Disruption of White Matter Integrity in Adult Survivors of Childhood Brain Tumors: Correlates with Long-Term Intellectual Outcomes

Abstract

Background: Although chemotherapy and radiation treatment have contributed to increased survivorship, treatment-induced brain injury has been a concern when examining long-term intellectual outcomes of survivors. Specifically, disruption of brain white matter integrity and its relationship to intellectual outcomes in adult survivors of childhood brain tumors needs to be better understood.

Methods: Fifty-four participants underwent diffusion tensor imaging in addition to structural MRI and an intelligence test (IQ). Voxel-wise group comparisons of fractional anisotropy calculated from DTI data were performed using Tract Based Spatial Statistics (TBSS) on 27 survivors (14 treated with radiation with and without chemotherapy and 13 treated without radiation treatment on average over 13 years since diagnosis) and 27 healthy comparison participants. Whole brain white matter fractional anisotropy (FA) differences were explored between each group. The relationships between IQ and FA in the regions where statistically lower FA values were found in survivors were examined, as well as the role of cumulative neurological factors.

Results: The group of survivors treated with radiation with and without chemotherapy had lower IQ relative to the group of survivors without radiation treatment and the healthy comparison group. TBSS identified white matter regions with significantly different mean fractional anisotropy between the three different groups. A lower level of white matter integrity was found in the radiation with or without chemotherapy treated group compared to the group without radiation treatment and also the healthy control group. The group without radiation treatment had a lower mean FA relative to healthy controls. The white matter disruption of the radiation with or without chemotherapy treated survivors was positively correlated with IQ and cumulative neurological factors.

Conclusions: Lower long-term intellectual outcomes of childhood brain tumor survivors are associated with lower white matter integrity. Radiation and adjunct chemotherapy treatment may play a role in greater white matter disruption. The relationships between white matter integrity and IQ, as well as cumulative neurological risk factors exist in young adult survivors of childhood brain tumors.

Fig 1. White matter differences between survivors and healthy controls.

(A) Significant white matter differences between survivors (in both RT and NRT groups) and healthy controls (HC) were found in the empirically-identified white matter regions from TBSS. White matter skeleton (color coded in green) is overlaid on a T₁ weighted image. Clusters of significantly lower fractional anisotropy (FA) for survivor group are in orange and red. (B) The plot of statistically significant correlations between intellectual performance and the white matter FA measured from the areas of left frontal pole (red), right frontal pole (blue) and corpus callosum (green). The open symbols represent survivors with radiation treatment and solid symbol represent those with no radiation treatment. LFP = left frontal pole (red), RFP = right frontal pole (blue), CC = corpus callosum (green). VIQ = verbal intelligence quotient, PIQ = performance intelligence quotient. a.u. = arbitrary units.

Fig 2. White matter differences between survivors treated with radiation therapy with or without chemotherapy and healthy controls.

(A) Significant white matter differences between survivors with radiation treatment with or without chemotherapy (RT) and healthy controls (HC) were found in the empirically-identified white matter regions from TBSS. White matter skeleton (color coded in green) is overlaid on a T₁ weighted image. Clusters of significantly lower fractional anisotropy (FA) for survivor group are in orange and red. (B) The plot of statistically significant correlations between intellectual performance and the white matter FA measured from the areas of left middle frontal (red) and left middle temporal (blue). LMF = left middle frontal (red), LMT = left middle temporal (blue). VIQ = verbal intelligence quotient, PIQ = performance intelligence quotient. a.u. = arbitrary units.

Fig 3. White matter differences between survivors with no radiation treatment and healthy controls.

Significant white matter differences between survivors with no radiation treatment (NRT) and healthy controls (HC) were found in the empirically-identified white matter regions from TBSS. White matter skeleton (color coded in green) is overlaid on a T₁ weighted image. Clusters of significantly lower fractional anisotropy (FA) for survivor group are in orange and red. No statistically significant correlation between intellectual performance and the white matter FA was found in these areas.

Fig 4. White matter differences between survivors with radiation treatment with or without chemotherapy (RT) and survivors without radiation treatment (NRT).

(A) Significant white matter differences were identified between the RT group and the NRT group in the empirically-identified white matter regions from TBSS. White matter skeleton (color coded in green) is overlaid on a T₁ weighted image. Clusters of significantly lower fractional anisotropy (FA) for the RT survivor group are in orange and red. (B) The plot of statistically significant correlations between intellectual performance and the white matter FA measured from the areas of anterior portion of corpus callosum (green), right middle temporal (red) and right middle frontal (blue) regions. CC = corpus callosum (green), RMF = right middle frontal (blue), RMT = right middle temporal (red). VIQ = verbal intelligence quotient, PIQ = performance intelligence quotient, a.u = arbitrary units.