Table 2.
The efficience of models to predict the treatment response in LARC petients.
| Training cohort | Validation cohort | |||||
|---|---|---|---|---|---|---|
| Score 3 | Score 6 | Score 7 | SVM1 | SVM2 | SVM3 | |
| DOWN-STAGING | ||||||
| AUC | 0.8630 (95% CI: 78.36–94.25%) |
0.8245 (95% CI: 73.07–91.83%) |
0.9297 (95% CI: 87.62–98.31%) |
0.8006 (95% CI: 65.28–94.84%) |
0.8462 (95% CI: 72.46–96.77%) |
0.8920 (95% CI: 79.40–99.01%) |
| Specificity | 0.82812 (95% CI: 68.750–95.312%) |
0.85938 (95% CI: 59.375–96.875%) |
0.90625 (95% CI: 73.438–98.438%) |
0.7778 (95% CI: 51.85–96.30%) |
0.7778 (95% CI: 51.85–96.30%) |
0.7778 (95% CI: 62.96–92.59%) |
| Sensitivity | 0.83333 (95% CI: 63.333–96.667%) |
0.73333 (95% CI: 56.667–93.333%) |
0.9000 (95% CI: 73.333–100.000%) |
0.8462 (95% CI: 61.54–100.00%) |
0.9231 (95% CI: 69.23–100.00%) |
0.9231 (95% CI: 76.92–100.00%) |
| Accuracy | 0.82979 (95% CI: 74.468–89.362%) |
0.81915 (95% CI: 69.149–89.362%) |
0.89362 (95% CI: 79.787–95.745%) |
0.8000 (95% CI: 65.00–92.50%) |
0.8250 (95% CI: 67.50–92.56%) |
0.8500 (95% CI: 70.00–92.50%) |
| PCR | ||||||
| AUC | 0.8361 (95% CI: 74.13–93.09%) |
0.8387 (95% CI: 74.83–92.91%) |
0.9078 (95% CI: 83.15–98.41%) |
0.8194 (95% CI: 69.08–94.79%) |
0.7581 (95% CI: 58.56–93.05%) |
0.8745 (95% CI: 74.82–99.49%) |
| Specificity | 0.86364 (95% CI: 63.64–100.00%) |
0.77273 (95% CI: 54.55–100.00%) |
0.86364 (95% CI: 72.73–100.00%) |
1.00 (95% CI: 80.00–100.00%) |
0.9000 (95% CI: 40.00–100.00%) |
0.9000 (95% CI: 50.00–100.00%) |
| Sensitivity | 0.77465 (95% CI: 54.93–91.55%) |
0.85915 (95% CI: 49.30–95.78%) |
0.88732 (95% CI: 69.01–97.18%) |
0.67742 (95% CI: 48.39–90.32%) |
0.67742 (95% CI: 32.26–100.00%) |
0.80645 (95% CI: 51.61–100.00%) |
| Accuracy | 0.78495 (95% CI: 64.52–89.25%) |
0.82796 (95% CI: 61.29–90.32%) |
0.88172 (95% CI: 75.27–94.62%) |
0.7561 (95% CI: 60.98–90.24%) |
0.73171 (95% CI: 48.78–92.68%) |
0.85366 (95% CI: 63.35–95.18%) |
| GOOD-RESPONSE | ||||||
| AUC | 0.8374 (95% CI: 75.95–91.53%) |
0.8039 (95% CI: 71.42–89.36%) |
0.9017 (95% CI: 83.30–97.05%) |
0.7920 (95% CI: 65.24–93.16%) |
0.7744 (95% CI: 62.65–92.23%) |
0.8972 (95% CI: 80.19–99.25%) |
| Specificity | 0.77083 (95% CI: 54.17–97.92%) |
0.8125 (95% CI: 52.08–93.75%) |
0.875 (95% CI: 70.83–97.92%) |
0.7143 (95% CI: 38.10–100.00%) |
0.7143 (95% CI: 38.10–95.24%) |
0.8571 (95% CI: 66.67–100.00%) |
| Sensitivity | 0.80435 (95% CI: 50.00–95.65%) |
0.73913 (95% CI: 56.52–95.65%) |
0.9130 (95% CI: 76.09–100.00%) |
0.8947 (95% CI: 47.37–100.00%) |
0.8421 (95% CI: 52.63–100.00%) |
0.8947 (95% CI: 68.42–100.00%) |
| Accuracy | 0.7766 (95% CI: 70.21–85.11%) |
0.7766 (95% CI: 69.15–85.11%) |
0.88298 (95% CI: 81.92–93.62%) |
0.7500 (95% CI: 65.00–87.50%) |
0.7750 (95% CI: 65.00–87.50%) |
0.8750 (95% CI: 75.00–95.00%) |