R help - Jun 2024 - unable to get barchart of censored subjects from ggsurvplot() in survminer package, if there is no predcitor

Hello. Running R 4.2.3 on Windows 10. Using survimer package version
survminer_0.4.9 published 2021-03-09.

I'm encountering an error with ggsurvplot() in the survminer package.
Email to the author/maintainer about 2 weeks ago has not yet resulted in
a reply.

I seem unable to produce a n.censor.table for
an unconditional (no predictor) survival curve. See code below.

Any advice?

Thanks.

--Chris Ryan, MD, MS, MSPH


library(survival)
library(survminer)
data(lung)

fit <- survfit(Surv(time, status) ~ 1, data = lung)


ggsurvplot(
           fit,
           data = lung,
           xlab = "Time in days",
           ncensor.plot = TRUE
                  )

## produces error "Error:
! Problem while converting geom to grob.
? Error occurred in the 1st layer.
Caused by error:
! Unknown colour name: strata"
about "uknown colour name strata"
## but there are no strata



## omit ncensor.plot, and no error

ggsurvplot(
           fit,
           data = lung,
           xlab = "Time in days",
       #   ncensor.plot = TRUE
                  )

## restore ncensor.plot = TRUE, but use a predictor in the model,
## and no error.

fit.2 <- survfit(Surv(time, status) ~ sex, data = lung)
ggsurvplot(
           fit.2,
           data = lung,
           xlab = "Time in days",
           ncensor.plot = TRUE
                  )

? Mon, 17 Jun 2024 15:53:16 -0400
"Christopher W. Ryan via R-help" <r-help at r-project.org>
?????:
> Caused by error:
> ! Unknown colour name: strata"
> about "uknown colour name strata"
> ## but there are no strata
rlang::last_trace() demonstrates that the error happens because some
code receives a string 'strata' and tries to decode it as a name of a
colour:

farver::decode_colour('strata')
# Error: Unknown colour name: strata

This appears to originate from the first layer of the ncensor.plot:

p <- ggsurvplot(
 fit,
 data = lung,
 xlab = "Time in days",
 ncensor.plot = TRUE
)
p$ncensor.plot$layers[[1]]$aes_params
# $colour
# [1] "strata"
# 
# $fill
# [1] "strata"

This can be traced further but I'm not sure how to fix the bug without
strata. Giving an explicit argument color = 'black' to ggsurvplot()
results in a warning but at least doesn't raise an error while drawing
the plot.

-- 
Best regards,
Ivan