Table 4.
RSF Model Performance. Mean results for performance metrics with associated bootstrap percentile 95% confidence intervals are displayed for deep learning and radiomics features with manually or automatically segmented masks. C-index: concordance index; IBS: integrated brier score.
| Modality | C-Index | IBS | Modality | C-Index | IBS | ||
|---|---|---|---|---|---|---|---|
| RSF model with deep learning features (Manual masks) | RSF model with deep learning features (Automated masks) | ||||||
| CT | 0.730 (0.687, 0.774) | 0.171 (0.139, 0.208) | CT | 0.739 (0.703, 0.793) | 0.159 (0.13, 0.193) | ||
| PET | 0.595 (0.543, 0.663) | 0.177 (0.15, 0.204) | PET | 0.568 (0.512, 0.620) | 0.185 (0.165, 0.202) | ||
| Clinical | 0.595 (0.529, 0.659) | 0.177 (0.15, 0.204) | Clinical | 0.587 (0.536, 0.658) | 0.185 (0.165, 0.202) | ||
| PET + CT Ensemble | 0.741 (0.695, 0.802) | 0.168 (0.141, 0.197) | PET + CT Ensemble | 0.740 (0.695, 0.792) | 0.165 (0.142, 0.189) | ||
| PET + CT + Clinical Ensemble | 0.737 (0.680, 0.804) | 0.169 (0.14, 0.199) | PET + CT + Clinical Ensemble | 0.732 (0.682, 0.783) | 0.166 (0.141, 0.193) | ||
| RSF model with radiomics features (Manual masks) | RSF model with radiomics features (Automated masks) | ||||||
| CT | 0.735 (0.687, 0.787) | 0.159 (0.135, 0.181) | CT | 0.718 (0.665, 0.771) | 0.160 (0.139, 0.187) | ||
| PET | 0.572 (0.520, 0.630) | 0.191 (0.174, 0.207) | PET | 0.526 (0.502, 0.565) | 0.211 (0.197, 0.242) | ||
| Clinical | 0.597 (0.542, 0.655) | 0.191 (0.174, 0.207) | Clinical | 0.584 (0.52, 0.631) | 0.211 (0.197, 0.242) | ||
| PET + CT Ensemble | 0.725 (0.689, 0.769) | 0.169 (0.153, 0.187) | PET + CT Ensemble | 0.703 (0.654, 0.764) | 0.179 (0.163, 0.204) | ||
| PET + CT + Clinical Ensemble | 0.717 (0.674, 0.757) | 0.168 (0.149, 0.189) | PET + CT + Clinical Ensemble | 0.700 (0.645, 0.754) | 0.176 (0.158, 0.206) | ||