rewirebio.iobenchmarks

Use caseDNA and genomesResearch and clinical research

Select a plasma ctDNA fragmentomics detection workflow

Which fragmentomics workflow detects tumour-derived plasma DNA under realistic low tumour fractions and fixed false-positive constraints?

In this release 1 evaluated endpoint · 1 tested configuration · Proxy evidence only · 2 recorded evidence gaps

Question and applicability

Inspect the DELFI repeated-cross-validation sensitivity-at-fixed-specificity evidence before selecting a workflow and validation design for the intended screening or diagnostic population.

Who this is for
  • Clinical researchers scoping the available diagnostic-genomics evidence
  • Computational researchers comparing exact evaluated configurations
Research setting

Research and clinical research. See Clinical research scope for what the evidence does not establish.

Biological setting
The DELFI stochastic gradient boosting classifier (GC-corrected total and short fragment coverage across 504 genomic bins, 39 arm Z-scores, mitochondrial representation) achieves 73% (152 of 208) sensitivity at a reported 98% specificity, under repeated 10-fold cross-validation (10 repeats) across 208 cancer patients (seven cancer types) and 215 healthy individuals.

Outside this use case

  • A combined mutation+DELFI configuration (115/126, 91%), a different subset/configuration not ingested here
  • Tumour-fraction-stratified sensitivity (unreported for this endpoint)

Inputs and expected output

Inputs you need

  • Plasma cell-free DNA whole-genome sequencing reads
  • The target specificity (false-positive) constraint for the intended use

Expected output

A sourced sensitivity figure at one reported specificity, from internal cross-validation; no prospective screening-validation claim.

Clinical research scope

Clinical applicability is not established as prospective screening performance: this is internal repeated cross-validation on one assembled cohort, not a separately held-out validation cohort or a prospective screening trial, and the cohort's clinically identified cancers/healthy comparators differ from an intended screening population.

Evaluated evidence

Evidence is grouped by its protocol. Relevance refers to the stated endpoint and context; it is separate from clinical validation and from the review method. Limits specific to each evaluation are listed with it.

DELFI 2019 internally cross-validated cancer detection protocol

Current source-reviewed mapping

Proxy evidence: transfer to this question is limited

A repeated 10-fold cross-validation (10 repeats) on one assembled cohort is a proxy for the broader plasma ctDNA fragmentomics detection decision and for this explicitly assessed assay endpoint; it is internal cross-validation, not a separately held-out validation cohort and not prospective screening validation.

Assessed endpoint
Sensitivity (73%, 152/208) at a reported 98% specificity for plasma cfDNA fragmentomics-based cancer detection, repeated 10-fold cross-validation (10 repeats)
Evaluation protocol
DELFI 2019 internally cross-validated cancer detection protocol
Computational task
A reviewed task relationship is not recorded for this protocol.
Input and population constraints
  • Inspect every linked evaluation's source locator and preserved conflicts before citing a result.
  • Do not combine this mapping's evaluations with any other protocol's results.

Limits on interpretation

  • 4 of 215 healthy individuals misclassified at source-labelled 98% specificity. Retain rounded printed specificity; do not recompute sensitivity.
  • Internal repeated cross-validation is not independent prospective screening validation; clinically identified cancers and healthy comparators differ from intended screening population.
  • DELFI composite features include CNAs/mtDNA, so this exact result is not pure fragmentation alone.
  • cfDNA signal is not purified ctDNA, tumour fraction limit is unreported for this endpoint.
  • Combined mutation+DELFI 115/126 (91%) is a different subset/configuration and is excluded.

Automated source review · 2026-10-07 · Claude Sonnet AMP-integration worker, bounded transcription of Codex-checked primary values; independently reviewed by Codex (workbench/amp-supervision/primary-review.md, integration-review-corrections.md)

Bounded primary-source transcription, independently Codex-checked. No new model execution, independent experimental reproduction, qualified human scientific review or clinical validation.

