CT learning detects gastric cancer downstaging after chemotherapy
Source check: Source checked. Not independent scientific confirmation.
Finding: Linear support-vector models analyzed baseline and restaging CT in a retrospective 292-patient gastric-cancer study. Models were evaluated in two separate testing cohorts against downstaging defined by baseline clinical and postoperative pathological stages; restaging AUCs were 0.922 and 0.850. The study did not demonstrate improved surgical decisions, and selected images, small same-centre tests and inconsistent abstract cohort labels limit interpretation.
Technology: Linear support-vector classifier; radiomic feature selection
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Linear support-vector classifiers learn relationships between selected CT texture features and recorded post-chemotherapy downstaging.
- Test and result
- Model development with nested cross-validation, followed by two held-out same-centre clinical testing cohorts and a postoperative pathology-based stage reference.
- Limitation / still unknown
- Prospective decision benefit and multicentre generalization remain unestablished; clinical baseline staging can be inaccurate, and low-quality images were excluded.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC8523390/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 21 Nov 2026. Review scope: Primary relevant model Methods, Results and limitations; exact primary identity and current primary PubMed correction/retraction relationships checked..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.