Skip to main content

Table 1 Predictive performance of the five classifiers for eight scenarios (A-H). Each value is the average of 1,000 trials

From: Subgroup identification for treatment selection in biomarker adaptive design

Scenario A B C D E F G H
Predictive biomarkers 10 15 20 10 10 10 10 15
Significances 7.025 9.27 11.761 7.432 12.426 5.274 12.313 17.039
True positives 4.147 6.443 8.865 4.487 9.495 2.202 9.267 14.002
DLDA Sensitivity 0.598 0.651 0.694 0.626 0.979 0.507 0.985 0.99
Specificity 0.991 0.993 0.993 0.995 0.999 0.94 0.983 0.99
PPV 0.745 0.788 0.809 0.791 0.993 0.475 0.97 0.981
NPV 0.958 0.963 0.968 0.96 0.998 0.946 0.992 0.995
Accuracy 0.952 0.959 0.964 0.958 0.997 0.897 0.984 0.99
ASD (1,1) Sensitivity 0.784 0.808 0.841 0.943 0.991 0.953 0.989 0.991
Specificity 0.691 0.689 0.686 0.142 0.644 0.083 0.551 0.533
PPV 0.253 0.256 0.268 0.112 0.283 0.104 0.51 0.5
NPV 0.967 0.97 0.976 0.824 0.999 0.849 0.992 0.994
Accuracy 0.7 0.7 0.702 0.222 0.679 0.169 0.682 0.671
ASD (1,2) Sensitivity 0.635 0.683 0.723 0.831 0.982 0.807 0.977 0.982
Specificity 0.931 0.927 0.923 0.357 0.899 0.284 0.853 0.838
PPV 0.537 0.56 0.579 0.148 0.647 0.117 0.774 0.762
NPV 0.959 0.964 0.97 0.903 0.998 0.904 0.991 0.993
Accuracy 0.902 0.903 0.903 0.404 0.907 0.336 0.89 0.882
ASD (2,1) Sensitivity 0.464 0.533 0.573 0.766 0.953 0.654 0.581 0.649
Specificity 0.982 0.982 0.982 0.767 0.976 0.754 0.973 0.972
PPV 0.623 0.656 0.685 0.32 0.839 0.27 0.846 0.868
NPV 0.944 0.951 0.956 0.968 0.995 0.952 0.86 0.882
Accuracy 0.93 0.937 0.941 0.767 0.974 0.744 0.856 0.875
ASD (2,2) Sensitivity 0.331 0.408 0.453 0.64 0.926 0.445 0.424 0.511
Specificity 1 1 1 0.95 0.999 0.951 0.999 0.999
PPV 0.673 0.719 0.761 0.635 0.987 0.509 0.864 0.904
NPV 0.932 0.939 0.945 0.96 0.993 0.941 0.816 0.843
Accuracy 0.933 0.941 0.945 0.919 0.992 0.901 0.827 0.853
  1. PPV positive prediction value, NPV negative prediction value