Table 4. Comparison of performances of TFPI2 and CA125 in discriminating CCC patients from other patient groups.
Discrimination | Biomarker | Sample set | No. of samples (No. of CCC) | AUC | 95% CI | Cut-off value | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) | Accuracy (%) |
---|---|---|---|---|---|---|---|---|---|---|---|
CCC vs other ovarian diseases | TFPI2 | Training set | 230 (29) | 0.893 | 0.812 to 0.972 | 345 pg/mL | 79.3 | 91.0 | 56.1 | 96.8 | 89.6 |
Validation set | 93 (7) | 0.817 | 0.597 to 1.000 | 71.4 | 91.9 | 45.5 | 97.5 | 90.3 | |||
CA125 | Training set | 230 (29) | 0.595 | 0.501 to 0.688 | 35 U/mL | 79.3 | 41.8 | 16.9 | 93.3 | 46.5 | |
Validation set | 93 (7) | 0.785 | 0.589 to 0.981 | 85.7 | 43.0 | 12.8 | 97.4 | 46.2 | |||
CCC vs BD + other EOCs | TFPI2 | Training set | 116 (29) | 0.854 | 0.761 to 0.952 | 345 pg/mL | 79.3 | 85.1 | 63.9 | 92.5 | 83.6 |
Validation set | 53 (7) | 0.797 | 0.558 to 1.035 | 71.4 | 85.7 | 50.0 | 94.7 | 83.7 | |||
CA125 | Training set | 116 (29) | 0.639 | 0.533 to 0.745 | 35 U/mL | 79.3 | 18.4 | 25.6 | 72.7 | 33.6 | |
Validation set | 53 (7) | 0.708 | 0.493 to 0.923 | 85.7 | 19.0 | 18.8 | 88.9 | 28.6 | |||
CCC vs EMS | TFPI2 | Training set | 100 (29) | 0.924 | 0.843 to 0.997 | 280 pg/mL | 82.8 | 93.0 | 82.8 | 93.0 | 90.0 |
Validation set | 24 (7) | 0.807 | 0.561 to 1.052 | 71.4 | 100 | 100 | 89.5 | 91.7 | |||
CA125 | Training set | 100 (29) | 0.700 | 0.584 to 0.817 | 35 U/mL | 79.3 | 43.7 | 36.5 | 83.8 | 54.0 | |
Validation set | 24 (7) | 0.803 | 0.545 to 1.060 | 85.7 | 23.5 | 31.6 | 80.0 | 41.7 | |||
All EOCs vs BN | CA125 | Training set | 230 (116*) | 0.814 | 0.760 to 0.869 | 35 U/mL | 81.0 | 55.3 | 64.8 | 74.1 | 68.3 |
Validation set | 93 (53*) | 0.761 | 0.664 to 0.857 | 75.5 | 62.5 | 72.7 | 65.8 | 69.9 |
Abbreviations: BD, borderline ovarian tumor; EMS, endometriosis; BN, benign ovarian tumor. Asterisks indicate numbers of EOC patients.