The use of synaptic biomarkers in cerebrospinal fluid to differentiate behavioral variant of frontotemporal dementia from primary psychiatric disorders and Alzheimer’s disease

Table 4 The AUC of the measured biomarkers as determined using ROC analysis

Diagnostic potential of the CSF biomarkers and selected diagnostic panel
	bvFTD vs PPD				bvFTD vs AD
Biomarker	AUC	CI	Specificity (%)	Sensitivity (%)	AUC	CI	Specificity (%)	Sensitivity (%)
NfL	0.95***	0.91–0.99	81.0	95.8	0.76***	0.66–.86	88.3	63.1
SNAP25	0.55	0.44–0.66	49.1	64.8	0.79***	0.70–0.87	81.7	68.4
Ng	0.53	0.43–0.64	53.0	52.1	0.69**	0.59–0.79	65.0	66.7
NPTX2	0.71***	0.62–0.81	70.2	66.2	0.63*	0.52–0.73	88.3	42.1
GluR4	0.56	0.45–0.67	70.4	46.5	0.56	0.45–0.66	44.0	68.5
Panel	0.96***	0.93-0.99	87.7	91.5	0.86***	0.79–0.92	98.2	61.1

Wald’s backward logistic regression was used to select the panel of CSF biomarkers. The biomarker panels selected were as follows: bvFTD vs PPD–NfL, NPTX2; bvFTD vs AD- NfL, SNAP25, Ng, GluR4. AUCs are shown for biomarkers alone. bvFTD, behavioral variant frontotemporal dementia; PPD, primary psychiatric disorders; AD, Alzheimer’s disease; AUC, area under curve; ROC, receiver operating characteristics; CI, confidence interval; NfL, neurofilament; SNAP25, synaptosomal-associated protein 25; Ng, neurogranin; NPTX2, neuronal pentraxin 2. *p < 0.05, **p < 0.01, ***p < 0.001

ISSN: 1758-9193