Development of thresholds and a visualization tool for use of a blood test in routine clinical dementia practice

Abstract

Introduction: We developed a multimarker blood test result interpretation tool for the clinical dementia practice, including phosphorylated (P-)tau181, amyloid-beta (Abeta)42/40, glial fibrillary acidic protein (GFAP), and neurofilament light (NfL).

Methods: We measured the plasma biomarkers with Simoa (n = 1199), applied LASSO regression for biomarker selection and receiver operating characteristics (ROC) analyses to determine diagnostic accuracy. We validated our findings in two independent cohorts and constructed a visualization approach.

Results: P-tau181, GFAP, and NfL were selected. This combination had area under the curve (AUC) = 83% to identify amyloid positivity in pre-dementia stages, AUC = 87%-89% to differentiate Alzheimer's or controls from frontotemporal dementia, AUC = 74%-76% to differentiate Alzheimer's or controls from dementia with Lewy bodies. Highly reproducible AUCs were obtained in independent cohorts. The resulting visualization tool includes UpSet plots to visualize the stand-alone biomarker results and density plots to visualize the biomarker results combined.

Discussion: Our multimarker blood test interpretation tool is ready for testing in real-world clinical dementia settings.

Highlights: We developed a multimarker blood test interpretation tool for clinical dementia practice. Our interpretation tool includes plasma biomarkers P-tau, GFAP, and NfL. Our tool is particularly useful for Alzheimer's and frontotemporal dementia diagnosis.

Keywords: Alzheimer; biomarker; blood test; dementia; glial fibrillary acidic protein; neurofilament light; phosphorylated tau; plasma.

FIGURE 1

Boxplots of plasma biomarkers according to their diagnostic group in the Amsterdam Dementia Cohort. Aβ, amyloid status; Abeta, amyloid beta; AD, Alzheimer's disease; DLB, dementia with Lewy bodies; FTD, frontotemporal dementia; GFAP, glial fibrillary acidic protein; MCI, mild cognitive impairment; NfL, neurofilament light; P‐tau, phosphorylated tau; SCD, subjective cognitive decline.

FIGURE 2

ROC curves with AUCs for the LASSO‐selected markers P‐tau181, GFAP, and NfL and their combination into a panel in the Amsterdam Dementia Cohort. ROC AUCs with 95% confidence intervals were computed with internal 10‐fold cross validation, for the clinical questions amyloid negative status (Aβ−) versus amyloid positive status (Aβ+) in the total cohort (A), Aβ− versus Aβ+ in the SCD and MCI set (B), AD versus FTD (C), controls versus FTD (D), AD versus DLB (E) and controls versus DLB (F). The panel consists of P‐tau181, GFAP, and age‐corrected NfL. NfL was corrected for age using a published formula. AD, Alzheimer's disease; AUC, area under the curve; DLB, dementia with Lewy bodies; FTD, frontotemporal dementia; GFAP, glial fibrillary acidic protein; MCI, mild cognitive impairment; NfL, neurofilament light; P‐tau, phosphorylated tau; ROC, receiver operating characteristics; SCD, subjective cognitive decline.

FIGURE 3

UpSet plots visualizing the proportion of Amsterdam Dementia Cohort individuals with certain combinations of normal/abnormal plasma biomarkers according to their diagnostic group for the six clinical questions. Participants were normal (gray dot) or abnormal (black dot) on their plasma biomarker results after applying Youden's index thresholds computed from receiver operating characteristics curves plotted for each clinical question. The proportion of participants with a certain combination of normal/abnormal plasma biomarker results according to their diagnostic groups and for the clinical questions amyloid negative status (Aβ−) versus amyloid positive status (Aβ+) in the total cohort (A), Aβ− versus Aβ+ in the SCD and MCI set (B), AD versus FTD (C), controls versus FTD (D), AD versus DLB (E) and controls versus DLB (F) are visualized. AD, Alzheimer's disease; DLB, dementia with Lewy bodies; FTD, frontotemporal dementia; GFAP, glial fibrillary acidic protein; MCI, mild cognitive impairment; NfL, neurofilament light (age‐corrected); P‐tau, phosphorylated tau; SCD, subjective cognitive decline.

FIGURE 4

Density plots visualizing the discriminative accuracy of P‐tau181, GFAP, and age‐corrected NfL interpreted in aggregation in the Amsterdam Dementia Cohort for six clinical questions. Probabilities of logistic regression analysis combining the LASSO‐selected markers P‐tau181, GFAP, and age‐corrected NfL were plotted and color coded for diagnostic group, including probability threshold lines at the Youden's index, at 90% specificity and at 90% sensitivity, according to their diagnostic groups and for the clinical questions amyloid negative status (Aβ−) versus amyloid positive status (Aβ+) in the total cohort (A), Aβ− versus Aβ+ in the SCD and MCI set (B), AD versus FTD (C), controls versus FTD (D), AD versus DLB (E), and controls versus DLB (F). Continuous probability density curves were applied to represent the probability thresholds for smoothed fitting. NfL was corrected for age using a published formula. AD, Alzheimer's disease; DLB, dementia with Lewy bodies; FTD, frontotemporal dementia; GFAP, glial fibrillary acidic protein; MCI, mild cognitive impairment; NfL, neurofilament light; P‐tau181, phosphorylated tau181; SCD, subjective cognitive decline.

FIGURE 5

Clinically implementable plasma result visualization tool. This figure presents two example results of individuals of our cohort: (A) male with MCI, age 65 years, and biomarker values of 4.15 pg/mL P‐tau181, 106 pg/mL GFAP and 21.8 pg/mL NfL; (B) female with FTD, age 65 years, and biomarker values of 1.52 pg/mL P‐tau181, 265 pg/mL GFAP, and 86.5 pg/mL NfL. These examples illustrate how the combination of the density plot, the UpSet plot, and the table with biomarker values together with their thresholds can be used for clinical interpretation of plasma biomarker results of new patients presenting at the memory clinic. FTD, frontotemporal dementia; MCI, mild cognitive impairment; NfL, neurofilament light; P‐tau181, phosphorylated tau181.