Neuroinflammation PET and long-term cognition and survival in symptomatic Alzheimer’s disease

Background: Neuroinflammation plays a key role in Alzheimer’s disease (AD) pathophysiology, but it is not clear how neuroinflammation contributes to disease progression. We aim to investigate the role of neuroinflammation on longitudinal cognition and survival in a unique cohort with PET imaging of translocator protein (TSPO) binding tracer [11C]PK11195 and long-term follow-up. We hypothesized that higher [11C]PK11195 binding would be associated with faster cognitive decline and higher mortality. Methods: 19 participants with AD dementia, 9 participants with MCI due to AD, and 21 healthy controls (HC) with historical dynamic [11C]PK11195 PET data were included. Principal component analysis was performed to identify relevant [11C]PK11195 patterns. An additional AD ROI consisting of temporal and parietal regions was investigated. [11C]PK11195 scores in the principal components (PCs) and AD ROI were compared between groups using ANOVA. Longitudinal MMSE covering a period up to 11 years was used to measure cognitive decline. We used linear mixed models with random subject-specific intercepts and slopes corrected for age, sex and syndrome diagnosis to investigate the association of neuroinflammation with cognition in MCI and AD. Survival data were available for all MCI and AD participants, up to 15.7 years after PET. To examine the influence of neuroinflammation on survival time, we used age, sex, and syndrome diagnosis adjusted cox proportional-hazards models. Results: Two PCs were retained. PC1 explained 55.4% of the variance and was most explained by [11C]PK11195 binding in the thalamus and entorhinal cortex. PC2 explained 15.3% of the variance and constituted of mostly the entorhinal cortex. There was no difference in [11C]PK11195 PET between AD, MCI and HCs (range F(2) = 0.157–1.231, P > 0.3). [11C]PK11195 did not predict longitudinal MMSE (PC1: β = 0.02, P = 0.73; PC2: β = 0.1, P = 0.44; AD ROI: β = 1.3, P = 0.57) or survival (PC1: HR = 0.90[95%CI: 0.80, 1.03], P = 0.13; PC2: HR = 0.96[0.75, 1.23], P = 0.72; AD ROI: HR = 0.02[0.00, 1.33], P = 0.06). Conclusions: Contrary to our hypothesis, we did not find evidence for [11C]PK11195 PET predicting long-term cognitive decline or survival. This may indicate that the level of [11C]PK11195 PET binding earlier in the disease trajectory is not directly linked to the long-term outcome.