Table 2. Prognostic accuracy of the four markers by time-dependent ROC curve analyses.
Marker | Peak accuracy (months) | AUC (95%) | Cut-off (%) | Sensitivity | Specificity | PPV(*) |
---|---|---|---|---|---|---|
EDN/RB | 39 to 55 | 0.68 (0.57-0.78) | 80 | 0.59 | 0.77 | 0.72 |
p60/CAF-1 | 21 to 27 | 0.69 (0.58-0.79) | 24 | 0.69 | 0.69 | 0.69 |
HJURP | 28 to 29 | 0.69 (0.59-0.79) | 6 | 0.92 | 0.46 | 0.63 |
PDLI4 | 39 to 55 | 0.65 (0.53-0.78) | 20 | 0.86 | 0.44 | 0.61 |
PPV : positive predictive value