Objective

To assess the relationships between fluid and imaging biomarkers of tau pathology and compare their diagnostic utility in a clinically heterogeneous sample.

Methods

Fifty-three patients (28 with clinical Alzheimer disease [AD] and 25 with non-AD clinical neurodegenerative diagnoses) underwent β-amyloid (Aβ) and tau ([¹⁸F]AV1451) PET and lumbar puncture. CSF biomarkers (Aβ₄₂, total tau [t-tau], and phosphorylated tau [p-tau]) were measured by multianalyte immunoassay (AlzBio3). Receiver operator characteristic analyses were performed to compare discrimination of Aβ-positive AD from non-AD conditions across biomarkers. Correlations between CSF biomarkers and PET standardized uptake value ratios (SUVR) were assessed using skipped Pearson correlation coefficients. Voxelwise analyses were run to assess regional CSF-PET associations.

Results

[¹⁸F]AV1451-PET cortical SUVR and p-tau showed excellent discrimination between Aβ-positive AD and non-AD conditions (area under the curve 0.92-0.94; ≤0.83 for other CSF measures), and reached 83% classification agreement. In the full sample, cortical [¹⁸F]AV1451 was associated with all CSF biomarkers, most strongly with p-tau (r = 0.75 vs 0.57 for t-tau and -0.49 for Aβ₄₂). When restricted to Aβ-positive patients with AD, [¹⁸F]AV1451 SUVR correlated modestly with p-tau and t-tau (both r = 0.46) but not Aβ₄₂ (r = 0.02). On voxelwise analysis, [¹⁸F]AV1451 correlated with CSF p-tau in temporoparietal cortices and with t-tau in medial prefrontal regions. Within AD, Mini-Mental State Examination scores were associated with [¹⁸F]AV1451-PET, but not CSF biomarkers.

Conclusion

[¹⁸F]AV1451-PET and CSF p-tau had comparable value for differential diagnosis. Correlations were robust in a heterogeneous clinical group but attenuated (although significant) in AD, suggesting that fluid and imaging biomarkers capture different aspects of tau pathology.

Classification of evidence

This study provides Class III evidence that, in a clinical sample of patients with a variety of suspected neurodegenerative diseases, both CSF p-tau and [¹⁸F]AV1451 distinguish AD from non-AD conditions.

Objective

The prognostic value of cerebrospinal fluid neurofilament light chain, total tau, phosphorylated tau₁₈₁, and amyloid beta_1-42 was examined in frontotemporal dementia subtypes.

Methods

We compared baseline biomarkers between 49 controls, 40 patients with behavioral variant frontotemporal dementia, 24 with semantic variant primary progressive aphasia, and 26 with nonfluent variant primary progressive aphasia. Linear mixed effect models were used to assess the value of baseline biomarkers in predicting clinical and radiographic change in patient cohorts over multiple yearly follow up visits.

Results

Neurofilament light chain concentrations were lowest in controls. Elevated baseline neurofilament light chain predicted faster worsening in clinical severity, frontotemporal volume and frontotemporal fractional anisotropy in patients with behavioral variant frontotemporal dementia and nonfluent variant primary progressive aphasia. High total tau similarly predicted faster progression in nonfluent variant primary progressive aphasia. In behavioral variant frontotemporal dementia, higher phosphorylated tau₁₈₁ predicted faster clinical progression whereas lower amyloid beta_1-42 predicted faster volumetric and fractional anisotropy reduction. Neurofilament light chain and phosphorylated tau₁₈₁ were of greater predictive value in patients with tau pathology as compared to TDP-43 pathology. Baseline neurofilament light chain correlated with baseline clinical severity and frontotemporal volume in behavioral variant frontotemporal dementia. Baseline total tau correlated with baseline clinical severity in semantic variant primary progressive aphasia.

