NFTs (600 nm) than SPs (510 nm) when using blue-violet excitation filter. Using multispectral microscopy, fluorescence signals of NFTs were com-pletely separated from those of SPs in AD brain section. Biodistribution study of THK-1188 showed sufficient brain uptake (4.8 %ID/g at 2 min post injection) in mice. Conclusions: Spectral imaging with THK-1188 can distinguish tau protein from A b in AD brain section. This technique would be usable for in vivo selective imaging of tau deposits in the brain. Background: P-glycoprotein (P-gp), an ABC efflux transporter encoded by the multidrug resistance 1 gene (MDR1; ABCB1), is highly expressed at the blood-brain barrier (BBB), where it acts as a gatekeeper to prevent entry of xenobiotics into the brain. Evidence from both animal and post-mortem hu-manstudiesimplicates P-gpasan importantamyloid-betaeffluxtransporter. In most Alzheimer’s Disease (AD) patients with late-onset disease, Abeta accumulates in the brainmainlybecause of reduced clearance,not increased production. We hypothesized that P-gp activity at the BBB is deficient in late-onset AD. If confirmed, therapeutic modulation of P-gp activity could represent a promising new target for delaying onset and progression of AD. Methods: Nine late onset AD cases and 9 age-matched controls were stud- ied using PET imaging to measure the activity of P-gp (11 C-verapamil) at theBBB and correctfor regional cerebral perfusion(rCBF,15 O-water).Sta- tistical parametric maps for 15 O-water and 11 C-verapamil were normalized to cerebellum (rCBFc and verapamil K1c). K 1, 11 C-verapamil tissue distribution clearance, is a function of rCBF and the tissue extraction ratio (K 1 ¼ rCBFER). To determine regional P-gp activity unconfounded by changes in rCBFc, we normalized K 1c to rCBFc and g enerated relative regional extraction ratio (ERc)-grouped statistical parametric maps (ratio of regional 11 C-verapamil K 1c and rCBFc). Randomise softwarewas used to perform non-parametric voxel by voxel unpaired unpaired 2-tailed t-test with multiple comparison correction, to compare parametric maps of AD subjectswiththoseof controls. Results: BilateralreductionsinP-gpactivity (p < 0.05) were found in parietotemporal and frontal lobes in patients with AD compared to controls, and activity was also decreased in the hippocam- pus and posterior cingulate. Many brain regions showed both reduced P-gp activity and rCBFc, but others showed disparate patterns, indicating that re- gional hypoperfusion was not always associated with reduced P-gp activity and suggesting that these measures report different aspects of AD. Conclu- sions: P-gp activity at the BBB is compromised in AD. Confirmation, in early stage disease, of a regional association of reduced P-gp activity with an increase in Abeta in brain and reduction in CSF would support its impor-tance for Abeta clearance and implicate it in the pathogenesis of AD. 67 on a 7T scanner at and 12, 30, and 44 (gadoteridol) co-registration and voxelwise evaluation of the Bloch equation for each variable TI dataset accounting for all RF and delays. IR-MPRAGE structural images were also acquired. A applied to R 1 maps to define a region of interest in the lateral ventricles superior to the fornix and free from any visual choroid plexus. PVH using the Fazekas CSF volume correlates inversely with baseline ventricular 1 H 2 O R 1 values (r ¼ consistent with a lower macromolecular content with ventricular dilatation. The expected age-related increase in CSF volume is also observed (r ¼ 0.74). Ventricular 1 H 2 O R 1 values increase at each time point after CR injection ( P and are significantly associated with PVH score (P ¼ 0.003; repeated measures ANOVA). Group (AD v CN) effects are not significant. Conclusions: Administration of CR increases 1 H 2 O R 1 in the ventricles of the elderly brain. In the context of a relatively intact blood-brain barrier, the temporal changes in R 1 likely reflect leakage of CR across the BCSFB (ca. 2 m M hr -1). The increased rate of R 1 change after CR administration in the pres- ence of PVHs PVHs and increased BCSFB permeability a Background: Definitive diagnosis of Alzheimer’s disease (AD) and non-AD tauopathies still relies on post-mortem examination. These diseases are often difficult to differentiate clinically due to overlapping phenotypes, especially at early stages. In vivo imaging with PETwill allow new insights into tau deposition, facilitating research into causes, diagnosis and treatment of tauopa- thies. We have characterized 18F-THK523 (THK523), a novel tau imaging ligand developedatTohoku UniversityinSendai, Japan, assessing its selectiv- ity and specificity for tau pathology both invitro and invivo. Methods: Apart from acute toxicity microdose evaluation and in vitro receptor binding screen providingdirectevidenceofalackofpharmacologicalactivityortoxicity,pre-clinical evaluation included in vitro binding studies, autoradiography (ARG) and histofluorescence (HF) analysis of human hippocampal sections, and in vivo microPET studiesintau and APP/PS1 transgenic mice. To date,9 human participants underwent bothTHK523 and PiB PET studies. Results: THK523 binds with higher affinity to recombinant tau compared with A b fibrils. ARG and HF analysis of human hippocampal sections demonstrated THK523 co-localized with immunoreactive tau pathology, but failed to highlight A b plaques. MicroPET studies revealed higher brain retention of THK523 in tau transgenic mice compared with their wild-type littermates or APP/PS1 mice. Initial human PET studies comparing THK523 and PiB have shown that THK523 does not bind to A b in AD. Despite a higher THK523 cortical retention, following the known regional distribution of tau deposits in AD, there were no strong distinctive visual features in the THK523 PET scans between AD, healthy controls and Semantic Dementia patients. Conclusions: Preliminary PET human studies suggest that despite fulfilling preclinical ligand criteria fortau,notbinding toAb and presenting higherandappropriate regional distribution in AD patients, the signal to noise ratio from THK523 is notsufficienttodistinguishclinicalgroupsbysimplevisualinspection.Further studies, assessing non-AD tauopathiessuchasProgressive SupranuclearPalsy and Corticobasal Syndrome, are underway to confirm these initial findings. Background: Due to well-known markedly increased susceptibility con- trast at ultra-high-field MR, this work is using 7T MR susceptibilty-weighted imaging (SWI) to better identify the histopathologic correlate of amyloid plaques containing iron and otherwise invisible subhippocampal structuresof humanpost-mortembraininpatientswithAlzheimer’sdisease. Methods: Post-mortem brain specimens of the frontal lobe and hippocam-puswereobtainedfrom8patients(meanage71.2 6 6.2years)withclinically diagnosed AD and 6 age-matched healthy controls (72.4 6 6.6 years) with-out AD. Coronal 1 w 3 cm thick brain slices were preserved and fixed in 2% agar Background: Alzheimer’s disease (AD) is one of the most common neuro- degenerative diseases. Neurofibrillary tangles (NFTs) are one of the patho-logical hallmarks found in AD brains. NFTs in the brain are tightly associated with the severity of dementia, indicating the contribution of NFTs to neuronal dysfunction. NFTs are a therapeutic target of AD for dis- ease modifying therapy. A PET tracer for imaging NFTs in the brain would be valuable in the development of new therapies for AD. Recently, [18