Table 6.
Random forest model results for tissue and saliva samples using either mgmiRs and demographic variables or the mgmiRs alone
| Model | AUC % (95% CI) | Sensitivity % [median (2.5–97.5%)] | Specificity % [median (2.5–97.5%)] | PPV | NPV |
|---|---|---|---|---|---|
| Tissue | |||||
| mgmiRs and demographics | 98.9 (97.1–100) | 96.3 (90.7-100) | 96.7 (90.0-100) | 98.2 | 93.8 |
| mgmiRs | 99.4 (98.6–100) | 93.7 (87.3–98.4) | 100 (100–100) | 100 | 88.2 |
| Saliva | |||||
| mgmiRs and demographics | 93.8 (87.3–100) | 85.7 (71.4–96.4) | 94.4 (86.1–100) | 93.3 | 90.0 |
| mgmiRs | 93.5 (87.2–99.7) | 78.6 (60.7–92.9) | 100 (100–100) | 100 | 85.7 |
Univariate and multivariable logistic regression with dichotomized variables