Objective: We hypothesized that vascular amyloid contributes to chronic brain ischemia, therefore amyloid burden measured by Pittsburgh compound B retention on positron emission tomography (PiB PET) would correlate with the extent of magnetic resonance imaging (MRI) white matter hyperintensities (WMH; or leukoaraiosis) in patients with high vascular amyloid deposition (cerebral amyloid angiopathy [CAA]) but not in patients with high parenchymal amyloid deposition (Alzheimer disease [AD]; mild cognitive impairment [MCI]) or in healthy elderly (HE) subjects. Methods: Forty-two nondemented CAA patients, 50 HE subjects, and 43 AD/MCI patients had brain MRI and PiB PET. Multivariate linear regression was used to assess the independent association between PiB retention and white matter disease volume, controlling for age, gender, apolipoprotein E genotype, and vascular risk factors within each group. Results CAA patients were younger than HE and AD subjects (68 ± 10 vs 73.3 ± 7 and 74 ± 7.4, p < 0.01) but had higher amounts of WMH (median = 21 vs 3.2 and 10.8 ml, respectively, p < 0.05 for both comparisons). Global PiB retention and WMH showed strong correlation (rho = 0.52, p < 0.001) in the CAA group but not in HE or AD. These associations did not change in the multivariate models. Lobar microbleed count, another marker of CAA severity, also remained as an independent predictor of WMH volume. Interpretation: Our results indicate that amyloid burden in CAA subjects (with primarily vascular amyloid) but not AD subjects (with primarily parenchymal amyloid) independently correlates with WMH volume. These findings support the idea that vascular amyloid burden directly contributes to chronic cerebral ischemia and highlights the possible utility of amyloid imaging as a marker of CAA severity.