This repository contains the scripts for the End-to-End Survival prediction from histopathological images. The main script to run the survival is Survival_Main.py and the following table shows the description for the main required variables for the script:
| Input Variable name | Description |
|---|---|
| -datadir_train | Path to the normalized tiles |
| -slide_dir | Path to the slide table, containing the name of patients and their corresponding slides |
| -clini_dir | Path to the clini table, containing the name of patients and their corresponding clinical data |
| -outputPath | Path to the folder to save the results |
| -checkPointName | Name which will be used to save the trained model |
| -evenCol | The name of event column in the clini table (OS\Death...) |
| -timeCol | The name of time column in the clini table (OS_time\follow_up...) |
| -timeInDays | if the measured follow-up time is in days, this variable should be defined as True. |
| -patientCol | The name of patient column in the clini table (PATIENT\ID...) |
| -trainFull | If the aim is to use all the patients to train a single model, this variable should be defined as True. |
If the variable -trainFull in the training script is defined as True, then the Survival_Deploy script can be used to evaluate the performance of the model on a new data set.
Some of the functions established in this repository are inspired by https://colab.research.google.com/github/sebp/survival-cnn-estimator/blob/master/tutorial_tf2.ipynb#scrollTo=azrczYYVvEQb.