Small somatic variant classifier with matched white-blood-cell background removal (GRAIL prototype), CCGA substudy 1

Overview

Small somatic variant classifier with matched white-blood-cell background removal (GRAIL prototype): As the SNV classifier, after removing variants also found in the participant's white blood cells (clonal haematopoiesis).

Consult the linked sources for architecture or protocol details. Missing evidence is not evidence of a missing capability.

Evaluations and results

3 evaluations · 3 results. Different protocols are not a single leaderboard.

Sorted by Accuracy (cancer signal origin among 127 jointly detected validation cancers) (higher is better). The best value in each column is highlighted. Decimals are rounded for display; each value links to the printed value and its source.

Tested configuration	Protocol and dataset	Accuracy (cancer signal origin among 127 jointly detected validation cancers)	Sensitivity at 98% specificity (training set, post hoc 98% specificity threshold)	Sensitivity at 98% specificity (validation set, post hoc 98% specificity threshold)
Small somatic variant classifier with matched white-blood-cell background removal (GRAIL prototype), CCGA substudy 1Evaluation	CCGA substudy 1 validation set: cancer signal origin accuracy among jointly detected cancers, CCGA substudy 1 validation cancers detected by all three representative classifiers (127)	35%	Not reported	Not reported
Small somatic variant classifier with matched white-blood-cell background removal (GRAIL prototype), CCGA substudy 1Evaluation	CCGA substudy 1 training set: cancer signal sensitivity at 98% specificity under 10-fold cross-validation, CCGA substudy 1 training set (1,414 analysable participants)	Not reported	36% (33%–39%)	Not reported
Small somatic variant classifier with matched white-blood-cell background removal (GRAIL prototype), CCGA substudy 1Evaluation	CCGA substudy 1 validation set: cancer signal sensitivity at 98% specificity, CCGA substudy 1 validation set (847 analysable participants)	Not reported	Not reported	33% (29%–38%)

All 3 result rows with coverage, uncertainty and sources

Filter evaluations

Tested model or methodProtocolDatasetEvaluation setupMetricEvidence origin

Applied filters: All linked evaluations

Exact evaluated configurations and original reported results
Tested configuration	Protocol and dataset	Finding	Evidence and details
Configuration: Small somatic variant classifier with matched white-blood-cell background removal (GRAIL prototype), CCGA substudy 1	Protocol: CCGA substudy 1 validation set: cancer signal origin accuracy among jointly detected cancers Dataset: CCGA substudy 1 validation cancers detected by all three representative classifiers (127)	35% Accuracy (cancer signal origin among 127 jointly detected validation cancers) percent · higher Uncertainty: Not reported by the source Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source SNV-WBC on CCGA substudy 1 validation set: cancer signal origin accuracy among jointly detected cancers ctdnajam-20261010-protocol-ccga1-validation-cso-accuracy Aggregation: Not reported Evaluation of cell-free DNA approaches for multi-cancer early detection · Results, cancer signal origin prediction: 35% (44/127)
Configuration: Small somatic variant classifier with matched white-blood-cell background removal (GRAIL prototype), CCGA substudy 1	Protocol: CCGA substudy 1 training set: cancer signal sensitivity at 98% specificity under 10-fold cross-validation Dataset: CCGA substudy 1 training set (1,414 analysable participants)	36% (33%–39%) Sensitivity at 98% specificity (training set, post hoc 98% specificity threshold) percent · higher Uncertainty: 95% CI 33 to 39. Clopper-Pearson exact binomial interval (STAR Methods, statistical analysis) Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source SNV-WBC on CCGA substudy 1 training set: cancer signal sensitivity at 98% specificity under 10-fold cross-validation ctdnajam-20261010-protocol-ccga1-training-cv-sens-98spec Aggregation: Not reported Evaluation of cell-free DNA approaches for multi-cancer early detection · Table 3, row 'SNV-WBC', training set, sensitivity and TP/total cancer samples 299/833
Configuration: Small somatic variant classifier with matched white-blood-cell background removal (GRAIL prototype), CCGA substudy 1	Protocol: CCGA substudy 1 validation set: cancer signal sensitivity at 98% specificity Dataset: CCGA substudy 1 validation set (847 analysable participants)	33% (29%–38%) Sensitivity at 98% specificity (validation set, post hoc 98% specificity threshold) percent · higher Uncertainty: 95% CI 29 to 38. Clopper-Pearson exact binomial interval (STAR Methods, statistical analysis) Coverage: Not reported scored / Not reported eligible	Author-reported evaluation · Source checked Methods, coverage and source SNV-WBC on CCGA substudy 1 validation set: cancer signal sensitivity at 98% specificity ctdnajam-20261010-protocol-ccga1-validation-sens-98spec Aggregation: Not reported Evaluation of cell-free DNA approaches for multi-cancer early detection · Table 3, row 'SNV-WBC', validation set, sensitivity and TP/total cancer samples 155/464

Source checking is not independent reproduction. Release 2026-10-10-cbb3da59bc08.

Use this configuration

How it works, versions and access

Evidence

Source checking verifies the cited claim or transcription. It does not establish independent reproduction.

Evidence table

Inspect claims, sources and review details

Trace each statement to its source and review. A context-only reference supports the record generally; it does not verify an individual field. Source checking does not reproduce an experiment.

One row per statement and cited source. Multiple citations are not independent evaluations. Shared locators are labelled explicitly.

0 evidence rows matching the loaded filters

Claims, original sources and review scope · Release 2026-10-10-cbb3da59bc08
Property and statement	Original source and location	Review and provenance

No evidence rows match these filters. Choose another scope or clear the search.

All evidence (JSONL) (gzip)How to read the evidence

Sources and history

Release 2026-10-10-cbb3da59bc08 · Record review: source checked

1 source record and release history

Evaluation of cell-free DNA approaches for multi-cancer early detection · Original source · Cancer Cell 40(12):1537-1549.e12, published 2022-12-12; publisher PDF (1-s2.0-S153561082200513X-main.pdf) as deposited by the Francis Crick Institute on figshare, 10.25418/crick.21731870.v1

Download this release (gzip)

Technical metadata and extraction receipts

Stable ID: ctdnajam-20261010-config-snv-wbc

areas: dna-genomes
contexts: clinical_research
method types: supervised_machine_learning
reported name: SNV-WBC
foundation model eligible: false
source locator: Table 2; STAR Methods, TS: SNV and SNV-WBC classifiers
parameters: Targeted sequencing of 507 genes in cfDNA (60,000x raw, 3,000x unique depth); matched white-blood-cell targeted sequencing
missing metadata: version: reason: unreported; note: Prototype classifier; no release or commit is printed

Related records