Table 4.
A. Prediction of disease recurrence status assuming a logistic regression model
Continuous/binary coding
Probability for model: likelihood ratio testAUC
Nomogram<0.0001/−0.81/−
GEMCaP:
    Fixed0.34/0.160.60/0.59
    Floating0.18/0.020.64/0.65
    Integrated0.19/0.030.63/0.65
Probability of GEMCaP as an independent predictor of status in addition to the nomogram:
GEMCaP (Binary):
    FixedNot significant0.80
    FloatingP = 0.020.85
    IntegratedP = 0.0550.84
B. Prediction of the added benefit of GEMCaP
Model
Nomogram scoreNomogram + floatingNomogram + integrated
Sensitivity60%77%77%
Specificity71%79%75%
Positive predictive value72%82%79%
Negative predictive value59%73%72%
Accuracy65%78%76%

NOTE: Due to the selection of the cut-point for the nomogram, a binary classification to predict disease status is not possible.