Longitudinal relaxographic imaging of white matter hyperintensities in the elderly

Background Incidental white matter hyperintensities (WMHs) are common findings on T 2 -weighted magnetic resonance images of the aged brain and have been associated with cognitive decline. While a variety of pathogenic mechanisms have been proposed, the origin of WMHs and the extent to which lesions in the deep and periventricular white matter reflect distinct etiologies remains unclear. Our aim was to quantify the fractional blood volume (v b ) of small WMHs in vivo using a novel magnetic resonance imaging (MRI) approach and examine the contribution of blood–brain barrier disturbances to WMH formation in the deep and periventricular white matter. Methods Twenty-three elderly volunteers (aged 59–82 years) underwent 7 Tesla relaxographic imaging and fluid-attenuated inversion recovery (FLAIR) MRI. Maps of longitudinal relaxation rate constant (R 1 ) were prepared before contrast reagent (CR) injection and throughout CR washout. Voxelwise estimates of v b were determined by fitting temporal changes in R 1 values to a two-site model that incorporates the effects of transendothelial water exchange. Average v b values in deep and periventricular WMHs were determined after semi-automated segmentation of FLAIR images. Ventricular permeability was estimated from the change in CSF R 1 values during CR washout. Results In the absence of CR, the total water fraction in both deep and periventricular WMHs was increased compared to normal appearing white matter (NAWM). The v b of deep WMHs was 1.8 ± 0.6 mL/100 g and was significantly reduced compared to NAWM (2.4 ± 0.8 mL/100 g). In contrast, the v b of periventricular WMHs was unchanged compared to NAWM, decreased with ventricular volume and showed a positive association with ventricular permeability. Conclusions Hyperintensities in the deep WM appear to be driven by vascular compromise, while those in the periventricular WM are most likely the result of a compromised ependyma in which the small vessels remain relatively intact. These findings support varying contributions of blood–brain barrier and brain-CSF interface disturbances in the pathophysiology of deep and periventricular WMHs in the aged human brain.