AUROC (GI cancer vs non-cancer, validation partition, mean over three split-runs) of Random Forest on 20,499 DMR-informed CpGs (caret) on EpiPanGI Dx plasma cfDNA targeted bisulfite methylation (GSE149438, 300 samples)
0.793 AUROC (GI cancer vs non-cancer, validation partition, mean over three split-runs)
Methods
- Tested configuration
- Random Forest on 20,499 DMR-informed CpGs (caret)
- Protocol
- Xu and Feng 2026: GI cancer vs non-cancer on GSE149438, three train-validation-test split-runs, DMR-informed CpG features
- Dataset
- EpiPanGI Dx plasma cfDNA targeted bisulfite methylation (GSE149438, 300 samples)
- Procedure
- Xu and Feng 2026: GI cancer vs non-cancer on GSE149438, three train-validation-test split-runs, DMR-informed CpG features
- Evaluation
- Random Forest on GSE149438 DMR-informed split-runs (Xu and Feng 2026)
- Coverage
- Not reported scored / Not reported eligible
- Uncertainty
- 95% CI 0.5946 to 0.9920. Across three split-runs, bounded to the valid range 0 to 1 (Table 3 footnote)
- Evidence
- Author-reported evaluation · source checkedAn Evaluation of DMR-Informed Fine-Tuning of Tissue Array-Pretrained CpGPT for Gastrointestinal Cancer Classification Using cfDNA Targeted Methylation · Table 3, row 'Random Forest', column 'Validation AUROC'
A source-checked result verifies the numerical transcription, not every model or protocol detail. Evaluation metadata: source checked. Source checked does not mean independently reproduced.
Reproduction
- Split
- Three train-validation-test partitions (seeds 42, 123, 456), about 10% test
- Adaptation
- Fine-tuned (CpGPT) or trained (classical) on each training partition
- Scoring implementation
- Not reported
No execution recipe has been verified for this exact configuration and evaluation. A benchmark's general instructions may use different inputs, splits or model settings.
Reproducing this published result requires matching its model configuration, data, split and scorer. Source checking or a successful smoke test does not establish score reproduction.
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.
1 evidence row matching the loaded filters
| Property and statement | Original source and location | Review and provenance |
|---|---|---|
| Reported result 0.7933 Individual claims | An Evaluation of DMR-Informed Fine-Tuning of Tissue Array-Pretrained CpGPT for Gastrointestinal Cancer Classification Using cfDNA Targeted Methylation Table 3, row 'Random Forest', column 'Validation AUROC' Version: medRxiv version 1, posted 2026-07-29 | source checked ["pdf-text-parse","independent-cell-check"] · 2026-10-12 author reported Audit detailsThe pinned bytes could not be re-verified: medRxiv regenerates this PDF, and a download on 2026-10-12 had a different SHA-256 (see the source's hash_scope). Compared all 15 Table 3 values and intervals, the asserted sentences and the footnote with the layout text of that copy, read for this review (the extractor's script was not run); all match. Field: Source artifact SHA-256: Hash scope: artifact_sha256 describes one retrieval (2026-10-11T21:49:59Z). medRxiv regenerates this PDF: a reviewer download on 2026-10-12 had 2,761,577 bytes, SHA-256 3d37b32e8ff782201ed8ee6677a01808d4f8c86eea70350a530dde7306cab7c9 and PDF CreationDate 2026-10-11T20:08:12-07:00. The reviewer's copy's text layer matched every extracted value. The licence does not allow archiving. Extraction artifact SHA-256: |
Sources and history
Release 2026-11-21-c8bea25e5abb · Record review: source checked
1 source record and release history
- An Evaluation of DMR-Informed Fine-Tuning of Tissue Array-Pretrained CpGPT for Gastrointestinal Cancer Classification Using cfDNA Targeted Methylation · Original source · medRxiv version 1, posted 2026-07-29
Technical metadata and extraction receipts
Stable ID: ctdnalit-20261011-cpgpt2026-result-random-forest-validation-auroc
- metric
- auroc
- metric qualifier
- GI cancer vs non-cancer, validation partition, mean over three split-runs
- metric direction
- higher
- unit
- unitless
- printed value
- 0.7933
- numeric value
- 0.7933
- source locator
- Table 3, row 'Random Forest', column 'Validation AUROC'
- review
- method: pdf-text-parse; independent-cell-check; reviewer: claude; reviewer note: Separate Claude review agent (reviewer-ctdna), independent of the extractor uc-ctdna; no human review claimed; date: 2026-10-12; artifact sha256: 3d37b32e8ff782201ed8ee6677a01808d4f8c86eea70350a530dde7306cab7c9; retrieval url: https://www.medrxiv.org/content/10.64898/2026.07.26.26358979v1.full.pdf; method note: Extracted by deterministic parse of the PDF text layer (pdftotext -layout) with caption, header, model labels and footnote asserted.; note: The pinned bytes could not be re-verified: medRxiv regenerates this PDF, and a download on 2026-10-12 had a different SHA-256 (see the source's hash_scope). Compared all 15 Table 3 values and intervals, the asserted sentences and the footnote with the layout text of that copy, read for this review (the extractor's script was not run); all match.
- uncertainty
- type: confidence_interval; lower: 0.5946; upper: 0.9920; level: 0.95; printed: (0.5946–0.9920); note: Across three split-runs, bounded to the valid range 0 to 1 (Table 3 footnote)