Search code examples
rggplot2plot

How to remove automated "strata"-text in ggsurvplot legend?

Please, find a data sample p below. I have produced the following ggsurvplot:

enter image description here

I would like to remove the encircled "Strata"-text from the legend, that prints automatically. I think it is redundant and disrupt the " graphic balance".

It prints when I run this script: 

#Fit the data
fit <- survfit(Surv(p$rfs, p$recurrence) ~ p$test, data=p)

#Plot

j <- ggsurvplot(
  fit,                     
  data = p, 
  risk.table = TRUE, 
  pval = TRUE,      
  pval.coord = c(0, 0.25),
  conf.int = T,         
  legend.labs=c("TERTp-wt (all)", "TERT-alt (all)"),  
  size=0.7,                    
  xlim = c(0,50),
  #alpha=c(0.4),
  conf.int.alpha=c(0.1),
  break.x.by = 6,    
  xlab="Time in months",
  ylab="Probability of recurrence-free survival",
  ggtheme = theme,             
  surv.median.line = "v",
  ylim=c(0,1),
  tables.theme=theme,
  surv.scale="percent",
  tables.col="strata",
  risk.table.col = "strata",
  risk.table.y.text = FALSE,
  tables.y.text = FALSE)

j$table <- j$table + labs(x = NULL, y = NULL)

j$plot <- j$plot+annotate("text", x = 14, y = 0.005, 
           label = "14 (95% CI: 10 - 24)", cex=3.3, vjust=0, 
           hjust = 1.1, fontface=2)

j

Data sample

p <- structure(list(rfs = c(38, 11.08, 49.5, 21.37, 73.5, 89, 0.72, 
64.3, 78, 22.9, 50.5, 99.8, 102.48, 114.35, 16.44, 53, 41, 96.2, 
113.42, 8.5, 25.7, 169.22, 1.97, 46.7, 71.5, 51.1, 88.5, 85.4, 
23, 42.7, 90.9, 102.61, 29.2, 48.2, 120, 69.5, 75.16, 48, 0.13, 
179.34, 70.19, 75.3, 22, 126.74, 69.8, 25.15, 42.35, 31.2, 2.04, 
59.9, 106.88, 23.6, 364.73, 67, 160.83, 22.68, 7.5, 67.4, 20, 
184.34, 72, 12, 386.19, 44.8, 42.9, 163.95, 63.4, 57.1, 0.46, 
66.9, 128, 1.02, 43.5, 93, 81.3, 47.9, 72.4, 96.2, 90.1, 99.1, 
90.8, 75.9, 88.01, 57.3, 97.2, 101.33, 136.27, 95.5, 97.1, 14, 
3.3, 72, 56, 140, 12, 12, 31, 45, 2.9, 48), recurrence = c(1L, 
0L, 1L, 0L, 0L, 1L, 0L, 0L, 1L, 0L, 1L, 0L, 0L, 0L, 0L, 0L, 1L, 
0L, 0L, 1L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 1L, 0L, 0L, 0L, 1L, 0L, 
0L, 0L, 0L, 0L, 1L, 0L, 0L, 0L, 1L, 0L, 0L, 1L, 0L, 0L, 0L, 0L, 
0L, 0L, 0L, 0L, 0L, 0L, 0L, 1L, 0L, 1L, 0L, 0L, 1L, 0L, 0L, 0L, 
0L, 0L, 0L, 0L, 0L, 1L, 0L, 0L, 1L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 1L, 1L, 1L, 0L, 1L, 1L, 1L, 1L, 
1L, 1L, 1L), test = c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 0L, 0L, 0L, 
0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 
0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L)), row.names = c(NA, 100L
), class = "data.frame")
Solution

  • An alternative way of removing the legend title is setting legend.title=element_blank() in the plot theme.

    Adding legend.title = "" to ggsurvplot() works well when the legend is on the top side of the plot. However, if the legend is positioned inside the plot, using legend.title = "" replaces the title with an empty string and therefore causes empty space between the legend box and the legend text. This may become visible if you have a background colour or your legend box overlaps with the survival lines.

    Therefore using element_blank() may be more useful when your legend is positioned inside the plot:

    j <- ggsurvplot(
  fit,                     
  data = p, 
  risk.table = TRUE, 
  legend = c(0.2, 0.13),
  risk.table.y.text = FALSE)

j$plot <- j$plot +
  theme(legend.title=element_blank())