huangzhii / SALMON

SALMON: Survival Analysis Learning with Multi-Omics Neural Networks
https://huangzhii.github.io/SALMON/
67 stars 22 forks source link

Kaplan-Meier survival plot #1

Closed bartmch closed 4 years ago

bartmch commented 4 years ago

Hi, Given the trained model and all the variables saved in .pickle format, how can I plot the Kaplan-Meier plot as shown in the article?

I tried multiple combinations but it's impossible to get the same monotonic decreasing curve as reported. Eg.: plotting os_test versus hazard_ratios_lbl_pred_all_test (with or without OS_event_test to censor the data)

Thank you!

huangzhii commented 4 years ago

Hi: Hazard ratios were derived from all five-fold testing sets. I used R library "survival" with overall survival time and event as input. You can check this https://cran.r-project.org/web/packages/survival/survival.pdf or this https://www.datacamp.com/community/tutorials/survival-analysis-R for more information. Hazards were used to differentiate the predicted OS event (either 1 or 0). "The Kaplan-Meier survival curves are generated by dichotomizing all testing patients to low risk and high risk groups via the median hazard ratio." is in the context of our paper. Thanks!