Skip to main content
. 2021 Jul 7;11:646190. doi: 10.3389/fonc.2021.646190

Table 3.

Predictive performance of machine-learning models with radiomic features, clinical features, and combined features in the independent test cohort.

Classifier Accuracy AUC Recall Precision F-score
RF_Radiomics 85.7% 0.941 0.875 0.808 0.840
RF_Clinic 42.9% 0.503 0.625 0.395 0.484
RF_Combination 82.1% 0.936 0.875 0.750 0.808