Brachial plexus MR imaging: accuracy and reproducibility of DTI-derived measurements and fibre tractography at 3.0-T

To estimate intrastudy, intraobserver and interobserver reproducibility of DTI-derived measurements and fibre tractography (FT) at 3.0 T MR imaging in subjects without known brachial plexus pathology. IRB approval and written informed consent were obtained. Forty healthy volunteers underwent bilateral 3.0-T DTI of the brachial plexus. Postprocessing included FT and analysis of fractional anisotropy (FA) and apparent diffusion coefficient (ADC). Four authors performed postprocessing and analysis independently and in different sessions at baseline and after 4 weeks. Non-parametric tests and Bland-Altman statistics were used. Minimum and maximum percent variability were 6% and 20% for FA (85%-93% reproducibility). For ADC minimum and maximum percent variability were 6% and 18% (86%-97% reproducibility). Quality of fibre tract was rated equal in 80% and slightly different in 20% of subjects. Minimum detectable differences between limb were 37% for FA and 32% for ADC. Intra- and inter-observer agreement were good. Evaluating the combined influence of the observer and of the repeated measurements the reproducibility was 81-92%. DTI of brachial plexus nerves is reliable. The healthy contralateral side can be used as an internal control considering that changes in FA and ADC values of less that 37% and 32% will not be clinically detectable with confidence.