Related Experiment Video
Updated: Aug 24, 2025
Identification of Disease-related Spatial Covariance Patterns using Neuroimaging Data
Published on: June 26, 2013
Discordance and Concordance Between Cerebrospinal and [18F]FDG-PET Biomarkers in Assessing Atypical and Early-Onset
Kely Mónica Quispialaya1, Joseph Therriault1, Antonio Aliaga1
1From the Translational Neuroimaging Laboratory (K.M.Q., J.T., M.Z., J.F.-A., F.L., S.G., P.V., P.R.-N.), The McGill University Research Centre for Studies in Aging, Douglas Hospital, McGill University, Montreal, Quebec, Canada; Research Institute of the McGill University Health Centre (K.M.Q., A.A., B.J.-C., B.G.), Montreal, Quebec, Canada; Departments of Medicine (K.M.Q., B.J.-C., B.G.) and Neurology and Neurosurgery (J.T., M.Z., J.F.-A., F.L., J.-P.S., S.G., P.V., P.R.-N.), McGill University, Montreal, Quebec, Canada; Montreal Neurological Institute (J.T., M.Z., J.F.-A., F.L., G.M., J.-P.S., P.R.-N.), Montreal, Quebec, Canada; Department of Radiochemistry (G.M.), McGill University, Montreal, Quebec, Canada; Department of Psychiatry (T.P.), Pittsburgh University, PA; and Department of Psychiatry (S.G., P.R.-N.), McGill University, Montreal, Quebec, Canada.
Background And Objectives:
To assess the concordance and discordance between the core Alzheimer disease (AD) CSF biomarkers and [18F]fluorodeoxyglucose (FDG)-PET patterns evaluated clinically in memory clinic patients who meet appropriate use criteria for AD biomarker investigations.
Methods:
We retrospectively assessed participants with atypical and/or early-onset dementia evaluated at a tertiary care memory clinic. All individuals underwent CSF evaluations for Aβ42, phosphorylated tau (P-tau181) and total tau, and brain [18F]FDG-PET. [18F]FDG-PET data were visually interpreted by 2 nuclear medicine experts as being consistent with AD or non-AD. CSF biomarker results were similarly grouped into AD biomarker positive/negative. Contingency tables and Kappa coefficients were used to establish the level of agreement and disagreement between CSF and [18F]FDG-PET results in all individuals.
Results:
One hundred thirty-six individuals had both [18F]FDG-PET and lumbar puncture performed as part of the early-onset and/or atypical dementia assessments. [18F]FDG-PET showed a pattern suggestive of AD in 43% of patients, while CSF biomarkers showed results consistent with AD in 57% of participants. In patients who met criteria for AD biomarker investigations, we found that [18F]FDG-PET was discordant with CSF AD biomarkers in nearly 20% of cases; 12% of individuals with [18F]FDG-PET scans consistent with AD had AD-negative CSF results; and 7% of individuals with [18F]FDG-PET scans not consistent with AD had AD-positive CSF results, potentially suggesting atypical AD variants or less advanced neurodegeneration. [18F]FDG-PET discriminated patients with an AD-positive CSF profile from patients with an AD-negative profile with a sensitivity and specificity higher than 80% (sensitivity: 81%, 95% CI = 71-88%, SP: 81%, 95% CI = 68-89%). Furthermore, [18F]FDG-PET had a positive predictive value of 87% (95% CI = 78-93%) and a negative predictive value of 72% (95% CI = 60-82%).
Discussion:
CSF and [18F]FDG-PET disagreed in nearly 20% of the cases studied in this clinical series. While CSF Aβ42 and P-tau181 biomarkers are specific for AD, the topographical information from [18F]FDG-PET may provide complementary information.
