Bayesian Two-Stage Design with Window-Cohort and Controlled Roll-on for Time-to-Event Estimand.

WCRBayesDesign

Bayesian Two-Stage Adaptive Design for Single-Arm Survival Trials

Overview

WCRBayesDesign provides simulation and optimization tools for Bayesian two-stage adaptive single-arm trials with time-to-event endpoints. All statistical calculations are carried out in a transformed-time space induced by a reference survival function $S_0$, while key results are reported on the original time scale for clinical interpretability.

Installation

# Install from source
install.packages("WCRBayesDesign_1.0.0.tar.gz", repos = NULL, type = "source")

Quick Start

library(WCRBayesDesign)

# Step 1 - Search feasible interim-analysis parameters
res_pIA <- find_Nw_pIA(
  tau = 24, theta_L = 0.62, theta_alt = 0.80,
  alpha_target = 0.05, beta_target = 0.20,
  rate = 5/12, X_grid = c(1, 2, 3),
  S0_dist = "weibull",
  S0_par = list(k = 1.2, lambda = 0.08)
)

# Step 2 - Optimise the two-stage design
opt <- two_stage_optimize_design(
  NwX_pIA_results = res_pIA,
  rate = 5/12, tau = 24,
  theta_L = 0.62, theta_alt = 0.80,
  alpha_target = 0.05, beta_target = 0.20,
  nsim = 2000,
  S0_dist = "weibull",
  S0_par = list(k = 1.2, lambda = 0.08),
  optimize = "ESS", ncores = 4
)

# Step 3 - Evaluate operating characteristics
oc <- oc_two_stage(
  N = 50, Nw = 30, X = 2,
  pIA = 0.5, pF = 0.05,
  rate = 5/12, tau = 24,
  theta_L = 0.62, theta_alt = 0.80,
  a0 = 0.01, b0 = 0.01, nsim = 5000,
  S0_dist = "weibull",
  S0_par = list(k = 1.2, lambda = 0.08)
)

Exported Functions

Supported Baseline Distributions

• Exponential (S0_dist = "exp")
• Weibull (S0_dist = "weibull")
• Log-normal (S0_dist = "lognormal")
• Log-logistic (S0_dist = "loglogistic")
• User-supplied (S0_dist = "custom")

License

GPL-3

Authors

• Zhongheng Cai (maintainer) — [email protected]
• Haitao Pan — [email protected].