Amyloid-PET–Positive Patient With bvFTD

PRACTICAL IMPLICATIONS

Positive amyloid PET does not exclude FTD pathology in the diagnostic workup of dementia patients.

Case

A 65-year-old man was referred to a local memory clinic with memory complaints, but clinical assessment found no abnormalities. When he presented 2 years later to our clinic, social disinhibition, reduced empathy, poor judgment, and hoarding had become obvious. He showed no insight. He had ischemic heart disease and was on preventive treatment. His mother died at the age of 97 suffering from dementia. Neurologic examination was normal. During neuropsychological examination, he exhibited verbal and behavioral disinhibition, inattention, emotional blunting, and unconcern. He had prominent difficulties in abstraction, set shifting, and sequencing with significant impact on memory tests (Table). A clinical diagnosis of behavioral variant FTD (bvFTD) was made. MRI (Figure, A) showed right- more than left-sided temporal atrophy, bilateral frontal and milder parietal atrophy. Fluorodeoxyglucose (FDG)-PET (Figure, B) demonstrated frontotemporal hypometabolism. Metabolism in the posterior cingulate was normal. He was homozygous for the APOE ε4 allele and negative for the C9orf72 expansion and mutations in MAPT, GRN, PSEN1, and APP. [18F]-Florbetapir PET (Figure, C) revealed increased tracer binding in all cortical regions corresponding to a centiloid value of 74%.

Table.

Longitudinal Results of Selected Neuropsychological Test Performance

Figure. MRI Brain, FDG-PET Brain, Florbetapir PET, and Histologic Slides With Amyloid, Tau, and TDP Pathology.

(A) MR and PET scans were obtained 3 years after clinical diagnosis. Coronal T1 MRI (in radiological orientation) showing right more than left temporal, bilateral frontal and milder parietal atrophy. (B) Color coded 3-dimensional stereotactic surface projection maps of FDG-PET demonstrating frontotemporal hypometabolism and preservation of posterior cingulate/precuneus (color bar indicating z-values of hypometabolism, top: medial projection, bottom: lateral projection; R: right hemisphere, L: left hemisphere). (C) Standardized uptake value ratios (SUVRs) compared to cerebellum of Florbetapir PET revealing increased tracer binding in all cortical regions. (D) The fusiform gyrus shows severe atrophy (hematoxylin-eosin, ×20). (E) Aβ peptide positive diffuse and neuritic plaques are shown in the occipital grey matter (immunoperoxidase, ×20). (F) Neurofibrillary tangles and neuropil threads in the entorhinal cortex (immunoperoxidase, phosphorylated tau, ×20). (G) TDP43 inclusions are documented in the dentate gyrus of hippocampus (immunoperoxidase, ×40). (H) Frontal lobe (immunoperoxidase, ×40). FDG = fluorodeoxyglucose.

Over the subsequent 3 years, his behavior deteriorated (Table), and formal neuropsychological assessment was no longer possible. He became incontinent without concern, developed grasp reflexes, and parkinsonism. He died at the age of 74 years, 9 years after clinical onset.

The brain showed considerable atrophy, neuronal loss, and gliosis in the temporal neocortex and amygdala with sparing of the hippocampus. Features were less prominent in the frontal cortex. The basal ganglia showed severe compromise of perforating arteries with calcification. The globus pallidus showed a lacunar infarct. All neocortical regions and subcortical gray matter contained diffuse subpial and perivascular amyloid plaques, but fewer neuritic and cored plaques. The occipital cortex showed a few cored plaques and severe amyloid angiopathy which was mild elsewhere, affecting mostly leptomeningeal vessels. Tau-related pathology was severe in CA1 and CA2 sectors and entorhinal cortex; the neocortex only showed few pretangles and tangles, scattered threads, and deposits in a few neuritic plaques with more prominent changes in the temporal and parietal lobes. A few tau deposits in neurons and subpial astrocytes were seen in the basal ganglia. Abnormal TDP-43 deposits consisted of neuronal cytoplasmic inclusions and dystrophic neurites (DN), which showed similar density in the hippocampus (dentate gyrus) and amygdala, although DN were more prominent in the temporal neocortex. TDP-43–related pathology was noticeable but milder in the frontal neocortex (Figure, D–H).

Pathologic features were in keeping with FTLD-TDP type A, Braak neurofibrillary tangle stage 2 (0–6),¹ Consortium to Establish a Registry for Alzheimer Disease neuritic plaque score B (sparse),¹ Thal phase 5 (Aβ plaque score 0–5),¹ Braak synuclein stage 0, and severe small vessel disease in the basal ganglia. LATE-NC stage 3 was considered in the differential diagnosis but believed to be unlikely, given the extent of TDP-43 in the temporal and frontal neocortex, mild Alzheimer disease (AD) changes, and absence of hippocampal sclerosis and lacunar infarcts being uncommon.

Discussion

As we previously reported,² this patient presented with a clinical and neuropsychological syndrome of bvFTD, but further investigations revealed conflicting results (MRI brain and FDG-PET in support, Florbetapir-PET against). We now report autopsy findings showing coexistent TDP-43 type A and AD pathologies and cerebrovascular disease in the context of APOE ε4 homozygosity.

Amyloid-PET imaging is becoming a standard clinical investigation. Recent studies³ document a frequent change of the clinical diagnosis when PET findings and clinical diagnosis seem to be incongruent. A diagnosis of bvFTD may be changed to AD when amyloid-PET is positive, a decision supported by cases diagnosed with bvFTD but found to have AD at postmortem.⁴ There may, however, be coexistent pathology or the scan report could be wrong (e.g., incorrect interpretation of off-target binding⁵). Patients with genetic- or autopsy-proven FTD are often “amyloid-positive” in CSF or PET or show amyloid pathology at postmortem.⁶ In patients with a clinical diagnosis of FTD who are APOE ε4 carriers, amyloid-PET scan is positive in 19% at age 60 and 43% at age 80.⁷

Using an evidence-based approach,⁴ our patient's probability of having AD, given a positive amyloid-PET, low pre-PET clinical probability of AD, his age, and APOE ε4 positivity, is only 0.2. Our patient's load of AD neuropathologic change was associated in 60% of patients in the National Alzheimer Coordinating Center Data Set with a Clinical Dementia Rating (CDR) sum of boxes score = 0 but 12% had a CDR sum of boxes score >12¹ demonstrating the variable clinicopathological relationships. According to the Vascular cognitive impairment neuropathology guidelines,⁸ there was <45% predictive probability that vascular pathology contributed to his cognitive impairment, given his age. Although the finding of TDP-43 type A pathology raises the question of LATE-NC, his clinical phenotype, age, MRI brain, FTD-PET, and distribution and severity of TDP-43 pathology were in keeping with FTLD.

This case emphasizes the importance of integrating clinical evaluation, patterns of brain atrophy and FDG-PET hypometabolism, potential genetic etiologies, pretest probability of amyloid positivity, and variability of clinicopathologic relationships before the final diagnosis.

Appendix. Authors

Study Funding

This case's brain is held by The Manchester Brain Bank, which is part of the Brains for Dementia Research program, jointly funded by Alzheimer Research UK and Alzheimer Society.

Disclosure

The authors report no disclosures relevant to the manuscript. Full disclosure form information provided by the authors is available with the full text of this article at Neurology.org/cp.