In:
Alzheimer's & Dementia, Wiley, Vol. 17, No. S9 ( 2021-12)
Abstract:
Tau‐PET has great potential in Alzheimer’s disease (AD) clinical trials as pre‐trial inclusion biomarker and as trial endpoint. However, it is unclear which Tau‐PET ROI is optimal for tracking treatment response in relevant trial populations, and how best to enrich trials if using Tau‐PET as an endpoint. We aimed to determine the sample required to achieve 80% power to observe a reduction in Tau‐PET increase when using different combinations of A/T/N biomarkers as trial inclusion measures. Tau‐PET outcome variables were selected based on ROIs that showed greatest increases using longitudinal data. Methods We included 179 cognitively unimpaired (CU) individuals (39% Aβ‐positive) and 103 MCI (62% Aβ‐positive) with longitudinal Tau‐PET ([ 18 F]RO948;baseline and after ∼18‐months [mean 18.81±4.02] ) from BioFINDER‐2. Annual change in Tau‐PET SUVR was calculated across event‐based modelling (EBM) (Figure 1) ROIs. For the power analysis, linear models were fit with the best ROI identified in the first step as outcome and different combinations of A (Amyloid‐PET), T (CSF/plasma P‐tau217, baseline Tau‐PET in outcome ROI), and N (CSF/plasma NfL, hippocampal volume) biomarkers as predictors. Model fit was assessed using change in AIC relative to no enrichment (i.e. all subjects). Results In CU, annual Tau‐PET increase was highest in EBM stage‐I (1.26%,95%CI [0.82‐1.70%]) (Figure 2); in MCI, the greatest change was seen in EBM stage‐II (1.89%, 95%CI[1.22‐2.56%] ). Using these ROIs as outcome variables (assuming selection of top 50% of individuals), regression models showed that plasma P‐tau217 provided the best model fit in CU (44% fewer participants,∆AIC=‐5,P 〈 0.01), followed by CSF P‐tau217 (‐32%, ∆AIC=‐3,P 〈 0.05). In MCI, Amyloid‐PET provided the best model fit (‐71%,∆AIC=‐8, P 〈 0.01), followed by plasma P‐tau217 (‐69%,∆AIC=‐6,P 〈 0.01) and CSF P‐tau217 (‐71%,∆AIC=‐6,P 〈 0.01). When combining biomarkers (Figure 3), plasma P‐tau217, CSF P‐tau217 and Amyloid‐PET proved the best model in CU (‐74%,∆AIC=‐3,P 〈 0.0001). In MCI, Amyloid‐PET with EBM‐II baseline Tau‐PET and plasma P‐tau217 proved the best model (‐92%,∆AIC=‐3,P 〈 0.0001). Combining plasma P‐tau217 with Amyloid‐PET (CU) or baseline Tau‐PET (MCI) provided similar results, however (69%,∆AIC=‐4,P 〈 0.0001; 87%,∆AIC=‐6,P 〈 0.0001, respectively). Conclusion Plasma P‐tau217 combined with either Amyloid‐PET (preclinical AD) and baseline Tau‐PET (prodromal AD) may prove optimal and feasible enrichment strategies for trials with Tau‐PET as endpoint.
Type of Medium:
Online Resource
ISSN:
1552-5260
,
1552-5279
Language:
English
Publisher:
Wiley
Publication Date:
2021
detail.hit.zdb_id:
2201940-6
Permalink