Table 2.

Performances of the logistic regression (LR) model, FIB-4, and non-alcoholic fatty liver disease fibrosis scores (NFS) in the diagnosis of advanced liver fibrosis.

Method	Accuracy	F1_S0-2	F1_S3-4	auROC	auPR	Specificity	Sensitivity
Discovery set
LR model	0.78	0.86	0.46	0.82	0.63	0.80	0.69
FIB4_1.45	0.73	0.81	0.52	0.79	0.58	0.80	0.58
FIB4_3.25	0.75	0.85	0.20	0.79	0.58	0.80	0.58
NFS_-1.455	0.68	0.74	0.57	0.75	0.49	0.80	0.47
NFS_0.676	0.74	0.84	0.19	0.75	0.49	0.80	0.47
Validation set 1
LR model	0.84	0.90	0.60	0.89	0.62	0.80	0.81
FIB4_1.45	0.82	0.88	0.60	0.85	0.60	0.80	0.71
FIB4_3.25	0.79	0.88	0.11	0.85	0.60	0.80	0.71
NFS_-1.455	0.77	0.84	0.59	0.85	0.57	0.80	0.74
NFS_0.676	0.80	0.88	0.17	0.85	0.57	0.80	0.74
Validation set 2
LR model	0.74	0.82	0.56	0.71	0.61	0.80	0.50
FIB4_1.45	0.65	0.75	0.43	0.63	0.54	0.80	0.38
FIB4_3.25	0.69	0.81	0.12	0.63	0.54	0.80	0.38
NFS_-1.455	0.46	0.45	0.48	0.59	0.39	0.80	0.25
NFS_0.676	0.65	0.78	0.15	0.59	0.39	0.80	0.25

FIB4_1.45 and FIB4_3.25 indicate FIB-4 with different thresholds of 1.45 and 3.25. NFS_-1.455 and NFS_0.676 indicate NFS with different thresholds of −1.455 and 0.676.