Table 2.

Performance of the immunodiagnostic model and AFP to detect HCC in different sets. The cutoff value of the immunodiagnostic model was set at the maximum Youden index when the specificity was > 90%. The cutoff value of AFP was 20 ng·mL⁻¹. +LR, positive likelihood ratio; AFP, alpha‐fetoprotein; AUC, area under curve; −LR, negative likelihood ratio; NPV, negative predictive value; PPV, positive predictive value; Se, sensitivity; Sp, specificity; TAAb, autoantibody to tumor‐associated antigen. AFP‐negative means the HCC patients with serum AFP concentration lower than 20 ng·mL⁻¹, AFP‐positive indicates that the HCC patients with serum AFP concentration not < 20 ng·mL⁻¹, PP refers the prediction probability (PP) value of the 6‐TAAb panel by the binary logistic regression.

	AUC (95%CI)	Se (%)	Sp (%)	Youden index	+LR	−LR	Accuracy (%)	Kappa	PPV (%)	NPV (%)
Training set
AFP	0.702 (0.640–0.759)	44.5	95.2	0.397	9.35	0.58	70.0	0.437	98.6	58.3
Model a	0.835 (0.797–0.873)	57.0	90.3	0.473	5.86	0.48	73.3	0.469	85.7	67.2
Model + AFP	0.923 (0.886–0.951)	75.3	98.7	0.740	59.45	0.25	81.9	0.623	99.3	60.9
Model in AFP‐negative group	0.831 (0.770–0.892)	52.8	88.6	0.414	4.63	0.53	69.4	0.402	84.2	61.9
Model in AFP‐positive group	0.861 (0.806–0.917)	64.4	91.2	0.556	7.35	0.39	77.6	0.548	79.7	82.8
Validation set
AFP	0.739 (0.663–0.805)	64.7	96.1	0.631	8.55	0.36	72.0	0.466	98.7	60.2
Model	0.788 (0.737–0.839)	43.3	88.1	0.314	3.62	0.64	67.2	0.317	75.3	64.8
Model + AFP	0.927 (0.890–0.964)	82.8	90.3	0.731	40.72	0.19	85.4	0.694	94.1	73.7
Model in AFP‐negative group	0.834 (0.755–0.914)	36.6	90.3	0.269	4.03	0.68	68.9	0.293	72.7	69.1
Model in AFP‐positive group	0.826 (0.755–0.897)	45.3	90.3	0.356	4.68	0.61	65.7	0.340	85.0	57.7

^aThe model in this table is the immunodiagnostic model established by the binary logistic regression.