survminer

The survminer package is organized into several functional categories for survival analysis:

Survival Curves

• ggsurvplot(): Draws survival curves with 'number at risk', cumulative events, and cumulative censored tables.
• arrange_ggsurvplots(): Arranges multiple ggsurvplots on a single page.
• ggsurvevents(): Plots event time distributions.
• ggsurvparametric(): Overlays parametric survival fits (from survreg or flexsurv) on Kaplan-Meier curves.
• surv_summary(): Creates a data frame summary from survfit results.
• surv_cutpoint(): Determines optimal cutpoints for continuous variables based on survival significance.
• surv_median_followup(): Computes median follow-up time using the reverse Kaplan-Meier method.
• pairwise_survdiff(): Performs multiple pairwise comparisons of survival curves with corrections for multiple testing.
• weighted_logrank(): Performs weighted (Fleming-Harrington) log-rank tests.

Modern Effect Measures

• ggrmst() / ggrmst_difference(): Restricted Mean Survival Time (RMST) visualization with area shading and difference annotation.
• gglandmark(): Landmark analysis to avoid immortal-time bias.
• ggmilestone(): Milestone survival at specific timepoints.

Diagnostics of Cox Model

• ggcoxzph(): Graphical test of proportional hazards (wrapper for plot.cox.zph()).
• ggcoxdiagnostics(): Displays goodness-of-fit graphs for Cox Proportional Hazards models.
• ggcoxfunctional(): Plots continuous explanatory variables against martingale residuals to help choose functional forms.
• ggcoxnph(): Non-proportional-hazards diagnostic panel.

Summary of Cox Model

• ggforest(): Draws forest plots for CoxPH models.
• ggforest_models(): Compares covariate hazard ratios across multiple Cox models.
• ggforest_subgroup(): Subgroup forest plots showing treatment hazard ratios and interaction p-values.
• ggadjustedcurves(): Plots adjusted survival curves for coxph models.

Competing Risks

• ggcompetingrisks(): Plots cumulative incidence curves.

Clinical Trial Data

• surv_adtte(): Prepares CDISC ADaM time-to-event (ADTTE) datasets.

The transition from version 0.5.2 to 1.0.0 is a major update that consolidates new functionality. While it introduces significant new features, it is designed to be backward compatible by keeping existing function defaults unchanged.

Key new features include:

• Publication Presets: Use ggsurvplot(preset = "publication" | ...) along with theme_surv_classic(), theme_surv_minimal(), or theme_surv_bold() for quick, high-quality plots.
• Modern Effect Measures: New functions for restricted mean survival time (ggrmst(), ggrmst_difference()), milestone survival (ggmilestone()), and landmark analysis (gglandmark()).
• Cox-model Tools: Subgroup and multi-model forest plots (ggforest_subgroup(), ggforest_models()), non-proportional-hazards diagnostic panels (ggcoxnph()), and enhanced ggadjustedcurves() with hazard-ratio annotations and number-at-risk tables.
• Enhanced Plotting: Composable single-ggplot returns via ggsurvplot(output = "ggplot"), parametric-over-KM overlays (ggsurvparametric()), and plotmath support in legend.labs.
• Specialized Helpers: CDISC ADTTE helper (surv_adtte()), batch weighted-log-rank wrapper (weighted_logrank()), and reverse-KM median follow-up (surv_median_followup()).

The customize_labels() helper function allows you to apply specific font styles (size, face, color) to various components of the ggsurvplot object, including the main plot, the risk table, and the ncensor plot. The ggsurvplot object is a list containing individual ggplot objects (e.g., $plot, $table, $ncensor.plot) which can be modified using standard ggplot2::labs() calls.

