Table 3. ROC analysis and calculus of sensitivity, specificity and predictive values.

			Controls (n = 25)		Cervical Cancer (n = 44)
Genes	AUC	Cut-off valuea	FPF	TNF	TPF	FNF	p-valueb	Sensitivity	Specificity	PPV	NPV	Youden indexc
CDKN2A	0.996	18	0	25	42	2	<1×10−10	0.95	1	100	92.6	0.95
CCNB2	0.995	58	0	25	43	1	<1×10−10	0.98	1	100	96.2	0.98
MKI67	0.995	79	0	25	43	1	<1×10−10	0.98	1	100	96.2	0.98
PRC1	0.995	80	0	25	43	1	<1×10−10	0.98	1	100	96.2	0.98
CDC2	0.995	85	0	25	42	2	<1×10−10	0.95	1	100	92.6	0.95
SYCP2	0.992	115	0	25	42	2	<1×10−10	0.95	1	100	92.6	0.95
NUSAP1	0.990	48	1	24	43	1	<1×10−10	0.98	0.96	97.7	96.0	0.94
PCNA	0.990	100	0	25	42	2	<1×10−10	0.95	1	100	92.6	0.95
TYMS	0.985	46	0	25	41	3	<1×10−10	0.93	1	100	89.3	0.93
CDC20	0.971	3	3	22	42	2	<1×10−10	0.95	0.88	93.3	91.7	0.83
CDKN3	0.970	83	1	24	41	3	<1×10−10	0.93	0.96	97.6	88.9	0.89
SMC4	0.960	431	1	24	40	4	<1×10−10	0.91	0.96	97.6	85.7	0.87
RFC4	0.905	221	4	21	42	2	<1×10−10	0.95	0.84	91.3	91.3	0.79
RRM2	0.905	103	5	20	41	3	3×10−9	0.93	0.8	89.1	87.0	0.73
TOP2A	0.866	128	5	20	43	1	<1×10−10	0.98	0.8	89.6	95.2	0.78
MCM2	0.846	121	4	21	40	4	2.5×10−9	0.91	0.84	90.9	84.0	0.75
ZWINT	0.827	59	7	18	39	5	1.1×10−6	0.89	0.72	84.8	78.3	0.61
CKS2	0.815	239	5	20	35	9	5×10−6	0.80	0.8	87.5	69.0	0.60
			TPF	FNF	FPF	TNF
CFD	0.982	478	24	1	2	42	<1×10−10	0.96	0.95	97.7	92.3	0.91
EDN3	0.968	42	23	2	4	40	<1×10−10	0.92	0.91	95.2	85.2	0.83
WISP2	0.926	151	24	1	10	34	2.1×10−8	0.96	0.77	97.1	70.6	0.73

AUC: area under the curve, FPF: false positive fraction, TNF: true negative fraction, TPF: true positive fraction, FNF: false negative fraction, PPV: Positive predictive value, NPV: Negative predictive value.

^aOptimal cut-off values (ng/ml) were selected according to the ROC analysis.

^bChi square test.

^cJ = sensitivity+specificity − 1.