# Pre-agreed update rules so an MCED test is not obsolete when its trial reads out

Source: https://onco.cc/ideas/idea-prev-mced-change-control-plan/  
OnCo record `idea-prev-mced-change-control-plan` (Idea). Data CC BY-NC 4.0, attribute "Data from OnCo (onco.cc)"; commercial use needs a licence.

## TL;DR

A cancer blood test improves every year, but a ten-year trial tests the old version. Regulators and sponsors could agree in advance how updates are validated and carried into the result.

## Summary

FDA's Predetermined Change Control Plan for AI devices offers a model. Propose an MCED-specific plan: the classifier may be updated on banked baseline samples under locked-test-set rules, each version must be shown non-inferior on a sequestered set, and the primary analysis uses the enrolment version with a pre-specified sensitivity analysis re-scoring banked plasma.

## Fields

- Kind: Idea
- Last checked: 2026-09-08
- Hypothesis: With banked plasma re-scored by an updated classifier under pre-agreed rules, estimated effect sizes fall within 10% of those from the enrolment version, allowing an updated test to inherit trial evidence.
- Rationale: Without this, each improved assay restarts the evidence clock, which is why the field keeps launching new observational studies instead of finishing randomised ones.
- Proposed test: Apply retrospectively in NHS-Galleri or PATHFINDER 2 banked plasma; publish agreement between classifier versions.
- Maturity: speculative
- Actor: regulator

## Sources

- Bottleneck evidence (The hardest cancers are found late): Crosby et al., Early detection of cancer (Science 2022): https://doi.org/10.1126/science.aay9040

## Connected records

- fronts: [Early Detection & Screening](https://onco.cc/fronts/early-detection/)
- technologies: [DNA methylation profiling](https://onco.cc/technologies/methylation-profiling/), [Multi-cancer early detection (MCED)](https://onco.cc/technologies/mced/)
- companies: [GRAIL](https://onco.cc/companies/grail/)
- trials: [NHS-Galleri](https://onco.cc/trials/nhs-galleri/), [PATHFINDER 2](https://onco.cc/trials/pathfinder-2/)
- bottlenecks: [AI that is built but not validated or deployed](https://onco.cc/bottlenecks/b-ai-validation/), [Regulatory divergence between regions](https://onco.cc/bottlenecks/b-regulatory-fragmentation/), [The hardest cancers are found late](https://onco.cc/bottlenecks/b-early-detection/)
- key papers: [Early detection of cancer](https://onco.cc/key-papers/paper-crosby-science/)

---
JSON: https://onco.cc/api/v1/entities/idea-prev-mced-change-control-plan.json