Table 2.

The workflow of building model.

Algorithm: Build model
Input: Training data set, test data set
Output: Trained model and performance of the model
1	for feature type in clinical data, H&E images, clinical + H&E images do
2	for 2-fold cross validation process do
3	if feature type = clinical data do
4	sort importance of features in Random forest;
5	train model on train data set;
6	compute AUC from ROC curve for one subset of cross validation;
7	draw ROC curve based on the results of 2-fold CV;
8	train the model with whole train data set;
9	test the performance of the model on test data set;
10	final;