Identifying the Best Marker Combination in CEA, CA125, CY211, NSE, and SCC for Lung Cancer Screening by Combining ROC Curve and Logistic Regression Analyses: Is It Feasible?
Table 8
Significance of the best marker combinations for lung cancer screening by ROC curve and logistic regression analyses.
AUC
CutofflogitP
Pcutoff
LogitP
SEN
SPE
PPV
NPV
+LR
−LR
Yden
Healthy subjects
CEA + CA125
0.863
−0.381
0.406
−2.465 + 0.577 CEA + 0.035 CA125
0.708
0.855
0.799
0.783
4.883
0.342
0.563
Pneumonia patients (male)
CEA + CY211
0.703
0.044
0.511
−0.237 + 0.008 CEA + 0.068 CY211
0.522
0.784
0.751
0.568
2.417
0.610
0.306
0
0.500
0.617
0.684
0.709
0.589
1.953
0.560
0.301
Pneumonia patients (female)
CEA
0.692
−0.257
0.436
−0.803 + 0.064 CEA
0.350
0.969
0.887
0.682
11.290
0.671
0.319
−0.628
0.348
0.592
0.713
0.589
0.716
2.063
0.572
0.305
∗Recalculated parameters after adjusting cutofflogitP.