Discrepant diagnostic classification with AV45-PET and FDG-PET of clinically diagnosed Alzheimer's patients in the Alzheimer's Disease Neuroimaging Initiative

norman l foster,richard king,angela wang,susan m landau,william j jagust,kewei chen,eric m reiman

DOI: https://doi.org/10.1016/j.jalz.2012.05.282

2012-01-01

Alzheimer s & Dementia

Abstract:In about 20% of clinically diagnosed Alzheimer's disease (AD) subjects in the Alzheimer's Disease Neuroimaging Initiative (ADNI) FDG-PET scans are not AD-like as are a similar proportion of amyloid PET scans. Recent NIA-AA diagnostic criteria consider biomarkers “uninformative” when results disagree. 70 ADNI subjects (43 men, 27 women, age: 75.1 ± 8.2 yrs, MMSE: 21.9 ± 4.2) with a clinical diagnosis of AD between 4/07–10/11 had both FDG-PET and amyloid scans. We classified FDG-PET scans as either AD-like or not AD-like using our previously published criteria based on an independent visual interpretation of Neurostat 3D-SSP metabolic and statistical maps and then consensus of two experienced raters. We classified amyloid PET (PIB-PET and AV45-PET) as either AD-like or not AD-like based on a quantitative evaluation of the ratio of cerebral cortical to cerebellar gray matter binding (SUVR) using a previously published threshold of 1.47 for PIB-PET which can be converted to a value of 1.28 for AV45-PET. FDG-PET and amyloid PET results agreed in 56/70 (80%) of subjects. In 49/70 (70%) both biomarkers were consistent with AD, but in 7/70 (10%) both biomarkers were not AD-like, thus by NIA-AA criteria biomarker information considered “dementia unlikely due to AD”. Biomarker results conflicted in 14/70 (20%) of subjects and thus would be considered “uninformative”. Different thresholds would not alter these results. In 8/14 (57%) cases FDG-PET was AD-like while the amyloid PET did not meet threshold. In 6/14 (43%) cases amyloid PET was consistent with AD, while FDG-PET results were indeterminate because they lacked any significant hypometabolism (3), ambiguous because metabolic changes were extensive and most prominent in frontotemporal areas (2), and limited to hypometabolism in the left anterior temporal lobe (1). Although most ADNI AD subjects with FDG-PET and amyloid PET have a high probability of AD etiology, a sizeable proportion does not. Imaging could improve diagnostic accuracy in clinical trials. Multiple biomarkers often may lead to conflicting results unexpected from theoretical disease constructs. A hierarchy for AD biomarkers should be considered based upon their ability to predict a person's clinical course and postmortem neuropathology.