within the healthy cervical spinal cord

Abstract

Purpose

To assess the within-participant reliability and measurement error in the determination of MTR in the healthy human cervical spinal cord.

Methods and materials

A total of twenty healthy controls (10 male, mean ± sd age: 33.9 ± 3.5 years, 10 females, mean ± sd age: 47.5 ± 14.4 years), with no family history of any neurological disorders or a contraindication to MRI scanning were recruited over a period of two months. Each participant was scanned twice with a 3T MRI scanner using standard MTI sequences. Spinal Cord Toolbox (v5.4) was used for image post-processing. Data were first segmented and then registered to a template and then MTR was computed. The within-participant coefficients of variation (CV%), single and average within-participants intraclass correlation coefficients (ICC) and Bland–Altman plots were determined for MT values over the volume between the 2nd and 5th cervical vertebrae for the total WM and for specific WM regions: dorsal column (DC), ventral column (VC) and lateral column (LC).

Results

MTR showed poor to excellent within-participant reliability for the total WM, DC, VC and LC with single/average ICC values of 0.03/0.06, 0.10/0.18, 0.39/0.75, and 0.001/0.002, respectively, and the CV% reported an acceptable variation with values less than 10%. The Bland–Altman plots showed good within-participant agreement between the scan-rescan values.

Conclusion

This study demonstrates that clinical trials using MTI technique are feasible and shows that quantitative MTI can monitor tissue changes in degenerative WM patients.

Implications for practice

MTI with its MTR index provide broad assessment of the integrity of white matter tissue and are being studied widely in brain as a diagnostic tool for the assessment of different neurological diseases.