Recently, while helping a colleague create plots, I uncovered some new issues and solutions related to ggpubr and survival analysis packages. Here are the records.
survminer Plot Saving Issue
The survminer package includes a convenient function called ggsurvplot, which allows for easy creation of survival curves with group significance tests, as well as frequency tables below the plot. However, this function returns a ggsurvplot object rather than a ggplot() object. This object contains survObj$plot for the survival curve and survObj$table for the frequency table. Therefore, if you need to adjust or modify the plot using ggplot syntax, you should operate on the objects within these two components.
ggpubr Still Has Bug with Significance Test Annotations in Faceted Plots
As of December last year, there is still a bug in ggpubr where box plots (other facets may also have this issue) append significance test annotations incorrectly after faceting. The data and annotation positions are incorrect. You need to manually draw individual plots and then combine them.
Table Annotations Inside the Plot
To be filled in later.
Multiple Ways to Implement Significance Annotations
Due to a complex requirement from a client, where significance tests should be based on a reference group with two groups instead of one, ggpubr‘s built-in functions cannot meet this need. I spent considerable time researching how to manually implement it, and different methods for various requirements are recorded here:
To be filled in later.
