Histology transformer distinguishes recorded glioma marker labels
Source check: Source checked. Not independent scientific confirmation.
Finding: ViT-WSI used whole-slide histology to classify recorded IDH1 status in pooled institutional and TCGA glioma data, reporting a patient-level ROC area of0.960. The labels included clinical immunohistochemistry and sequencing references. This is retrospective label-discrimination evidence, not a new mutation discovery or replacement for molecular testing.
Technology: Vision transformer; weakly supervised multiple-instance learning
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Pretrained patch embeddings and a learned attention aggregator produce slide representations, fine-tuned for recorded marker-label classification and aggregated across a patient’s slides.
- Test and result
- The wider institutional dataset had5216 slides from1211 patients, with biomarker-defined subsets pooled with TCGA glioma slides for marker modeling. Methods explicitly keep patients distinct across splits; reported IDH1 AUC is patient-aggregated.
- Limitation / still unknown
- TCGA was pooled rather than reserved as a wholly untouched external test. Molecular predictors were initialized from a subtype model, whose overlap with downstream test cases was not independently reconciled. The paper describes both ten random70:30 splits and tenfold evaluation; tuning containment is incomplete. Immunostaining proxies and direct sequencing are heterogeneous references. No causal genotype mechanism, real clinical benefit or independent replication is established.
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/PMC9839963/ · https://doi.org/10.1016/j.isci.2022.105872
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: Actual relevant primary computational/clinical Methods, model/result context and limitations read; no operational biological protocols.
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.