Evaluated configurations

Each configuration below belongs to this protocol. Inspect its inputs, population and scoring conditions before comparing it with another evaluation.

DELFI 2019 internally cross-validated cancer detection tested configuration

Author-reported evaluation · Source checked

DELFI 2019 internally cross-validated cancer detection; bounded primary-source candidate.

Inspect results, conditions and reproduction (1 recorded result)
Population and split
208 patients with seven cancer types and 215 healthy individuals used for classifier; broader study abstract 236 cancers/245 healthy is not classifier denominator. · 10-fold cross-validation repeated 10 times; feature selection and model estimation on training folds only
Inputs and adaptation
Plasma cfDNA whole-genome fragmentation and copy-number features; composite cfDNA includes background and tumour-derived signal · Supervised cohort-specific gradient boosting; not a foundation model
Evaluation budget
Not reported
Runtime and memory

Runtime and memory measurements are not reported in this evaluation. A study budget is not a runtime or memory measurement.

Recorded results for this configuration
MetricValueCoverageUncertainty and source
sensitivity_at_reported_98_percent_specificity73%

percent · higher

Not reported scored / Not reported eligible

95% CI 67%–79%; sensitivity intervals from 2,000 bootstrap replicates

Result provenance

Evaluation methods, evidence and reproduction

No execution recipe has been verified for this exact configuration and evaluation. Inspect its methods and original run documentation before attempting reproduction.

Open the protocol's results and comparison checks →

Mapping sources and review metadata

Mapping use-case-mapping-amp-20261007-issue15 · revision 1

Add Codex-checked primary-source protocol evidence from the bounded AMP intake (rewire.it#365).

Reviewed evidence fingerprint cd4db57e178c0edf3c49a8cbf8877bc769168345ab18d7cee3bbd32ee831f892

Limitations and missing evidence

These gaps apply to the question as a whole. Absence of evidence is not a zero score.

  • 4 of 215 healthy individuals were misclassified at the source-labelled 98% specificity; the printed specificity is retained rather than recomputed.
  • cfDNA signal reflects total circulating DNA, not purified ctDNA; tumour-fraction limits are unreported for this endpoint.

Planned work

These plans do not contribute measured results or evaluated winners above.

Contribute evidence or propose a correction

Evidence collection plan

Collecting evidence

Mapped evidence already covers 1 evaluated endpoint, in the evaluated evidence above. The status above describes only the specific comparison in this plan, which remains open; it does not mean no evidence has been collected. The plan defines a comparison to investigate; it does not establish model performance or suitability.

Comparison question

Which fragmentomics workflow detects tumour-derived plasma DNA under realistic low tumour fractions and fixed false-positive constraints?

Baselines, outcomes and validation requirements

Baselines to include

  • The conventional/author-introduced workflow measured in the linked primary source(s).

Outcomes to measure

  • The declared endpoint in this use case's active mapping(s); see evidence_gaps for what remains open.

Validation requirements

  • Independent held-out population matched to the intended clinical setting.
  • Qualified human scientific review before any clinical-validation claim.

Next collection task

A dedicated rewire-benchmarks protocol/run task with an independent, prospectively collected held-out validation cohort.

Sources and review

Automated source review · 2026-10-07 · Claude Sonnet AMP-integration worker, bounded transcription of Codex-checked primary values; independently reviewed by Codex (workbench/amp-supervision/primary-review.md, integration-review-corrections.md)

Bounded primary-source transcription, independently Codex-checked. No new model execution, independent experimental reproduction, qualified human scientific review or clinical validation.

Release provenance and downloads

Release 2026-10-07-1448159e6a81

Use-case input digest d60fd7f669bfec7bd34ec6e5080d8e1cb4e8186f8286ead60888848f1e20e001

Download questions, collection plans and review metadata (JSON) · Verify release checksums

Question use-case-plasma-ctdna-fragmentomics. Any numerical results on this page come from this release's existing evaluation records.