White matter integrity is associated with cognitive processing in patients treated for a posterior fossa brain tumor

Abstract

Children treated for posterior fossa tumors experience reduced cognitive processing speed and, after imaging, show damage to white matter (WM) tracts in the brain. This study explores relationships between white matter microstructure, assessed by fractional anisotropy (FA), and speed of cognitive processing using tract-based spatial statistics (TBSS). At 36 months after treatment with radiotherapy and chemotherapy, 40 patients completed an MRI examination and neuropsychological evaluation. Patients were matched with healthy control subjects based on age, sex, and race. Individual FA values were extracted from examinations for all voxels identified as having significant association between processing speed and FA using TBSS. The regions were labeled anatomically, and fiber tracts were grouped into larger fiber bundle categories based on their anatomical and functional associations. Analyses were performed between mean skeletal FA values in each of the fiber bundles and each of the cognitive processing scores controlling for age. Children 3 years after treatment for posterior fossa brain tumors demonstrate significantly lower processing speed associated with decreased FA, compared with their healthy peers. Commissural fibers in the corpus callosum were negatively affected by disease and therapy with detrimental consequence on patients' cognitive processing. Diffusion tensor imaging of the white matter tracts in the brain is relevant to determining potential mechanisms underlying clinically meaningful change in cognitive performance. Neuroprotective strategies are needed to preserve critical functions.

Fig. 1.

Fractional anisotropy (FA) skeleton mask (green). Top: Voxels where FA is significantly higher (P < .01) in controls when compared to patients (red). Middle: Voxels where FA is significantly associated with overall processing speed (red). Bottom: Voxels that are significantly different between patients and controls in their association with processing speed (red).

Fig. 2.

Proposed conceptual model of the relationship between age at the time of diagnosis, fractional anisotropy (FA) of the corpus callosum, and overall processing speed. R² values indicate the amount of variance in the outcome variable accounted for by the predictor variable. The model suggests that impact of age at diagnosis on overall processing speed may be more directly explained by variations in corpus callosum FA.