Purpose There are no comprehensive guidelines for the use of FDG PET in the following three clinical scenarios: (1) diagnostic work-up of patients with idiopathic Parkinson’s disease (PD) at risk of future cognitive decline, (2) discriminating idiopathic PD from progressive supranuclear palsy, and (3) identifying the underlying neuropathology in corticobasal syndrome. Methods We therefore performed three literature searches and evaluated the selected studies for quality of design, risk of bias, inconsistency, imprecision, indirectness and effect size. Critical outcomes were the sensitivity, specificity, accuracy, positive/negative predictive value, area under the receiving operating characteristic curve, and positive/negative likelihood ratio of FDG PET in detecting the target condition. Using the Delphi method, a panel of seven experts voted for or against the use of FDG PET based on published evidence and expert opinion. Results Of 91 studies selected from the three literature searches, only four included an adequate quantitative assessment of the performance of FDG PET. The majority of studies lacked robust methodology due to lack of critical outcomes, inadequate gold standard and no head-to-head comparison with an appropriate reference standard. The panel recommended the use of FDG PET for all three clinical scenarios based on nonquantitative evidence of clinical utility. Conclusion Despite widespread use of FDG PET in clinical practice and extensive research, there is still very limited good quality evidence for the use of FDG PET. However, in the opinion of the majority of the panellists, FDG PET is a clinically useful imaging biomarker for idiopathic PD and atypical parkinsonism associated with dementia.

Brain β-amyloid is a hallmark of Alzheimer disease (AD)^1- 3 and is quantifiable in vivo through specific positron emission tomography (PET) radiotracers binding to β-amyloid with 96% to 97.8% sensitivity and 88.9% to 100% specificity.^4,5The F 18–labeled tracers (eg, [¹⁸F]-florbetapir, [¹⁸F]-flutemetamol, and [¹⁸F]-florbetaben) now overcome the limited clinical applicability of carbon 11–labeled Pittsburgh Compound B⁶ with their longer half-life (110 minutes). This evidence of analytical validity and feasibility makes amyloid PET a promising pathophysiologic marker for diagnosing AD according to most recent criteria.^7- 9 However, the clinical utility of amyloid PET is still under investigation worldwide. The presence of brain amyloidosis in patients with clinically diagnosed AD or lack thereof in individuals without AD is not warranted. In a recent meta-analysis,¹⁰ the prevalence of positive scans was 88% in patients with AD, 51% in patients with dementia with Lewy bodies (DLB), 30% in those with cerebrovascular disease (CVD), 12% in patients with frontotemporal lobar degeneration (FTLD), 38% in those with corticobasal degeneration (CBD), and 24% in healthy elderly individuals serving as controls. Consequently, amyloid PET, although approved for clinical use, is still not reimbursed by national health care systems. In the United States, a wide collection of clinical cases using amyloid PET, funded by private and public sources,¹¹ perform amyloid PET scans in cases consistent with the pertinent Appropriate Use Criteria,^12,13 collecting quantitative evidence on the incremental value of the examination for a large sample. To date, the published evidence on the added value of amyloid PET has important limitations.