Detecting Amyloid–β Pathology in Subjective Cognitive Decline using novel CSF and plasma biomarkers

Abstract Background Cerebrospinal fluid (CSF) and blood‐based biomarkers are essential tools for detecting individuals with cognitive complaints and at‐risk of Alzheimer's disease (AD). However, their utility in detecting underlying AD pathology in subjective cognitive decline (SCD) remains to be fully established, and direct comparisons between different biomarkers and platforms are needed. This study aimed to compare the performance of CSF and blood‐based biomarkers for detecting Amyloid‐b (Ab) pathology in individuals with SCD, including core AD and AD‐related pathophysiology biomarkers Method We studied individuals with SCD from the β‐AARC cohort, which enrolled individuals seeking medical advice for cognitive complaints. Ab‐positivity (A+) was defined as CSF Aβ42/Aβ40<0.062 (Lumipulse). The following CSF biomarkers were compared: p ‐tau181, t‐tau, α‐synuclein, GFAP, IL‐6, NfL, Neurogranin, S100B, sTREM2, YKL40, and tau variants (NTA‐tau, p ‐tau205, p ‐tau235, and p ‐tau231). Plasma biomarkers included Aβ42/Aβ40, p ‐tau181, p ‐tau217, GFAP, NfL, NPTX2, SNAP25 and YKL40. Calculated ratios of amyloid and tau biomarkers in CSF and plasma were also analyzed. Biomarkers were analyzed using either Lumipulse, Simoa or Elecsys platforms. Diagnostic performance was assessed using Receiver Operating Characteristic (ROC) analysis, with sensitivity, specificity and negative (NPV) and positive predictive value (PPV) derived from the optimal Youden's Index cutoff. Result We analyzed 143 individuals with SCD (25% A+; Table 1) Beyond AD CSF core biomarkers (Aβ42/Aβ40, p ‐tau181), CSF p ‐tau231, p ‐tau205, and p ‐tau235 showed the largest effect sizes among significant biomarkers ( p <0.05, FDR corrected, Table 2), and achieved an NPV>0.95 for detecting amyloid status in CSF. In plasma (Table 3), p ‐tau217 and p ‐tau217/Aβ42 by either Lumipulse, Elecsys or Simoa, as well as p ‐tau181/Aβ42 (Elecsys), and Aβ42/Aβ40 (Lumipulse) exhibited the highest discrimination accuracies (AUC >0.90, NPV >0.95). However, only CSF p ‐tau181/Aβ42 (Lumipulse), showed a PPV>90%. Conclusion Several CSF and plasma biomarkers distinguished A+ from A− individuals with cognitive complaints but normal cognition. Furthermore, plasma biomarkers are effective for ruling out Aβ pathology in SCD individuals, showing high NPVs. However, due to modest PPVs, their utility for confirmation is limited in low A+ prevalence settings like in the β‐AARC study.