Copyright: ©Author(s) 2026.
World J Radiol. Mar 28, 2026; 18(3): 117599
Published online Mar 28, 2026. doi: 10.4329/wjr.v18.i3.117599
Published online Mar 28, 2026. doi: 10.4329/wjr.v18.i3.117599
Table 4 Model efficacy in predicting clinical response to vedolizumab
| Model | AUC | Sensitivity | Specificity | PPV | NPV |
| Training cohort | |||||
| Combined model | 0.757 (0.65-0.863) | 72.4% (42/58) | 74.3% (26/35) | 82.4% (42/51) | 61.9% (26/42) |
| Radiomics model | 0.646 (0.522-0.769) | 67.2% (39/58) | 62.9% (22/35) | 75.0% (39/52) | 53.7% (22/41) |
| Clinical model | 0.731 (0.626-0.835) | 55.2% (32/58) | 82.9% (29/35) | 84.2% (32/38) | 52.7% (29/55) |
| Validation cohort | |||||
| Combined model | 0.721 (0.56-0.883) | 76.2% (16/21) | 70.0% (14/20) | 72.7% (16/22) | 73.7% (14/19) |
| Radiomics model | 0.819 (0.687-0.951) | 81.0% (17/21) | 75.0% (15/20) | 77.3% (17/22) | 78.9% (15/19) |
| Clinical model | 0.79 (0.654-0.927) | 52.4% (11/21) | 90.0% (18/20) | 84.6% (11/13) | 64.3% (18/28) |
- Citation: Zhang XY, Li YK, Tian ZB, Guo QY, Liu JN, Liu RQ, Ren KY. Predictive model for vedolizumab efficacy in moderate-to-severe ulcerative colitis based on computed tomography-derived body compositions and nutritional inflammatory markers. World J Radiol 2026; 18(3): 117599
- URL: https://www.wjgnet.com/1949-8470/full/v18/i3/117599.htm
- DOI: https://dx.doi.org/10.4329/wjr.v18.i3.117599