Table 3.
Efficacy of fecal SDC2, FIT, and serum CEA for CRC and APL diagnosis.
| No. | Sensitivity (%) | Specificity (%) | AUC (95% CI) | P value | |
|---|---|---|---|---|---|
| SDC2 | 518 | 75.3 | 81.4 | 0.784 (0.746–0.818) | < 0.0001 |
| FIT | 323 | 90.2 | 62.8 | 0.765 (0.715–0.810) | < 0.0001 |
| CEA | 282 | 50.6 | 77.5 | 0.684 (0.627–0.738) | < 0.0001 |
| SDC2 + FIT1 | 308 | 68.8 | 95.0 | 0.880 (0.839–0.914) | < 0.0001 |
| SDC2 + FIT + CEA2 | 202 | 70.0 | 96.3 | 0.905 (0.856–0.941) | < 0.0001 |
| SDC2 + FIT3 | 308 | 95.3 | 53.9 | 0.746 (0.694–0.794) | < 0.0001 |
| SDC2 + FIT + CEA3 | 202 | 97.5 | 48.8 | 0.731 (0.665–0.791) | < 0.0001 |
FIT included fecal Hb and TF. The sensitivity and specificity of detecting quantitative CEA alone depend on logistic regression. 1Use logistic regression to build prediction curves and ROC curve analysis to calculate the area under the curve. 2CEA results exceeding 5ng/ml are considered positive, use logistic regression to build prediction curves and ROC curve analysis to calculate the area under the curve. 3Result was considered positive if any one of them has a positive result.