Performance of the model
| High- vs low-risk . | Training . | Validation . | BV-DHAP . | BV-ICE . |
|---|---|---|---|---|
| Sensitivity | 61.5 | 36.4 | 61.5 | 61.5 |
| Specificity | 88.5 | 89.4 | 93.3 | 83.3 |
| PPV | 61.5 | 61.5 | 72.7 | 53.3 |
| NPV | 88.5 | 75.0 | 89.4 | 87.5 |
| High- vs low-risk . | Training . | Validation . | BV-DHAP . | BV-ICE . |
|---|---|---|---|---|
| Sensitivity | 61.5 | 36.4 | 61.5 | 61.5 |
| Specificity | 88.5 | 89.4 | 93.3 | 83.3 |
| PPV | 61.5 | 61.5 | 72.7 | 53.3 |
| NPV | 88.5 | 75.0 | 89.4 | 87.5 |
Performance of the model shown for the training and validation cohorts. The training cohort consists of the BV-DHAP and BV-ICE studies of which the model performance is also shown separately. The optimal cutoff for high- vs low-risk groups is based on the percentage of PFS events in the training cohort which was 23%.
BV, brentuximab vedotin; DHAP, dexamethasone, high-dose cytarabine and cisplatin; ICE, ifosfamide, carboplatin, etoposide; NPV, negative predictive value; PPV, positive predictive value.