An increase of blood proteins in the CSF is indicative of Alzheimer’s disease (AD), which may imply increased blood brain barrier (BBB) permeability, either through impaired active transport or vessel leakage. Evidence for impairment of BBB function in AD comes studies demonstrating the presence of serum proteins such as albumin and immunoglobulin in the cerebrospinal fluid (1), though the role of the BBB remains controversial (2). Magnetic resonance imaging (MRI) is used extensively to study and diagnose AD, and the use of dynamic contrast enhanced (DCE) MRI (3) can quantify the leakage of a gadolinium (Gd) chelate contrast agent from the blood vessels into the extracellular-extravascular space (EES). Existing BBB permeability data indicates that there is evidence of an age-related deficit that is further amplified in dementia groups such as AD (4,5). To evaluate if vessel leakage in AD could be detected using gadolinium (Gd) chelate based dynamic contrast enhanced (DCE) MRI.DCE 3T MRI data was acquired on a Siemens Verio (Siemens Healthcare, Erlangen Germany) using a 32-channel head coil from 6 subjects with previous diagnoses of AD, 2 subjects with vascular dementia (VaD), and 7 healthy controls (HC) group-matched for age and gender. After initial localizers, a variable flip angle FLASH acquisition used 8 RF flip angles (2°, 4°, 8°, 10°, 14°, 18°, 22°, 26°) with a resolution of 1.5x1.5x3mm3, TR=3.89ms, TE=1.2ms, BW=800Hz/pixel, and partial Fourier in both phase encoding directions. A dual temporal resolution DCE acquisition was employed using identical parameters except a flip angle of 14°, rapidly acquiring 42 volumes in the first 5 minutes of the contrast injection, and then one 4-NEX volume approximately every five minutes until at least 50 minutes had elapsed. Contrast injection was begun 15s into the acquisition, total scan time was 5m:04s. A saturated double-angle method (SDAM) sequence was used to map the transmit RF (B1) field (6). Data was acquired across five repeated alternating acquisitions at each of two flip angles, 60° and 120°. Images were averaged and smoothed with an isotropic Gaussian filter of 8mm width over the entire field of view prior to calculation of the B1 field. Structural T1, T2, FLAIR, and time-of-flight MRA scans were also acquired for radiological review to rule out other pathologies. The extended Tofts model was used after transforming B1-corrected T1 values into contrast agent concentration in order to generate Vp, Ktrans and kep maps. The T1w volume was segmented into 10 ROIs: caudate nucleus, thalamus, cerebellar white and grey, frontal cortex white and grey, corpus callosum, cerebro-spinal fluid, as well as automated definitions of global grey matter and white matter ROI. Voxels located within one internal carotid artery were used to define an input function for each scan. DCE-MRI mapping was successful in all 15 subjects, resulting in Ktrans, Vp, and kep maps of the whole brain. No inter-group effects were measured in any ROI. Analysis revealed a detectable concentration (7) range in both GM and WM of 0.006 to 3.0 mM.DCE-MRI did not detect passive leakage of the Gd chelate through the BBB. This implies either the leakage of the BBB is slower than the detectable limit, or that the permeability to a small (560 Dalton) molecule is not indicative of the permeability to blood proteins, perhaps due to an active transport mechanism. This experiment was designed to maximize the sensitivity towards small leakage values, but was unable to detect differences in quantified measures of BBB permeability. The VaD group was included as a positive control; the expected permeability difference may not have been detected as this group contained only 2 subjects.