Cerebral Blood Flow and Aβ-Amyloid Estimates by WARM Analysis of [11C]PiB Uptake Distinguish among and between Neurodegenerative Disorders and Aging

Anders B. Rodell, Graeme O’Keefe, Christopher C. Rowe, Victor L. Villemagne, Albert Gjedde

Background: We report results of the novel Washout Allometric Reference Method (WARM) that uses estimates of cerebral blood flow and amyloid load from the same [(11)C]Pittsburgh Compound B ([(11)C]PiB) retention maps in brain to distinguish between patients with different forms dementia, including Alzheimer's disease, and healthy volunteers. The method introduces two approaches to the identification of brain pathology related to amyloid accumulation, (1) a novel analysis of amyloid binding based on the late washout of the tracer from brain tissue, and (2) the simultaneous estimation of absolute cerebral blood flow indices (sCBF) from the early accumulation of the tracer in brain tissue. Objective: We tested the hypothesis that a change of cerebral blood flow is correlated with the degree of tracer [(11)C]PiB retention, reflecting dendritic spine pathology and consequent inhibition of brain energy metabolism and reduction of blood flow by neurovascular coupling in neurodegenerative disorders, including Alzheimer's disease. Methods: Previously reported images of [(11)C]PiB retention in brain of 29 subjects with cognitive impairment or dementia [16 Alzheimer's Disease (AD), eight subjects with dementia with Lewy bodies (DLB), five patients with frontotemporal lobar degeneration (FTLD), five patients with mild cognitive impairment, and 29 age-matched healthy control subjects (HC)], underwent analysis of PiB delivery and retention by means of WARM for quantitation of [(11)C]PiB's binding potentials (BPND) and correlated surrogate cerebral blood flow (sCBF) estimates, based on the [(11)C]PiB images, compared to estimates by conventional Standard Uptake Value Ratio (SUVR) of [(11)C]PiB retention with cerebellum gray matter as reference. Receiver Operating Characteristics (ROC) revealed the power of discrimination among estimates. Results: For AD, the discriminatory power of [(11)C]PiB binding potential (BPND) by WARM exceeded the power of SUVR that in turn exceeded the power of sCBF estimates. Differences of [(11)C]PiB binding and sCBF measures between AD and HC both were highly significant (p < 0.001). For all the dementia groups as a whole, sCBF estimates revealed the greatest discrimination between the patient and HC groups. WARM resolves a major issue of amyloid load quantification with [(11)C]PiB in human brain by determining absolute sCBF and amyloid load measures from the same images. The two parameter approach provides key discriminary information in AD for which [(11)C]PiB traditionally is used, as well as for the distinct flow deficits in FTLD, and the marked parietal and occipital lobe flow deficits in DLB. Conclusion: We conclude that WARM yields estimates of two important variables that together discriminate among patients with dementia, including AD, and healthy volunteers, with ROC that are superior to conventional methods of analysis. The distinction between estimates of flow and amyloid load from the same dynamic emission tomograms provides valuable pathogenetic information.