# 1. Modify labels using standard ggplot2 labs() on the list components
ggsurv$plot <- ggsurv$plot + labs(
  title    = "Survival curves",                     
  subtitle = "Based on Kaplan-Meier estimates",  
  caption  = "created with survminer"             
)

# 2. Apply global font styles using customize_labels()
# Format for font arguments: c(size, face, color)
ggsurv <- customize_labels(
  ggsurv,
  font.title    = c(16, "bold", "darkblue"),         
  font.subtitle = c(15, "bold.italic", "purple"), 
  font.caption  = c(14, "plain", "orange"),        
  font.x        = c(14, "bold.italic", "red"),          
  font.y        = c(14, "bold.italic", "darkred"),      
  font.xtickslab = c(12, "plain", "darkgreen")
)

Use xlim and break.time.by to adjust the X-axis view without affecting the underlying survival estimates. For the risk table, risk.table.y.text.col = TRUE colors the text annotations, and risk.table.y.text = FALSE replaces text names with bars in the legend of the risk table.

ggsurvplot(
   fit,                     # survfit object
   data = lung,             # data used to fit
   risk.table = TRUE,      # show risk table
   pval = TRUE,             # show p-value of log-rank test
   conf.int = TRUE,         # show confidence intervals
   xlim = c(0,500),         # present narrower X axis
   xlab = "Time in days",   # customize X axis label
   break.time.by = 100,     # break X axis in time intervals
   ggtheme = theme_light(), # customize plot and risk table with a theme
   risk.table.y.text.col = T, # colour risk table text annotations
   risk.table.y.text = FALSE # show bars instead of names in legend of risk table
)

For maximum customization, ggsurvplot() allows adding a number of censored subjects plot (ncensor.plot) and median survival pointers (surv.median.line).

ggsurv <- ggsurvplot(
           fit,                     
           data = lung,             
           risk.table = TRUE,       
           pval = TRUE,             
           conf.int = TRUE,         
           palette = c("#E7B800", "#2E9FDF"),
           xlim = c(0,500),         
           xlab = "Time in days",   
           break.time.by = 100,     
           ggtheme = theme_light(), 
           risk.table.y.text.col = T,
           risk.table.height = 0.25, 
           risk.table.y.text = FALSE,
           ncensor.plot = TRUE,      # plot the number of censored subjects at time t
           ncensor.plot.height = 0.25,
           conf.int.style = "step",  # customize style of confidence intervals
           surv.median.line = "hv",  # add the median survival pointer.
           legend.labs = 
             c("Male", "Female")    
        )
ggsurv

You can highly customize ggsurvplot() using several parameters to control the visual appearance of the survival curve and its associated tables:

• size: Change line size.
• palette: Custom color palettes (vector of colors).
• conf.int: Add confidence interval (TRUE/FALSE).
• pval: Add p-value of log-rank test (TRUE/FALSE).
• risk.table: Show risk table (TRUE/FALSE).
• risk.table.col: Color risk table by strata.
• legend.labs: Change legend labels.
• risk.table.height: Adjust the height of the risk table.
• ggtheme: Change the ggplot2 theme (e.g., theme_bw(), theme_light()).
• xlim: Set X axis limits (does not affect survival estimates).
• xlab: Customize X axis label.
• break.time.by: Set time intervals for X axis breaks.
• risk.table.y.text.col: Color risk table text annotations.
• risk.table.y.text: If FALSE, shows bars instead of names in the risk table legend.
• ncensor.plot: Plot the number of censored subjects at time t.
• ncensor.plot.height: Height of the censor plot.
• conf.int.style: Customize confidence interval style (e.g., "step").
• surv.median.line: Add median survival pointer ("hv").

ggsurvplot(
  fit, 
  data = lung, 
  size = 1,                 
  palette = c("#E7B800", "#2E9FDF"),
  conf.int = TRUE,          
  pval = TRUE,              
  risk.table = TRUE,        
  risk.table.col = "strata",
  legend.labs = c("Male", "Female"),    
  risk.table.height = 0.25, 
  ggtheme = theme_bw(),
  xlim = c(0,500),
  xlab = "Time in days",   
  break.time.by = 100,
  risk.table.y.text.col = TRUE,
  risk.table.y.text = FALSE
)