Table 4. Results of applying RF to predict outcome of the paclitaxel signature for the METABRIC Discovery patient set.
| Type of
treatment |
Survival
years (as threshold) |
# Patients |
K (number
of genes to be used in random selection) |
Accuracy
(True Positive - TP) (%) |
Precision | F-Measure | MCC | AUC |
|---|---|---|---|---|---|---|---|---|
| Chemotherapy
(CT) |
3 | 22 | 7 | 61.1 | 0.617 | 0.612 | 0.224 | 0.444 |
| 4 | 7 | 66.7 | 0.643 | 0.646 | 0.189 | 0.715 | ||
| 5 | 19 | 66.7 | 0.722 | 0.687 | 0.189 | 0.571 | ||
| Hormone therapy
(HT) |
3 | 185 | 19 | 77.0 | 0.780 | 0.775 | 0.018 | 0.524 |
| 4 | 9 | 79.1 | 0.733 | 0.710 | 0.084 | 0.527 | ||
| 5 | 9 | 68.9 | 0.533 | 0.601 | -0.133 | 0.594 | ||
| CT and/or HT | 3 | 221 | 9 | 80.2 | 0.677 | 0.734 | -0.07 | 0.389 |
| 4 | 19 | 54.8 | 0.554 | 0.551 | -0.143 | 0.395 | ||
| 5 | 7 | 60.5 | 0.567 | 0.579 | 0.016 | 0.479 |
Paclitaxel gene panel consisted of 19 genes ( ABCB1, ABCB11, ABCC1, ABCC10, BAD, BBC3, BCL2, BCL2L1, BMF, CYP2C8, CYP3A4, MAP2, MAP4, MAPT, NR1I2, SLCO1B3, TUBB1, TUBB4A, TUBB4B).