AI Science Tracker
EN

Random forests classify CT-derived COVID severity in a retrospective cohort

Shahid Beheshti University of Medical Sciences · 26 Nov 2021 · medicine

Discovery

Random forests classify CT-derived COVID severity in a retrospective cohort

Source check: Source checked. Not independent scientific confirmation.

Finding: A random forest used patient-level aggregates of CT texture features to distinguish four COVID severity groups in 956 patients from the MosMed dataset. Reported overall accuracy was 90.95%. Eight-fold validation was selected after comparing fold configurations, which can make performance optimistic. This retrospective severity-classification study does not establish prospective triage safety or clinical benefit.

Technology: Random forest

Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.

AI contribution
A 40-tree random forest learns severity relationships from aggregated CT texture measurements, compared with LDA and nearest-neighbor classifiers.
Test and result
Patient-level CT feature aggregation, internal cross-validation and severity-reference comparisons.
Limitation / still unknown
External validation and prospective clinical utility remain unestablished; selection among validation configurations can bias performance.
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/PMC9015452/

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 Methods/results/limitations and exact front/canonical identity; current primary PubMed linked correction/retraction metadata 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.

Open in the full catalog for filters, charts and related discoveries. Browse medicine discoveries.

Related catalog entries

Links use documented relationships where available, plus shared catalog context. A shared topic or model family does not establish a scientific dependency.