Variables | AUROC | Sensitivity | Specificity | Accuracy | PPV | NPV |
---|---|---|---|---|---|---|
Training cohort | ||||||
 PCS-nomogram | 0.939 (0.913, 0.964) | 82.3% (75.3%, 87.6%) | 94.5% (90.2%, 97.0%) | 89.1% (85.2%, 92.0%) | 92.4% (86.5%, 95.8%) | 86.9% (81.5%, 90.9%) |
 Traditional model | 0.783 (0.734, 0.831) | 68.0% (61.1%, 75.0%) | 75.3% (68.5%, 81.0%) | 72.0% (67.0%, 76.6%) | 69.0% (61.0%, 75.9%) | 74.5% (67.7%, 80.2%) |
Internal validation cohort | ||||||
 PCS-nomogram | 0.895 (0.857, 0.932) | 85.3% (78.8%, 90.1%) | 84.9% (78.9%, 89.4%) | 85.1% (80.9%, 88.5%) | 82.6% (75.8%, 87.7%) | 87.4% (81.6%, 91.5%) |
 Traditional model | 0.791 (0.742, 0.839) | 63.3% (55.4%, 70.6%) | 83.8% (77.7%, 88.5%) | 74.5% (69.5%, 78.9%) | 76.6% (68.4%, 83.2%) | 73.2% (66.7%, 78.8%) |
External validation cohort | ||||||
 PCS-nomogram | 0.893 (0.855, 0.931) | 88.4% (82.4%, 92.5%) | 78.8% (71.8%, 84.4%) | 83.5% (79.0%, 87.2%) | 80.1% (73.5%, 85.4%) | 87.5% (81.1%, 91.9%) |
 Traditional model | 0.727 (0.672, 0.783) | 64.5% (56.7%, 71.6%) | 75.0% (67.8%, 81.1%) | 69.8% (64.6%, 74.6%) | 71.4% (63.5%, 78.3%) | 68.6% (61.4%, 75.0%) |