Table 2 Prediction models from the logistic regression and the ROC analysis results of the combined (a) and individual metabolites (b).
From: Urine metabolomics based prediction model approach for radiation exposure
Prediction models | AUC | SEa | 95% CIb | Youden index (J) | Sensitivity | Specificity | |
|---|---|---|---|---|---|---|---|
(a) | |||||||
P1 | p = [1/1 + e−{−9.026 + 4.319 * taurine}] | 0.989 | 0.0084 | 0.916–1.00 | 0.891 | 96.3 | 92.86 |
P2 | p = [1/1 + e−{−4.161 + 3.697 * taurine−0.894 * citrate}] | 0.995 | 0.0058 | 0.925–1.00 | 0.963 | 96.3 | 100.00 |
P3 | p = [1/1 + e−{−5.057 + 6.006 * taurine-0.500 * citrate-5.335 * α keto glutarate}] | 0.997 | 0.0032 | 0.930–1.00 | 0.963 | 96.3 | 100.00 |
P4 | p = [1/1 + e−{−0.498 + 13.771 * taurine-3.412 * citrate-34.461 * α keto glutarate + 515.183 * fumarate}] | 0.999 | 0.0018 | 0.933–1.00 | 0.964 | 100.00 | 96.4 |
(b) | |||||||
Metabolites | AUC | SEa | 95% CIb | Youden index (J) | Sensitivity | Specificity | |
|---|---|---|---|---|---|---|---|
1 | Citrate | 0.966 | 0.033 | 0.878–0.996 | 0.927 | 96.3 | 96.43 |
2 | Taurine | 0.989 | 0.0084 | 0.916–1.00 | 0.891 | 96.3 | 92.86 |
3 | Fumarate | 0.976 | 0.0192 | 0.893–0.999 | 0.927 | 96.3 | 96.43 |
4 | α-Ketoglutarate | 0.955 | 0.0298 | 0.862–0.993 | 0.892 | 100.00 | 89.29 |
5 | Creatine | 0.954 | 0.0238 | 0.861–0.992 | 0.777 | 77.8 | 100.00 |
6 | Succinate | 0.954 | 0.0256 | 0.861–0.992 | 0.817 | 88.89 | 92.86 |
