To exclude or not to exclude: White matter hyperintensities in diffusion tensor imaging research

Objective: A practical methodological issue for diffusion tensor imaging (DTI) researchers is determining what to do about incidental findings, such as white matter hyperintensities (WMHI). The purpose of this study was to compare healthy control subjects with or without WMHIs on whole brain DTI. Method: Participants were 30 subjects (age=37.7, SD=11.3, Range=18-60; 70% female) who had no known developmental, general medical, neurological or psychiatric condition that could have had an adverse affect on brain morphology. Results: MRI (3 Tesla) revealed, at minimum, a WMHI in eight subjects (26.7%). Fractional anisotropy (FA) was calculated for 19 regions of interest (ROI). Frequency distributions of FA scores for the 19 ROIs were calculated. The 10th percentile for each ROI was selected as a cut-off score. Having four or more low FA scores occurred in 16.7%. More subjects with incidental findings met criterion for low FA scores (37.5%), compared to 9.1% of subjects with no findings. When subjects with minor WMHIs were retained and only those with multiple incidental findings were excluded, 8.3% of the retained subjects met criterion for low FA scores compared to 50.0% of the excluded subjects. Conclusions: The decision to include or exclude subjects who have incidental findings can influence the results of a study.