Biomarkers

Area of Science:

• Neurology
• Nuclear Medicine
• Biomarkers

Background:

• The Centiloid (CL) scale standardizes amyloid PET imaging across studies.
• Varying positivity thresholds complicate risk identification for cognitive decline and specific pathologies.
• This study aims to define optimal CL thresholds for a three-level classification system.

Purpose of the Study:

• To establish optimal Centiloid (CL) thresholds for a three-level classification system (NEGATIVE, GRAY-ZONE, POSITIVE).
• To validate these thresholds in distinguishing tau pathology, cognitive status, and clinical outcomes.
• To improve risk stratification and personalized treatment strategies in memory clinic cohorts.

Main Methods:

• 580 participants underwent amyloid PET scans; subsets had tau PET and CSF analysis.
• Participants classified as NEGATIVE (CL<12), GRAY-ZONE (12≤CL≤37), or POSITIVE (CL>37).
• Statistical analyses (Chi-squared, Kruskal-Wallis, linear mixed-effects models, ROC) assessed associations and prognostic implications.

Main Results:

• Predefined thresholds effectively stratified participants, with 44% of GRAY-ZONE showing discordant visual classification.
• CSF and tau levels differed significantly across groups (p<0.01), with GRAY-ZONE showing intermediate characteristics.
• Optimal CL thresholds (e.g., 13 for decline, 14/51 for tau) were identified via ROC analysis, linking lower thresholds to early tau and higher thresholds to advanced pathology.

Conclusions:

• Moving beyond binary amyloid classification is crucial, with at least three meaningful levels identified.
• The GRAY-ZONE group exhibits distinct clinical and biomarker profiles.
• Refined CL thresholds aid in identifying individuals at risk for cognitive decline and specific tau pathologies, crucial for personalized therapies.