Table 2.
Method | Predictors | Missing data | AUROC | |
---|---|---|---|---|
mean | SEM | |||
Logistic regression | EuroSCORE II | N/A | 0.651 | 0.012 |
Logistic regression | STS score | N/A | 0.646 | 0.007 |
Logistic regression | A | Mean imputation | 0.689 | 0.011 |
Logistic regression | A and J | Mean imputation | 0.717 | 0.011 |
Logistic regression | A | Marginalization | 0.682 | 0.011 |
Logistic regression | A and J | Marginalization | 0.723 | 0.010 |
Ours (3D ResNet-5022) | A and (I or J) | Marginalization | 0.722 | 0.011 |
Ours (Swin UNETR19) | A and (I or J) | Marginalization | 0.725 | 0.010 |
We report the mean and standard error of the mean (SEM) for 10-fold cross-validation.
A tabular patient factors, I unprocessed CT image, J CT measurements.