Change legend labels in ggcompetingrisks plot in r

Question

I have made a cumulative incidence plot using cmprsk::cuminc and then ggcompetingrisks for plotting. I want to change legend labels for the group. I've tried with legend.lab = c("A", "B", "C") . This works when I do a ggsurvplot without competing risk. But it does not work now. Any suggestions?

My code is like:

    fit <- cuminc(df$Time, df$Event, group = df$genotype)

    p <- ggcompetingrisks(fit, multiple_panels = FALSE, palette = "black",
                          legend.title = c("genotype"),
                          legend.labs = c("A", "B", "C"))

The thing is that the legend show both the Event (0, 1) and the Group (1, 2, 3). I want only the group to be shown in the legend and I want it to be named A, B and C..

Please help me!!

Answer 1

If I get your fit correct, group will be plotted as linetype so you can use scale_linetype_manual() to set the title of the legend, and the names. To turn of the other legend about event, you can use guides()

See below for an example using some simulated data:

library(survminer)
library(ggplot2)
library(cmprsk)

df = data.frame(
Time = rexp(100),
genotype = factor(sample(1:3,100,replace=TRUE)),
Event = factor(sample(0:1,100,replace=TRUE),0:1,c('no event', 'death'))
)

fit <- cuminc(df$Time, df$Event, group = df$genotype)

ggcompetingrisks(fit)
p <- ggcompetingrisks(fit, multiple_panels = FALSE)

p + scale_linetype_manual(name="genotype",values=1:3,labels=c("A","B","C"))+
guides(col="none")

Please do dput(df) and paste the output as part of your question if the df I have above is different from yours. This way others can help as well.

If you need thicker lines, do:

old_geom <- ggplot2:::check_subclass("line", "Geom")$default_aes
update_geom_defaults("line", list(size = 1.5))
p + scale_linetype_manual(name="genotype",values=1:3,labels=c("A","B","C"))+
    guides(col="none")
update_geom_defaults("line", list(size = old_geom$size))