Deep Dive into RPACT Cloud

Where Statistical Innovation Meets Open Source Excellence

Daniel Sabanés Bové, Friedrich Pahlke

RPACT

April 20, 2026

Outline

  • RPACT Cloud
  • Q&A
  • Outlook to Future Developments
  • Screenshots for reference, we will do the live demo

RPACT Cloud

Why RPACT Cloud?

What are the advantages of using RPACT Cloud instead of directly using the R package rpact or other software e.g. PASS or East?

  • Code-based: RPACT Cloud produces R code that can be exported as part of reports, and can then be fully transparently version-controlled, reviewed and audited, and can be used for regulatory submission purposes
  • Web-based: No local installation required on statisticians’ laptops, accessible from any device with internet access, with free demo at rpact.cloud
  • Company installation: Companies can install RPACT Cloud on their own servers, ensuring data security and compliance with internal policies
  • Continuous development: We are constantly developing and improving RPACT Cloud based on user feedback and the latest methodological research
  • Fully validated: License includes installation qualification tests for the GUI (Shiny test suite), adding on top of the rpact package validation
  • Cost-effective: Installation license with a flat rate per company, no per-user fees, and no hidden costs for updates or included support

Do we get the same results?

Does RPACT Cloud give the same results as other software e.g. PASS or East?

  • As part of the development of rpact and therefore also with RPACT Cloud building on top of rpact, we as well as pharma companies have done extensive comparisons with other software, including East and SAS. The results of these comparisons are permanently integrated in the unit tests of rpact as part of the installation qualification test suite. Therefore we can guarantee that we give numerically equal results.
  • Recently an academic group from France did an East/nquery vs. rpact comparison, the results are available here
    • Conclusion: “R provides a credible, transparent, and reproducible alternative to commercial software for many standard sample size calculations in oncology. For most practical two‑arm designs, alignment with East and nQuery is excellent.”
    • They note limitation of rpact for one-arm survival designs as well as binary designs with exact tests. We are aware of these limitations and would be happy to add them to rpact and RPACT Cloud if there is sufficient interest or a corresponding collaboration project

Questions and Answers

Thank you for your questions you sent us in advance!

  • Events projection: Does the product support events projection?
    • Currently not yet, but this could well be added in a collaboration with you
  • Complex adaptive design and Simulation: Does the product support advanced design and simulation functions (e.g. sample size re-estimation, phase II/III seamless design, population enrichment design)
    • Yes, sample size re-estimation is supported via conditional power (start via Plan a Trial > Calculation target > Power simulation)
    • Yes, Phase II/III seamless designs are supported with inverse-normal combination, Fisher combination test as options
    • Population enrichment design will be implemented this year

Questions and Answers (cont’d)

  • Regulatory acceptance?
    • FDA has received training in rpact
    • RPACT Cloud builds on top of the rpact package. Both are fully validated and have comprehensive validation documentation, which are available for customers and can be used for regulatory audit purposes.
    • As mentioned, 4 major pharmas have switched from East to RPACT Cloud. In addition, e.g. Roche has been using only rpact since 2019 for all of their phase II and III group sequential and adaptive designs, and has successfully submitted multiple regulatory applications with rpact-based designs.

Additional FAQ

  • Can you visualize and compare different development plans (including probability of success, expected overall duration until approval, etc.)?
    • This is not yet available in RPACT Cloud, but also this one could well be added in a “clinical development plan” module in a collaboration with you.
  • Are all four major outcome types (binary, continuous, time-to-event, count) supported?
    • Yes, all four major outcome types are supported in RPACT Cloud.
  • Is borrowing from historical control data supported?
    • Borrowing from historical control data is not yet available in RPACT Cloud. A first step in this direction will be the addition of Go/No-go decision criteria also supporting Bayesian designs using MAP (meta-analytic predictive) priors, which will be implemented this year. More advanced borrowing methods could well be added in a collaboration with you.
  • Are multiple endpoints supported, e.g. ORR and OS for oncology trials?
    • Multiple endpoints are not yet supported in RPACT Cloud, but this is also something that could well be added in a collaboration with you.

Future of RPACT Cloud

In the near future you will see in RPACT Cloud:

  • Count data simulations, complementing the currently available sample size and power functionality
  • Enrichment designs, which are currently not yet available and are planned for implementation by the end of 2026
  • Go/No-go decision criteria operating characteristics for Phase 2 trials (frequentist and Bayesian)
  • A gradual transition from an rpact-focused GUI towards a flexible platform for clinical trial design, including statistical methodologies beyond rpact, such as graphical multiple testing procedures
  • Exploration of an integrated AI chat interface, potentially powered by the upcoming “RPACT Code” technology

“RPACT Code” (working title)

In 2026 we are developing “RPACT Code” (working title), a next-generation AI system that helps statisticians design clinical trials using natural language, while ensuring that all results are computed deterministically and reproducibly in R.

Unlike typical AI coding tools, this system will:

  • use real statistical computation (to get correct, instead of hallucinated, results)
  • access our proprietary retrieval-augmented generation (RAG) system over the full set of documentation, source code, tests and validation documentation of our R packages rpact and crmPack
  • combine LLMs, RAG, and structured execution for maximum reliability and efficiency
  • be designed from the start for enterprise and regulated environments

RPACT Cloud – Demo

RPACT Cloud – Basics

RPACT Cloud – Explore Parameter Levels

RPACT Cloud – Settings

RPACT Cloud – Quick Start Menu

RPACT Cloud – Design Tabs

RPACT Cloud – Getting Help

RPACT Cloud – Copy And Paste

RPACT Cloud – Output Options

RPACT Cloud – Reporting

RPACT Cloud – Export

RPACT Cloud – Plotting

RPACT Cloud – Output Table

RPACT Cloud – Design Comparison

Summary

Summary

  • RPACT Cloud is our web-based graphical user interface for rpact, with a focus on ease of use and accessibility
  • We are continuously developing and improving our software based on user feedback and the latest methodological research
  • This year we are developing “RPACT Code” (working title), a next-generation agentic AI system that helps statisticians design clinical trials using natural language, while ensuring that all results are computed deterministically and reproducibly in R

References