DOI: 10.1002/alz.081729 ISSN: 1552-5260

Association of FDG‐PET with co‐pathologies in autopsy‐proven AD

Marianne Chapleau, Nidhi Mundada, Salvatore Spina, William W. Seeley, Lea T. Grinberg, Daniel Schonhaut, Howard J. Rosen, Joel H. Kramer, Bruce L Miller, Renaud La Joie, Gil D. Rabinovici
  • Psychiatry and Mental health
  • Cellular and Molecular Neuroscience
  • Geriatrics and Gerontology
  • Neurology (clinical)
  • Developmental Neuroscience
  • Health Policy
  • Epidemiology



The presence of multiple pathologies is the rule and not the exception in Alzheimer’s disease (AD). The goal of this study was to assess whether 18F fluorodeoxyglucose (FDG) PET can provide information about the presence of non‐AD pathologies in autopsy‐proven AD patients.


Our cohort included 55 patients with antemortem FDG‐PET and autopsy‐proven intermediate/high ADNC from the UCSF Neurodegenerative Diseases Brain Bank. All patients underwent FDG‐PET. Standardized uptake value ratio (SUVR) images were computed using the pons as reference region. W‐maps (adjusted Z‐score maps) were created, comparing each patient to a group of 187 controls (age range: 20‐96, 40% males). Voxel‐wise correlations were computed to verify the unique contribution of each neuropathology to cerebral hypometabolism, while controlling for age, sex, time between FDG‐PET scan and death, scanner type, Thal and Braak staging, and the presence of all other co‐pathologies. Associations between SUVR values in specific region of interests (ROIs) were also assessed. The ROIs included were the following: Landau AD ROI for AD pathology, Cingulate Island Sign (CIS) for Lewy body disease, Inferior/Medial Temporal Lobe (IMT) and IMT/Frontal Supraorbital (IMT/FSO) ratios for limbic‐predominant age‐related TDP43 encephalopathy neuropathologic change (LATE‐NC).


Mean age was 63.8 ± 10.5, mean time between FDG and death was 6.2 ± 2.8, and mean MMSE was 21.1 ± 7.1 at the time of PET (Table 1). Group averaged W‐map showed the typical temporo‐parietal AD pattern; hypometabolism in these regions was associated with higher Thal phase and Braak stage (Figure 1). Lewy body disease (LBD), argyrophilic grain disease (AGD), aging‐related tau astrogliopathy (ARTAG), cerebral amyloid angiopathy (CAA) and vascular brain injury (VBI) were associated with hypometabolism in small discrete regional clusters (Figure 2). The Landau AD ROI was associated with both Thal (rs = ‐0.325, p = .015) and Braak stages (rs = ‐0.447, p<.001). The CIS sign was not associated with the presence of LBD (t(53) = 1.01, p = 0.319). Both the IMT (t(53) = ‐2.25, p = .028) and IMT/FSO (t(53) = ‐3.40, p = .001) ratios were associated with the presence of LATE‐NC.


Co‐pathologies make a relatively small contribution to hypometabolism in patients with intermediate‐high ADNC. Novel biomarkers are needed for precise detection of co‐pathologies.

More from our Archive