Table 3 DeLong’s test P values for pairwise AUC comparisons
Model_1 | Model_2 | Dataset | AUC_1 | AUC_2 | p |
|---|---|---|---|---|---|
Radiomics_Combined | Radiomics_Tumor | Training Cohort | 0.996 | 0.995 | 0.652 |
Radiomics_Combined | Radiomics_Tumor | Testing Cohort | 0.800 | 0.787 | 0.825 |
Radiomics_Combined | Radiomics_Tumor | Validation Cohort Ⅰ | 0.707 | 0.677 | 0.729 |
Radiomics_Combined | Radiomics_Tumor | Validation Cohort Ⅱ | 0.693 | 0.641 | 0.473 |
Radiomics_Combined | Radiomics_Tumor | Validation Cohort Ⅲ | 0.676 | 0.656 | 0.812 |
Radiomics_Combined | Radiomics_Tumor | Validation Cohort Ⅳ | 0.751 | 0.664 | 0.313 |
Deep Learning | Radiomics_Combined | Training Cohort | 1.000 | 0.996 | 0.017* |
Deep Learning | Radiomics_Combined | Testing Cohort | 0.818 | 0.800 | 0.819 |
Deep Learning | Radiomics_Combined | Validation Cohort Ⅰ | 0.732 | 0.707 | 0.757 |
Deep Learning | Radiomics_Combined | Validation Cohort Ⅱ | 0.696 | 0.676 | 0.793 |
Deep Learning | Radiomics_Combined | Validation Cohort Ⅲ | 0.764 | 0.693 | 0.267 |
Deep Learning | Radiomics_Combined | Validation Cohort Ⅳ | 0.720 | 0.751 | 0.721 |
