rk.survival: Modern Survival Analysis for RKWard

rk.survival upgrades the survival analysis capabilities of RKWard by integrating the powerful survival and survminer packages. It allows users to perform rigorous survival analysis and generate publication-quality, ggplot2-based visualizations (Kaplan-Meier curves, Forest plots, and Diagnostic plots) via a simple GUI.

Features

This package installs a new submenu in RKWard: Analysis > Survival Analysis (survminer).

• Kaplan-Meier Analysis:

  • Fit survival curves (survfit) and plot them instantly (ggsurvplot).
  • Visuals: Add Risk Tables, P-Values (Log-Rank), Confidence Intervals, and Censoring marks with a click.
  • Themes: Support for scientific journals (JCO, Lancet, etc.).

• Cox Proportional Hazards:

  • Fit multivariate Cox models (coxph).
  • Forest Plot: Automatically generate a ggforest plot to visualize Hazard Ratios.

• Cox Diagnostics:

  • Test the Proportional Hazards assumption using Schoenfeld residuals (cox.zph).
  • Visuals: Plot residuals to identify time-varying effects.

Requirements

1. A working installation of RKWard.
2. The following R packages:

   install.packages(c("survival", "survminer", "ggplot2"))
   

3. The R package devtools (for installation from source).

Installation

1. Open R in RKWard.
2. Run the following commands in the R console:

local({
## Preparar
require(devtools)
## Computar
  install_github(
    repo="AlfCano/rk.survival"
  )
## Imprimir el resultado
rk.header ("Resultados de Instalar desde git")
})

3. Restart RKWard to ensure the new menu items appear correctly.

Usage & Examples

To test these plugins, we will use the standard lung cancer dataset included in R.

Step 0: Prepare Data

Run this code in the RKWard console to load the data and label the factors (so the plots look nice):

library(survival)
data(lung)

# Convert 'sex' from 1/2 to Male/Female labels
lung$sex <- factor(lung$sex, levels = c(1, 2), labels = c("Male", "Female"))

Example 1: Kaplan-Meier Curve

Compare survival rates between Men and Women.

1. Navigate to Analysis > Survival Analysis (survminer) > Kaplan-Meier Analysis.
2. Time Variable: Select time.
3. Status Variable: Select status.
4. Grouping Variable (Strata): Select sex.
5. Options: Check "Show Risk Table" and "Show P-Value".
6. Click Submit.
   • Result: A professional survival curve showing that Females have a significantly better survival probability (p < 0.01), with a table of "Number at risk" below the x-axis.

Example 2: Cox Proportional Hazards

Analyze the effect of Age, Sex, and Performance Score on survival.

1. Navigate to Analysis > Survival Analysis (survminer) > Cox Proportional Hazards.
2. Time Variable: Select time.
3. Status Variable: Select status.
4. Covariates: Select age, sex, and ph.ecog (ECOG performance score).
5. Options: Ensure "Generate Forest Plot" is checked.
6. Save Model as: cox_model.
7. Click Submit.
   • Result:
     • Text: A summary table of coefficients.
     • Graph: A Forest Plot showing Hazard Ratios. Note that sexFemale is to the left of the line (HR < 1), indicating a protective effect.

Example 3: Model Diagnostics

Check if the "Proportional Hazards" assumption holds true for our model.

1. Navigate to Analysis > Survival Analysis (survminer) > Cox Diagnostics.
2. Select Cox Model: Choose the cox_model created in the previous step.
3. Click Submit.
   • Result: Three plots (one for each variable: age, sex, ph.ecog). If the red trend line is roughly horizontal, the assumption holds. If it slopes significantly, the effect of that variable changes over time.

Author

Alfonso Cano Robles

• Email: alfonso.cano@correo.buap.mx

Assisted by Gemini, a large language model from Google.