predictTTE

predictTTE predicts time-to-event milestones from interim ADTTE data.

What it does

• Fits parametric enrollment, event, and dropout models.
• Simulates future event trajectories with tumor assessment windows.
• Returns:
  • predicted dates (with confidence intervals) for target event counts;
  • predicted event counts (with confidence intervals) for target dates;
  • an interactive cumulative event plot with prediction-start and cutoff markers;
  • a CI-ribbon plot for target-date event count predictions.

Installation

if (!requireNamespace("devtools", quietly = TRUE)) {
    install.packages("devtools")
}
devtools::install_github(
  "YuquanW/predictTTE",
  build_vignettes = TRUE,
  dependencies = TRUE
)

If the package was already installed without vignettes, reinstall it with build_vignettes = TRUE, then run:

vignette(package = "predictTTE")

Example

library(predictTTE)

# Use packaged auto-loaded example data (.rda)
data <- example_adtte

res <- predict_tte(
  data = data,
  planned_total_n = 300,
  target_events = c(200, 250),
  target_dates = as.Date(c("2026-02-01", "2026-05-01", "2026-08-01")),
  nsim = 1000,
  assessment_intervals_days = c(40, 80),
  assessment_cut_days = c(365.25),
  enrollment_model = "exponential",
  event_model = "exponential",
  dropout_model = "exponential",
  ci_level = 0.95,
  fixed_parameters = TRUE,
  seed = 123
)

res$pred_dates
res$pred_event_counts
res$plot
res$pred_event_plot

# one of target_events / target_dates can be NULL
# (but not both NULL)

Project structure

• R/: package functions
• data/: example ADTTE dataset (.rda, auto-loaded)
• inst/extdata/: anonymized reference CSV files
• tests/testthat/: unit tests
• vignettes/: long-form package guide