survival analysis of tcga clinical data

Question

[akutasame]

Hi all,

I want to do survival analysis of clinical data but I am not sure about censoring TCGA clinical data? I have ''daysto_lastfollowup", "daysto_death", "vital status" and "overall_survival_months". What I did is

Method 1:

daysToEvent <- rep(NA, nrow(tcgaClinical))
daysToEvent[vitalStatus == "Alive"] <- daysToLastFollowup[vitalStatus == "Alive"]
daysToEvent[vitalStatus == "Dead"] <- daysToDeath[vitalStatus == "Dead"]
eventStatus <- rep(NA, nrow(tcgaClinical))
eventStatus[vitalStatus == "Alive"] <- 1
eventStatus[vitalStatus == "Dead"] <- 0

tcgaOS <- Surv(daysToEvent/30, eventStatus == 0)
rownames(tcgaOS) <- rownames(tcgaClinical)

Or

Method 2:

eventStatus <- rep(NA, nrow(tcgaClinical))
eventStatus[vitalStatus == "Alive"] <- 1
eventStatus[vitalStatus == "Dead"] <- 0

tcgaOSM <- Surv(tcgaClinical$OS_MONTHS, eventStatus == 0)
rownames(tcgaOSM) <- rownames(tcgaClinical)

Which method makes more sense ? Thank you

[akutasame]

That was helpful, Thank you

Answer 1

[MatthewP]

Hello, I recommend you download clinical data from cBioPortal instead of TCGA/GDC. If you download clinical data from cBioPortal you will see fields Overall Survival (Months) and Overall Survival Status thats what you need for OS(Overall survival) analysis.

As OS the event is dead, so alive is censored data, you should give value 1 to dead and 0 to alive. However I recommend use 1 and 2, give value 2 to dead and 1 to alive.

Here, some tutorial may be useful for you:
https://www.emilyzabor.com/tutorials/survival_analysis_in_r_tutorial.html#introduction
https://www.datacamp.com/community/tutorials/survival-analysis-R
http://www.sthda.com/english/wiki/survival-analysis-basics