Table 4.
Efficacy of fecal SDC2, FIT, and serum CEA for early-stage CRC and APL diagnosis.
| No. | Sensitivity (%) | Specificity (%) | AUC (95%CI) | P value | |
| SDC2 | 443 | 69.9% | 81.4% | 0.756 (0.714–0.796) | < 0.0001 |
| FIT | 253 | 86.3% | 62.8% | 0.745 (0.687–0.798) | < 0.0001 |
| CEA | 203 | 59.0% | 60.8% | 0.622 (0.551–0.689) | 0.0018 |
| SDC2 + FIT1 | 246 | 60.6% | 95.0% | 0.856 (0.806–0.897) | < 0.0001 |
| SDC2 + FIT + CEA1 | 156 | 66.2% | 96.3% | 0.891 (0.832–0.936) | < 0.0001 |
| SDC2 + FIT2 | 246 | 93.9% | 53.9% | 0.739 (0.680–0.793) | < 0.0001 |
| SDC2 + FIT + CEA2 | 156 | 97.3% | 48.8% | 0.730 (0.654–0.798) | < 0.0001 |
1Use logistic regression to build prediction curves and ROC curve analysis to calculate the area under the curve. 2Result are considered positive if any one of them has a positive result.