The abstract is read at the publisher; the summary is JClub's.
A classifier using genome-wide rbcDNA read-depth features detects early-stage gastric cancer with 83%–90% overall sensitivity and 82%–89% sensitivity for stage I across three multicenter cohorts.
- Why it matters: Early gastric cancer detection is constrained by invasive endoscopy and limited sensitivity of existing blood biomarkers, creating a need for a non-invasive approach.
- What they did: Shallow whole-genome sequencing of rbcDNA from 1–2 mL peripheral blood in 865 participants identified 1,000 reproducible read-depth features; a dimensionally reduced machine-learning model was trained with a fixed 90% specificity cutoff and tested in independent cohorts.
- The result: The model reached 84% AUC in nested cross-validation and 95% AUC in the test cohort, and it distinguished early-stage GC from precancerous lesions and other tumors while detecting ~80% of GC cases negative for tumor markers.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.