Comparison of post-stroke white matter assessment using disconnectome-symptom mapping versus quantitative diffusion MRI

Indirect structural disconnection-symptom mapping allows white matter impairment to be determined without the need for multi-directional diffusion (MDDW) imaging for each individual. Although widely used this method has not been validated. We analyzed a multicenter dataset obtained from 166 individuals in the chronic stage after stroke with upper limb impairment quantified with Fugl-Meyer upper extremity score (FMUE) comprising stroke lesion maps and MDDW imaging. White matter integrity was quantified (1) by diffusion-tensor-imaging-based fractional anisotropy in preselected tracts (fractional anisotropy method; FAM) and (2) by extracting a percentage of tract disconnection by masking each tract of a predefined tractography atlas using the individual map (disconnection-symptom mapping; DSM). We also calculated a lateralization index for the fractional anisotropy between both hemispheres. The following tracts were tested: corticospinal tract (CST), superior lateral fasciculus (SLF) and corpus callosum (CC) but also optic radiation (OR) as a control tract. Both methods (FAM, DSM) showed comparable results for the association of white matter integrity of the CST with FMUE. DSM showed a strong association with FMUE likely because of the number of participants who failed to show an overlap of the tracts and lesion masks (for CST: n = 57 out of 166; for CC: n = 103 out of 166) whereas with FAM these participants could be used for further analyses. On the first view, our data support the use of white matter integrity quantification based on DSM in individuals with chronic stroke. However, at least one-third-of cases (for CC even worse) showed no overlap of lesion and tract resulting in artificially high associations with clinical parameters.