Table 2 Comparison of pre- and on-treatment biomarkers for predicting response and outcome. Evaluation of the performance of the top 10 pre- and on-treatment genes identified for predicting pathological response in the NEO dataset
Response accuracy | Response AUC | Outcome (log-rank) | ||||
|---|---|---|---|---|---|---|
NEO | I-SPY | NEO | I-SPY | NEO | I-SPY | |
On-treatment | ||||||
AAGAB | 100% | 78% | 1.00 | 0.63 | 0.048 | 0.0036 |
ZNF165 | 88% | 54% | 0.91 | 0.57 | 0.26 | 0.70 |
KRTCAP3 | 79% | 52% | 0.85 | 0.56 | 0.81 | 0.49 |
RFC2 | 79% | 40% | 0.85 | 0.35 | 0.51 | 0.44 |
C20orf151 | 70% | NA | 0.75 | NA | 0.36 | NA |
ARF5 | 70% | 43% | 0.75 | 0.36 | 0.0038 | 0.20 |
BSPRY | 70% | 48% | 0.75 | 0.49 | 0.47 | 0.19 |
NGRN | 58% | NA | 0.66 | Na | 0.53 | Na |
CHST7 | 29% | 46% | 0.21 | 0.52 | 0.65 | 0.40 |
SLC18B1 | 25% | Na | 0.18 | NA | 0.55 | NA |
Pre-treatment | ||||||
IGF1R | 74% | 63% | 0.76 | 0.69 | 0.36 | 0.11 |
CTNNB1 | 71% | 49% | 0.73 | 0.46 | 0.60 | 0.40 |
SLC20A2 | 71% | 56% | 0.72 | 0.57 | 0.063 | 0.56 |
HMGCL | 68% | 47% | 0.67 | 0.45 | 0.10 | 0.97 |
ST6GALNAC5 | 68% | 52% | 0.69 | 0.53 | 0.6 | 0.28 |
AAGAB | 65% | 57% | 0.65 | 0.58 | 0.71 | 0.20 |
C1orf51 | 62% | NA | 0.61 | NA | 0.12 | NA |
KRTCAP3 | 62% | 54% | 0.63 | 0.57 | 0.78 | 0.78 |
SETDB2 | 50% | 49% | 0.48 | 0.51 | 0.29 | 0.15 |
FADS2 | 29% | 48% | 0.27 | 0.5 | 0.14 | 0.73 |