Aims: We assessed the utility of plasma neurofilament light chain (NfL) and phosphorylated-tau217 (p-tau217) in patients attending memory clinic. We (1) described the initial cohort of participants recruited; (2) described biomarker levels by diagnostic group; and (3) described performance of biomarkers in classifying participants by diagnostic group.
Methods: Patients were recruited from a memory clinic in Australia. In addition to routine clinical assessment, participants provided a blood sample, which was analysed for NfL and p-tau217 on the Roche Elecsys platform. Gold standard diagnosis was clinical diagnosis following standard-of-care multidisciplinary assessment.
Results: Seventy-eight participants were recruited (mean[SD] age 77.5[8.0], 55.1% female). Median(IQR) time from first visit to diagnosis was 147(77-185) days. Twenty-nine (37.2%) participants were diagnosed with dementia, the most common subtype being Alzheimer’s disease. Thirty-one participants had mild cognitive impairment (MCI) and 13 had subjective cognitive impairment (SCI) or normal cognition. Median(IQR) NfL in those with dementia was 3.55pg/mL(2.94-5.63), 2.96 pg/mL(2.46-3.93) in MCI, and 2.39 pg/mL(1.91-4.27) in SCI/normal cognition. Median(IQR) p-tau217 was 0.58pg/mL(0.46-0.85) in Alzheimer’s dementia, 0.25pg/mL(0.17-0.44) in other dementia subtypes, 0.44pg/mL(0.24-0.62) in non-amnestic MCI, 0.61 pg/mL(0.43-0.70) in amnestic MCI and 0.21pg/mL(0.15-0.26) in SCI/normal cognition. The AUC-ROC for p-tau217 in classification of presumed underlying amyloid-beta pathology (AD dementia or amnestic MCI) versus other participants was 0.74(95%CI 0.62-0.82). Using Youden’s index, a cut-off p-tau217>0.27pg/mL had sensitivity of 96.4% and specificity of 50.0%.
Conclusions: In this pilot real-world study of patients without amyloid imaging or CSF, plasma NfL and p-tau217 showed promising associations and classification utility compared to gold-standard clinical assessment.