Relationship between spectral analysis, SUV and SUV Pons ratio as a measure of cerebral glucose metabolic rate in Alzheimer's disease

Paul Edison,Ajayverma Macharouthu,Zuzana Walker,Clive Ballard,Grazia Daniela Femminella,Joseph Nowell,Aparna Prasanna,Gregor Russell,Simon Thacker,Kehinde Junaid,Hilary Archer,Elizabeth Coulthard,Carol Bannister,Salman Karim,Sanara Raza,George Tadros,Craig W Ritchie,Clive Holmes,Lucy Knight,Vandana Mate,Paul Koranteng,Nicholas R Livingston,Stefan Carver,Andrew Donaldson,Anthony Peter Passmore,Sajeev Kshemendran,Bernadette Mcguinness,Naghma Malik,Brady Mcfarlane,Ramin Nilforooshan,Eleanor G Blunt,Ben Underwood,Basil H Ridha ,John Harrison ,Robert M Lawrence ,David J Brooks

doi:10.1002/alz.046068

Abstract

AbstractBackgroundCerebral metabolic rate of glucose (CMRglc) can be measured with fluorodeoxyglucose (FDG) positron emission tomography (PET) using arterial blood sampling. However, blood sampling is an invasive methodology for subjects. Standard uptake value (SUV) is a non‐invasive method to assess the cerebral glucose metabolism. Here we evaluate the relationship between semi‐quantitative measurements such as SUV and Pons ratio with a quantitative measurement such as Spectral analysis (SA) in order to assess their surrogate use in clinical trials.MethodsTo assess cerebral glucose metabolism by a non‐invasive and simplified method, the standardized uptake value (SUV) and used the Pons RATIO (where target region was divided by pons as a reference region, which requires no arterial input function or blood sampling. The SA was performed using blood sampling to create the input function. The study participants included 63 subjects (mean ± SD age, 71 ± 8 years) from the Evaluating Liraglutide in Alzheimer’s Disease (ELAD) trial. Regional CMRglucose (CMRglc) and regional SUV were measured using FDG PET, and the correlation between CMRglc and SUV together with CMRglc and Pons ratio were estimated for the temporal lobe, medial temporal lobe and hippocampusResultsIn this AD population, the SA and SUV in the temporal lobe (r=0.635, p= >0.001), medial temporal lobe (0.533, p= >0.001) and hippocampus (r=0.539, p= >0.001) had a positive correlation and were statistically significant. The correlation coefficient between SA and Pons ratio for the temporal lobe (r=0.166, p= 0.194), medial temporal lobe (r=0.146, p=0.146) and the hippocampus (r=0.193, p=0.129) was positive but not statistically significant.ConclusionThe SUV correlated better with cerebral glucose metabolic rate using SA than Pons RATIO in AD subjects, while Pons RATIO did not perform well. Therefore, SUV analysis maybe more reliable analysis method compared to Pons RATIO for FDG PET in clinical trials in AD subjects.

Full Text