Interpretation

High cerebrospinal fluid neurofilament light chain predicts more aggressive disease in behavioral variant frontotemporal dementia and nonfluent variant primary progressive aphasia. Total tau, phosphorylated tau₁₈₁, and amyloid beta_1-42 also predict some measures of disease aggressiveness in frontotemporal dementia.

Background

Important insights into the early pathogenesis of Alzheimer's disease can be provided by studies of autosomal dominant Alzheimer's disease and Down syndrome. However, it is unclear whether the timing and spatial distribution of amyloid accumulation differs between people with autosomal dominant Alzheimer's disease and those with Down syndrome. We aimed to directly compare amyloid changes between these two groups of people.

Methods

In this cross-sectional study, we included participants (aged ≥25 years) with Down syndrome and sibling controls who had MRI and amyloid PET scans in the first data release (January, 2020) of the Alzheimer's Biomarker Consortium-Down Syndrome (ABC-DS) study. We also included carriers of autosomal dominant Alzheimer's disease genetic mutations and non-carrier familial controls who were within a similar age range to ABC-DS participants (25-73 years) and had MRI and amyloid PET scans at the time of a data freeze (December, 2020) of the Dominantly Inherited Alzheimer Network (DIAN) study. Controls from the two studies were combined into a single group. All DIAN study participants had genetic testing to determine PSEN1, PSEN2, or APP mutation status. APOE genotype was determined from blood samples. CSF samples were collected in a subset of ABC-DS and DIAN participants and the ratio of amyloid β42 (Aβ42) to Aβ40 (Aβ42/40) was measured to evaluate its Spearman's correlation with amyloid PET. Global PET amyloid burden was compared with regards to cognitive status, APOE ɛ4 status, sex, age, and estimated years to symptom onset. We further analysed amyloid PET deposition by autosomal dominant mutation type. We also assessed regional patterns of amyloid accumulation by estimated number of years to symptom onset. Within a subset of participants the relationship between amyloid PET and CSF Aβ42/40 was evaluated.

Findings

192 individuals with Down syndrome and 33 sibling controls from the ABC-DS study and 265 carriers of autosomal dominant Alzheimer's disease mutations and 169 non-carrier familial controls from the DIAN study were included in our analyses. PET amyloid centiloid and CSF Aβ42/40 were negatively correlated in carriers of autosomal dominant Alzheimer's disease mutations (n=216; r=-0·565; p<0·0001) and in people with Down syndrome (n=32; r=-0·801; p<0·0001). There was no difference in global PET amyloid burden between asymptomatic people with Down syndrome (mean 18·80 centiloids [SD 28·33]) versus asymptomatic mutation carriers (24·61 centiloids [30·27]; p=0·11) and between symptomatic people with Down syndrome (77·25 centiloids [41·76]) versus symptomatic mutation carriers (69·15 centiloids [51·10]; p=0·34). APOE ɛ4 status and sex had no effect on global amyloid PET deposition. Amyloid deposition was elevated significantly earlier in mutation carriers than in participants with Down syndrome (estimated years to symptom onset -23·0 vs -17·5; p=0·0002). PSEN1 mutations primarily drove this difference. Early amyloid accumulation occurred in striatal and cortical regions for both mutation carriers (n=265) and people with Down syndrome (n=128). Although mutation carriers had widespread amyloid accumulation in all cortical regions, the medial occipital regions were spared in people with Down syndrome.

Interpretation

Despite minor differences, amyloid PET changes were similar between people with autosomal dominant Alzheimer's disease versus Down syndrome and strongly supported early amyloid dysregulation in individuals with Down syndrome. Individuals with Down syndrome aged at least 35 years might benefit from early intervention and warrant future inclusion in clinical trials, particularly given the relatively high incidence of Down syndrome.

Funding

The National Institute on Aging, Riney and Brennan Funds, the Eunice Kennedy Shriver National Institute of Child Health and Human Development, the German Center for Neurodegenerative Diseases, and the Japan Agency for Medical Research and Development.