Accuracy of FDG-PET brain scan to differentiate sporadic bvFTD of mild severity from late onset psychiatric disorders

Background: Positron emission tomography with [18F] fluorodeoxyglucose (FDG-PET) brain scan accurately differentiates behavioral variant frontotemporal dementia (bvFTD) from other neurodegenerative disorders. However, the accuracy of FDG-PET to differentiate bvFTD from primary psychiatric disorders (PPD) remains understudied, with prior studies reporting high false positive rates. Methods: Retrospective data from the DIagnostic and Prognostic Precision Algorithm for behavioral variant Frontotemporal Dementia (DIPPA-FTD) database included 508 sporadic bvFTD and 152 late-onset PPD cases (onset at age 45–75 years). 226 cases had an FDG-PET scan completed at baseline with a visual rating for bvFTD diagnosis (Yes, No, or Ambiguous). Follow-up clinical diagnosis at least 1 year later was used as the gold standard to calculate sensitivity and specificity of FDG-PET at baseline. Results: Follow up consisted of 198 probable bvFTD cases (87.6%) and 28 PPD cases (12.4%). FDG-PET had a sensitivity of 87% and a specificity of 93% to identify bvFTD cases, with a diagnostic accuracy of 88%. Within the PPD group, 7.1% had false positives and 14.3% had ambiguous scans. Rates of false negatives (normal or ambiguous PET) in bvFTD were 13.1% and were 2.6 times more common in males. Conclusion: FDG-PET had excellent diagnostic accuracy to differentiate between sporadic bvFTD and late onset PPD, with a low true false positive rate in PPD. False negatives were more common in males with bvFTD. These results support the use of FDG-PET as an important step in the diagnostic algorithm for patients with late onset behavioral changes, although results can be misleading in some cases.