Table 4 Performance AU-ROC curve of the BOADICEA model and ML algorithms (with standard deviation) predicting breast cancer lifetime risk from simulated datasets (n = 2500) and Swiss clinic-based sample (n = 112,587 women from 2481 families)
Dataset | BOADICEA model | ML: random forest | ML: logistic regression | ML: adapt boosting | ML: linear model | ML: K-nearest neighbors | ML: linear discriminant | ML: quadratic discriminant | ML: MCMC GLMM |
|---|---|---|---|---|---|---|---|---|---|
A.Sim_no_signal | 0.5103 | 0.5020 (0.0197) | 0.5093 (0.0210) | 0.5029 (0.0177) | 0.5151 (0.0190) | 0.5254 (0.0199) | 0.5094 (0.0241) | 0.5002 (0.0216) | 0.5075 (0.0201) |
B.Sim_ atifical_signal | 0.5392 | 0.9101 (0.0148) | 0.9233 (0.0172) | 0.9321 (0.0122) | 0.6659 (0.0164) | 0.9301 (0.0159) | 0.9109 (0.0187) | 0.9244 (0.0166) | 0.9219 (0.0151) |
C.Sim_ atifical_signal + 20% missing | 0.5022 | 0.8977 (0.0183) | 0.9100 (0.0293) | 0.9291 (0.0156) | 0.6407 (0.0257) | 0.9232 (0.0180) | 0.8982 (0.0276) | 0.9209 (0.0297) | 0.9088 (0.0219) |
D.Sim_ atifical_signal + 20% missing +imputation | 0.5115 | 0.9028 (0.0127) | 0.9203 (0.0157) | 0.9299 (0.0110) | 0.6463 (0.0147) | 0.9276 (0.0140) | 0.9035 (0.0159) | 0.9220 (0.0141) | 0.9154 (0.0137) |
Swiss clinic-based sample | 0.5931 | 0.8535 (0.0214) | 0.8271 (0.0189) | 0.9017 (0.0162) | 0.6921 (0.0202) | 0.8377 (0.0156) | 0.7899 (0.0188) | 0.8369 (0.0192) | 0.8932 (0.0149) |