Table 5.
Model | Cohorts | AUC | Accuracy | Sensitivity | Specificity | Precision | Recall | F1 |
---|---|---|---|---|---|---|---|---|
Clinical | Training | 0.760 | 0.641 | 0.844 | 0.622 | 0.172 | 0.844 | 0.256 |
Test | 0.777 | 0.741 | 0.778 | 0.736 | 0.269 | 0.778 | 0.4 | |
Radiomicsa | Training | 0.829 | 0.835 | 0.719 | 0.846 | 0.303 | 0.719 | 0.426 |
Test | 0.846 | 0.660 | 0.944 | 0.625 | 0.239 | 0.944 | 0.382 | |
Combined | Training | 0.847 | 0.867 | 0.656 | 0.887 | 0.351 | 0.656 | 0.457 |
Test | 0.857 | 0.796 | 0.833 | 0.792 | 0.334 | 0.833 | 0.476 |
Abbreviation: SPN solitary pulmonary nodules, IMA invasive mucinous adenocarcinoma, LADC lung adenocarcinoma, AUC area under the curve, INMA invasive non-mucinous adenocarcinomas.
Represents models were constructed using support vector machine (SVM).