TABLE 3.
Model | AUC | ACC | Sensitivity | Specificity | p‐Value |
---|---|---|---|---|---|
Clinic‐radiological | 0.743 | 0.657 | 0.731 | 0.632 | 0.028 |
Radiomics | 0.795 | 0.820 | 0.808 | 0.829 | 0.021 |
MM‐Net | 0.774 | 0.814 | 0.769 | 0.829 | 0.043 |
PCMM‐Net | 0.843 | 0.824 | 0.818 | 0.816 | ‐ |
Note: The p‐value signifies the contrast of AUCs between PCMM‐Net and the clinico‐radiological model, as well as between PCMM‐Net and Clinic‐radiological, and further between PCMM‐Net and MM‐Net.
Abbreviations: AUC, area under the curve; ACC, Accuracy; MM‐Net, Multi‐Modal MR Images Convolution Neural Network; PCMM‐Net, Prior Clinico‐Radiological Features Informed Multi‐Modal MR Images Convolution Neural Network.