Background: Diffusion tensor parameters can be analysed by fitting regions of interest (ROIs) to selected brain structures. The clinical usefulness of these measurements is influenced by their reproducibility and validity.
Objective: To investigate the reproducibility of fractional anisotropy (FA) and mean diffusivity (MD) measurements.
Material and methods: Seventy-six infants were imaged once at term-equivalent age. We measured several brain regions. Reproducibility was assessed using intraclass correlation coefficient and Bland-Altman method.
Results: Intra-observer reproducibility was excellent for FA in the calcarine cortex (right) and frontal white matter (left), and for MD in the corpus callosum (anterior), internal capsule, corona radiata, putamen, frontal white matter, optic radiation (left), thalamus (right) and calcarine cortex (right). Inter-observer reproducibility was excellent for FA in the corpus callosum (posterior) and for MD in the internal capsule and corona radiata (right). Inter-observer reproducibility was poor for FA in frontal and posterior white matter (right) and for MD in the inferior colliculus (right). Reproducibility was fair to good in other areas. The Bland-Altman plots showed no considerable bias, and variance was independent of the mean value.
Conclusion: Reproducibility of ROI measurement was fair to good for both FA and MD.