Alzheimer's Imaging Consortium

Background: Higher cerebrospinal fluid noradrenergic metabolic turnover has been associated with higher levels of Alzheimer's disease pathology in cognitively impaired individuals. It remains unclear whether there is a specific anatomic vulnerability to metabolic alterations within the locus coeruleus (LC) and whether this hypermetabolism relates to steeper rates of pathology accumulation. Here, we overcome a spatial limitation of Positron Emission Tomography (PET) imaging in small nuclei with a dedicated Magnetic Resonance Imaging (MRI)-guided framework to recover PET resolution. This method allowed us to investigate the [18F]Fluorodeoxyglucose (FDG)-PET signal, reflecting energetic metabolic demand of the LC and its association with serial beta-amyloid and tau-PET.

Methods: We included 78 amyloid-positive, cognitively impaired participants from the ADNI-3 cohort (Table 1). The LC was identified using a dilated Keren template registered to each T1 (Fig 1A). The LC mask was subdivided into a rostral, middle, and caudal part to investigate region-specific associations (Fig 1B). We used an MR joint-entropy-penalty algorithm to quantify PET signal via deconvolution based on spatially-variant blur kernels. The deconvolution step was stabilized using an anatomical-based joint-entropy prior (Fig 1C). The FDG-PET signal was referenced to the pons, excluding the LC (Fig 1D). Amyloid positivity was determined as a centiloid score = 21, reflecting moderate neuritic plaques, on florbetaben and florbetapir scans. Linear mixed effects models investigated relationships between LC-FDG, global beta-amyloid and meta-temporal tau burden (flortaucipir) over time. Age and sex were included as covariates.

Results: Higher LC FDG-PET signal was associated with higher meta-temporal tau-PET at baseline (T = 2.87, p = 0.00463) and accumulation over time (T = 2.257, p = 0.0273; Fig 2A&C). Higher LC FDG-PET signal was associated with higher global amyloid at baseline (T = 2.066, p = 0.041) and faster amyloid accumulation over time in the dementia group (T = -2.74, p = 0.00904; Fig 2B&D), following a rostro-caudal gradient.

Conclusion: In a cohort with cognitive impairment, we observed steeper accumulation of Alzheimer's disease pathology in individuals with higher LC FDG-PET signal, with a specific amyloid-related vulnerability for the rostral parts of the LC. Future steps will be to include cognitively unimpaired individuals to investigate the relationship between LC FDG-PET signal and pathology accumulation in early stages of the disease process.