This is a test version of Biostars. For the public version, visit https://www.biostars.org.
A error in drawing nomogram for multivariate cox regression with rms package

I kept getting the following error when i tried to draw a nomogram with rms package:

Error in lims[[i]] : subscript out of bounds

I've eliminated all the missing data from the dataset and transformed all the variables into continuous variables, but nothing really changed. I'd appreciate it if my question is answered.

regression nomogram. r cox rms
• 3,531 views • link
updated by GenoMax GenoMax • written by Dude • Dude

Please improve your post by adding your code and example of your data. As it stands we can't help you.

• 0 views • link
written by Basti • Basti

Oh sorry, I just uploaded the code and data you might need. Thank you so much!!

• 0 views • link
written by Dude • Dude

1 answer

The code I used is listed below:

fit_cox <- coxph(Surv(surviving_days,vital == 1) ~ gender + ajcc_pathologic_stage + risk_factor, data = clinical)
fit_cox <- rms::cph(Surv(surviving_days,vital == 1) ~ gender + stages + risk_factor, data = clinical)
summary(fit_cox)
Survival1 <- function(x)Survival(365, x)
Survival2 <- function(x)Survival(730, x)
nomo <- nomogram(fit_cox, fun = list(Survival1, Survival2),
                 fun.at = c(0.05,  seq(0.1, 0.9, by = 0.05), 0.95),
                 funlabel = c("1 year survival", "2 year survival"))
plot(nomo)

the data to construct fit_cox was uploaded in the link below:

https://github.com/Datapioneer/Question.git

thank you!

• 0 views • link
written by Dude • Dude

When calling rms() you indicated stages wheras there is only a variable called ajcc_pathologic_stage

• 0 views • link
written by Basti • Basti

Yes, I just noticed it. But when I fit cox model with the corrected rms::cph() command, there was no error message, but when I ran the command "summary(fit_cox)", I got amessage like this :

Error in summary.rms(fit_cox) : adjustment values not defined here or with datadist for gender ajcc_pathologic_stage risk_factor

and then I kept running down to the "nomogram()" function, and I got a message like this:

Error in value.chk(at, i, NA, -nint, Limval, type.range = "full") :
variable gender does not have limits defined by datadist

Another question is there any problem with fitting the model with "coxph()" function? I always get the message of "Error in lims[[i]] : subscript out of bounds" when running nomogram.

Thank you.

• 0 views • link
written by Dude • Dude

Hi it might be late to answer, but add surv = T inside rms::cph(), and it should run fine then : like:

fit_cox <- rms::cph(Surv(surviving_days,vital == 1) ~ gender + stages + risk_factor, data = clinical, surv = T)
• 0 views • link
updated by GenoMax GenoMax • written by StartR • StartR

Log in to answer this question.

More posts like this