. 2011 Aug 14;6:83–93. doi: 10.4137/BMI.S7513

Table 2.

Marker correlations and AUC improvement with 2-marker panel.

Secondary marker	AUC	Primary marker
		CA125			HE4			CA72-4
		Correlations¹	ΔAUC²	ΔAUC_Pr³	Correlations¹	ΔAUC²	ΔAUC_Pr³	Correlations¹	ΔAUC²	ΔAUC_Pr³
		Correlations¹	ΔAUC²	ΔAUC_Pr³
CA-125	0.896	–	–		–	–		–	–
HE4	0.797	0.72, 0.16	0.0009	0.002	–	–		–	–
CA72-4	0.753	0.58, −0.05	0.001	0.0003	0.42, −0.01	0.031	0.017	–	–
Kallikrein-6	0.685	0.56, 0.12	0.0001	0.0004	0.57, 0.32	0.001	0.0003	0.43, −0.05	0.002	0.024
CA15–3	0.668	0.61, 0.10	0.002	0.004	0.44, 0.15	0.0001	0.001	0.53, 0.0	0.0007	0.009
Mesothelin	0.665	0.43, −0.08	0.001	0.0005	0.61, 0.35	0.0002	0.0000	0.27, 0.10	0.017	0.027
B7-H4	0.654	0.46, −0.01	0.006	0.0001	0.36, 0.07	0.014	0.006	0.41, −0.01	0.007	0.020
MIF	0.605	0.19, 0.05	0.0005	0.0002	0.30, 0.14	0.0004	0.0000	0.26, 0.04	0.0002	0.003
MMP7	0.598	0.51, 0.07	0.0007	0.003	0.49, 0.44	0.0003	0.002	0.46, −0.10	0.0006	0.004
Prolactin	0.598	−0.11, 0.02	0.011	0.007	−0.25, −0.06	0.034	0.020	−0.15, 0.02	0.017	0.018
Spondin	0.597	0.25, 0.04	0.0004	0.0001	0.18, 0.25	0.001	0.001	0.08, −0.01	0.003	0.011
IGF-II	0.584	[−0.16, −0.07]	0.0002	0.0001	[−0.21, −0.20]	0.0006	0.0001	[−0.12, 0.04]	0.003	0.006
IGFBPII	0.571	0.46, 0.05	0.0002	0.003	0.42, 0.33	0.0009	0.002	0.31, −0.09	0.0003	0.002
Eotaxin	0.569	[0.04, 0.04]	0.005	0.006	[0.10, 0.04]	0.026	0.012	[−0.06, −0.04]	0.004	0.008
sVCAM-1	0.559	[−0.16, 0.00]	0.0004	0.0000	[−0.05, 0.11]	0.002	0.003	[−0.08, −0.08]	0.0002	0.002

Notes: Top 15 markers in terms of univariate AUC are listed.

^1.

Correlations (cases, controls) of primary marker with secondary marker. Brackets [,] denote that secondary marker is down-regulated in cases. Primary markers are all up-regulated in cases.

^2.

Observed increase in AUC for (optimal) combination of primary and secondary marker over that achieved with primary marker alone; AUCs computed non-parametrically. Bold indicates increase in AUC of 0.01 or more.

^3.

Predicted increase in AUC, based on assuming log MVN distribution, for (optimal) combination of primary and secondary marker over that achieved with primary marker alone.