Extending ggsurvplot() Capabilities

[kassambara]

Capabilities of the current version:

• plots curves and survival tables
• computes and displays pvalue
• computes and displays survival medians

New features to be added:

• Supports a list of survfit objects: ggsurvplot(fit.list) ==> Returns a list of ggsurvplot. see this #197

• Plots survival curves for a data set grouped by one or two grouping variables. For example one might like to plot the survival curves of patients treated by drug A vs patients treated by drug B in a dataset grouped by TP53 mutation and/or RAS mutation:

fit <- survfit(Surv() ~ treatmeent)
ggsurvplot(fit, group.by = "tp53.mutation")

Returns a list of ggsurplots. One can (optionally) arrange ggsurvplots using arrange_ggsurvplots()

• Supports faceting. Computes and displays specific pvalue for each panel: ggsurvplot(fit, facet.by = "tp53.mutation"). see this 190

• New option 'add.all' . One might like to plot survival curves of patients stratified by grouping variables, and to add on the same plot, the survival curves of 'all' patients (NULL model): ggsurvplot(fit, add.all = TRUE). User request at #98

• Plotting survival curves from a data frame containing survival curves summary.

• Combine multiple survfits object on the same plot. User request at: https://github.com/kassambara/survminer/issues/195

pfs <- survfit( Surv(pfs.time, pfs.status) ~ 1, data = data)
os <- survfit( Surv(os.time, os.status) ~ 1, data = data)
fit <- list(PFS = pfs, OS = os)
ggsurvplot(fit, combine = TRUE)

[kassambara]

The new ggsurvplot function would look like this:

ggsurvplot(fit, data, color = NULL, palette = NULL, group.by = NULL, 
    facet.by = NULL, combine = NULL, add.all = NULL, .....){

    if(is.list(fit))
      ggsurvplot_list(fit,  data...)
      else if(!is.null(group.by))
      ggsurv_plot(fit, data, group.by, ...)
      .......

    }

To keep things simple, we need a family of ggsurvplot functions:

• ggsurvplot_core(fit, data): the core function used by the other functions
• ggsurvplot_list(fit = list(fit1, fit2), data )
• ggsurvplot_facet(fit, data, facet.by)
• ggsurvplot_groupby(fit, group.by)
• ggsurvplot_addAll(fit, data)
• ggsurvplot_combine(fit = list(fit1, fit2))

These functions will be internally called by the main function ggsurvplot() depending on the situation.

Other small functions needed:

• surv_group_by(): Create a Grouped Dataset for survival analysis.
• surv_fit(): Create Survival Curves. Accepts a list of formula and/or data. wrapper arround survfit
• surv_pvalue(fit, method = "survdiff"). Returns the pvalue of a fit. Accepts also a list of fits
• surv_median(fit). Returns the median of survival curves. Accepts also a list of fits

[kassambara]

We need a more generic function surv_plot(), that will be a wrapper around the ggsurvplot_*() family functions.

The final list would be:

1. For visualization:

• surv_plot()
• ggsurvplot(fit, data): the core function used by the other ggsurvplot_*() functions
• ggsurvplot_list(fit = list(fit1, fit2), data )
• ggsurvplot_facet(fit, data, facet.by)
• ggsurvplot_group_by(fit, group.by)
• ggsurvplot_add_all(fit, data)
• ggsurvplot_combine(fit = list(fit1, fit2))

2. For fitting and describing

• surv_group_by(): Create a Grouped Dataset for survival analysis.
• surv_fit(): Create Survival Curves. Accepts a list of formula and/or data. Accepts a grouped data sets. Wrapper arround survfit. Returns a list of survfit objects when inputs are a list of formula and/or data frame.
• surv_pvalue(fit, method = "survdiff"). Returns the pvalue of a fit. Accepts also a list of fits. Returns a list of p-value when the input is a list of fits.
• surv_median(fit). Returns the median of survival curves. Accepts also a list of fits. Returns a list when input is a list.

[kassambara]

ggsurvplot() help page reorganized and new functions added: New functions added:

• surv_group_by()
• surv_fit()
• surv_median()
• surv_pvalue()
• ggsurvplot_df()
• ggsurvplot_list()
• ggsurvplot_group_by()
• ggsurvplot_add_all()
• ggsurvplot_combine()
• ggsurvplot_facet_by()
• ggsurvtable()