Relevance of plasma biomarkers to imaging biomarkers

Abstract Background Recent advances in blood‐based biomarkers have shown promising abilities to estimate brain pathological statuses related to amyloid‐β (Aβ) deposition, phosphorylated tau deposition, and neurodegeneration using minimally invasive methods. The objective of this study was to evaluate the potential of blood‐based ATN classification, by analyzing relevance of plasma biomarkers to imaging biomarkers. Method There were 161 subjects, consisting of 62 cognitively normal (CN), 13 MCI, 40 Alzheimer’s dementia (AD) and 46 non‐Alzheimer’s dementia (non‐AD) individuals. Plasma Aβ composite biomarker was measured using immunoprecipitation‐mass spectrometry assay by combining plasma ratios of APP 669‐711 /Aβ 1‐42 and Aβ 1‐40 /Aβ 1‐42 . Plasma levels of p‐tau 181, glial fibrillary acidic protein (GFAP) and neurofilament light (NfL) were measured using the Simoa Ⓡ platform. PiB‐PET, THK5351‐PET, FDG‐PET and structural MRI images, and their representative semiquantitative values; PiB‐SUVR, THK‐SUVR, FDG‐PET scores (representing the degree of AD‐like glucose hypometabolism), and VSRAD ® scores (representing the degree of medial temporal atrophy), respectively, were used to analyze the relationships with the plasma biomarkers. Result The ROC analyses showed that AUCs of plasma Aβ composite, p‐tau 181, GFAP and NfL to discriminate Aβ status (PiB‐positive vs. negative) were 0.934, 0.835, 0.720 and 0.583, respectively. Whereas AUCs of those to discriminate clinical status (CN/MCI vs. dementia) were 0.649, 0.655, 0.800 and 0.803, respectively. Correlations to PiB‐SUVR were stronger in Aβ composite (r = 0.657) and p‐tau 181 (r = 0.551). Both p‐tau and GFAP showed higher correlations to THK‐SUVR (r = 0.444 and 0.416, respectively), and to FDG‐PET scores (r = 0.533 and 0.541, respectively) compared with other biomarkers. Correlations to VSRAD scores were only significant in NfL (r = 0.319). NfL also showed the highest correlation to MMSE scores (r = ‐0.583) (all the above r values were p < 0.001). These results were supported by multiple regression analyses between each imaging data and each plasma biomarker values using SPM8 adjusting for age and sex. Conclusion The results demonstrated reasonable relationships between the plasma biomarkers and imaging biomarkers, suggesting clinical utility of blood‐based ATN classification. GFAP and NfL are expected to reflect different aspects of neurodegeneration like FDG‐PET and structural MRI.