This vignette covers changes between versions 0.1.0 and 0.2.0.


Hadley Wickham’s ggplot2 version 2.0 revolution, at the end of 2015, triggered many crashes in dependent R packages, that finally led to deletions of few packages from The Comprehensive R Archive Network. It occured that survMisc package was removed from CRAN on 27th of January 2016 and R world remained helpless in the struggle with the elegant visualizations of survival analysis. Then a new tool - survminer package, created by Alboukadel Kassambara - appeared on the R survival scene to fill the gap in visualizing the Kaplan-Meier estimates of survival curves in elegant grammar of graphics like way. This blog presents main features of core ggsurvplot() function from survminer package, which creates the most informative, elegant and flexible survival plots that I have seen!

During the development of RTCGA package (about which I wrote here) we encountered a need to provide the simplest possible interface to estimates of survival curves for biotechnologists and the discovery of ggsurvplot() was a bull’s-eye! Many have tried to provide a package or function for ggplot2-like plots that would present the basic tool of survival analysis: Kaplan-Meier estimates of survival curves, but none of earlier attempts have provided such a rich structure of features and flexibility as survminer. On basis of estimates of survival curves one can infere on differences in survival times between compared groups, so survival plots are very useful and can be seen in almost every publication in the field of survival analysis and time to event models.


After regular installation

biocLite("RTCGA.clinical") # data for examples

we can create simple estimates of survival curves just after we put survfit (survival package) object into ggsurvplot() function. Let’s have a look at differences in survival times between patients suffering from Ovarian Cancer (Ovarian serous cystadenocarcinoma) and patients suffering from Breast Cancer (Breast invasive carcinoma), where data were collected by The Cancer Genome Atlas Project.

survivalTCGA(BRCA.clinical, OV.clinical,
             extract.cols = "admin.disease_code") -> BRCAOV.survInfo
fit <- survfit(Surv(times, patient.vital_status) ~ admin.disease_code,
               data = BRCAOV.survInfo)
# Visualize with survminer
ggsurvplot(fit, data = BRCAOV.survInfo, risk.table = TRUE)