# Let the trial learn: response-adaptive allocation across many combination arms

Source: https://onco.cc/ideas/idea-tr2-bandit-allocation/  
OnCo record `idea-tr2-bandit-allocation` (Idea). Data CC BY-NC 4.0, attribute "Data from OnCo (onco.cc)"; commercial use needs a licence.

## TL;DR

As results come in, the trial sends more new patients to the arms that are working and fewer to those that are not, so more people benefit and bad arms die faster.

## Summary

Multi-armed bandit and Bayesian response-adaptive randomisation, used in I-SPY 2 and in the REMAP-CAP platform during COVID-19, allocate patients towards the arms that are performing best while maintaining control of false positives. With ten or more combination arms, adaptive allocation shortens time to identify winners and reduces the number of patients on futile arms. Regulators accept these designs with pre-specified simulation of operating characteristics.

## Fields

- Kind: Idea
- Last checked: 2026-09-08
- Hypothesis: Adaptive allocation across ten combination arms identifies the best arm with 30% fewer patients than equal randomisation at the same error rates, in simulation and in a live platform.
- Rationale: REMAP-CAP found effective COVID-19 treatments faster than fixed designs. Oncology platforms have been slower to adopt because of endpoint latency; ctDNA or pathological response as intermediate endpoints removes that obstacle.
- Proposed test: Implement in a neoadjuvant platform using pathological response as the adaptive endpoint; compare patients-to-decision with a fixed-randomisation shadow analysis.
- Maturity: early-clinical
- Actor: research

## Sources

- Bottleneck evidence (Too many combinations to test): Palmer & Sorger, Combination cancer therapy can confer benefit via patient-to-patient variability without drug additivity or synergy (Cell 2017): https://doi.org/10.1016/j.cell.2017.11.009

## Connected records

- ideas: [Kill combination arms early using circulating tumour DNA, before waiting for scans](https://onco.cc/ideas/idea-tr2-ctdna-futility-gates/), [Pre-surgery platform trials that test combinations on pathological response in months](https://onco.cc/ideas/idea-tr2-neoadjuvant-combo-platform/)
- terms: [Basket, umbrella, and platform trials](https://onco.cc/terms/basket-umbrella-platform/), [Pathologic complete response (pCR)](https://onco.cc/terms/pcr/)
- bottlenecks: [Too many combinations to test](https://onco.cc/bottlenecks/b-combination-space/), [Trial design, endpoints and cost](https://onco.cc/bottlenecks/b-trial-design/)
- key papers: [Combination Cancer Therapy Can Confer Benefit via Patient-to-Patient Variability without Drug Additivity or Synergy](https://onco.cc/key-papers/paper-palmer-cell/)
- roadmaps: [Trial modernisation roadmap: the randomised trial → platforms and adaptive designs → decentralised, pragmatic and always-on](https://onco.cc/roadmaps/trial-modernisation-roadmap/)

---
JSON: https://onco.cc/api/v1/entities/idea-tr2-bandit-allocation.json