O4‐09‐02: Diagnostic value of molecular imaging in a memory clinic setting

Abstract

[11C]Pittsburgh compound-B (PIB) and [18F]fluorodeoxyglucose (FDG) PET measure fibrillar Aß load and glucose metabolism respectively. We assessed the diagnostic value of these tracers in an unselected memory clinic population. Paired dynamic (90 minutes) PIB and static FDG (15 minutes) PET scans were performed in 154 patients, shortly after completing a standard dementia work-up in a memory clinic. PET images were visually assessed by a nuclear medicine physician and results were reported to the cognitive neurologists. Outcome measures were (change in) clinical diagnosis and confidence in that diagnosis after disclosing PET results. At baseline, patients received the following diagnoses: Alzheimer's disease (AD, n = 66), mild cognitive impairment (MCI, n = 30), subjective memory complaints (SMC, n = 15), frontotemporal dementia (FTD, n = 18), dementia with Lewy bodies (DLB, n = 5), other forms of dementia (n = 10), and other psychiatric (n = 6) or neurological diseases (n = 4). PIB scans were positive in 40/66 patients with clinical AD (61%), 5/18 patients with clinical FTD (28%), 4/5 patients with DLB (80%), and 3/10 patients with other dementias (30%). FDG uptake patterns matched the clinical diagnosis in 58% of patients with clinical AD, and in 33% of patients with clinical FTD. PET results led to a change in diagnosis in 35 (23%) patients. A change in clinical diagnosis only occurred when diagnostic certainty was lower than 90% prior to PET. Furthermore, percentage diagnostic alterations after PET increased with decreasing pre-PET diagnostic confidence. Across groups, diagnostic confidence increased from 71±17% before to 87±16% after PET (P<0.001). Two year clinical follow-up in a subsample (n = 39) showed that PIB and FDG predicted progression to AD for patients with MCI, and that the diagnosis of dementia established after PET remained unchanged in 96% of the patients.