Table 3.

Cut-off scores and sensitivity/specificity values of fluid biomarkers for discriminating the prodromal AD or AD dementia from normal subjects

Characteristics	Normal (n = 50) vs. Prodromal AD (n = 23) groups					Normal (n = 50) vs. AD dementia (n = 28) groups
	Cut-off	Sen (%)	Spe (%)	AUC	p value	Cut-off	Sen (%)	Spe (%)	AUC	p value
Plasma factors
FXIIa activity, U/ml	>  26.3	73.9	74.0	0.783	<  0.001	>  27.2	78.6	78.0	0.906	<  0.001
FXIa activity, U/ml	>  1.13	52.2	52.0	0.566	0.367	>  1.17	57.1	58.0	0.659	0.020
FXa activity, U/ml	>  0.76	56.5	54.0	0.573	0.316	>  0.82	71.4	70.0	0.779	<  0.001
Kallikrein activity, U/ml	>  1.21	56.5	56.0	0.609	0.138	>  1.29	64.3	64.0	0.772	<  0.001
CSF biomarkers
Aβ1–42 levels, pg/ml	<  832.4	87.0	88.0	0.962	<  0.001	<  721.7	89.3	98.0	0.979	<  0.001
t-Tau levels, pg/ml	>  252.6	65.2	66.0	0.701	0.006	>  315.3	85.7	86.0	0.954	<  0.001
p-Tau181 levels, pg/ml	>  46.1	65.2	64.0	0.690	0.009	>  46.1	78.6	78.0	0.870	<  0.001
BK levels, pg/ml	>  46.1	54.5	56.5	0.656	0.073	>  52.8	74.1	73.9	0.783	0.001
Ratios
Aβ1–42 / FXIIa	<  33.8	87.0	86.0	0.965	<  0.001	<  27.4	100.0	100.0	1.000	<  0.001
Aβ1–42 / FXIa	<  742.5	82.6	82.0	0.930	<  0.001	<  635.4	96.4	96.0	0.993	<  0.001
Aβ1–42 / FXa	<  1,091	82.6	82.0	0.933	<  0.001	<  941.8	96.4	96.0	0.994	<  0.001
Aβ1–42 / Kallikrein	<  680.9	87.0	86.0	0.943	<  0.001	<  584.8	96.4	96.0	0.994	<  0.001

Statistically-derived optimal cut-off value was determined with the best balance between sensitivity (Sen) and specificity (Spe) values. Discrimination of prodromal AD and AD dementia from normal group was evaluated by receiver operator characteristics (ROC) curve analysis and quantified by the area under the curve (AUC) using SPSS software version 24.0