AI-assisted cardiac measurements show an in-hospital mortality association
Source check: Source checked. Not independent scientific confirmation.
Finding: In the retrospective international WASE-COVID cohort, left-ventricular longitudinal strain was associated with in-hospital mortality after adjustment. Left-ventricular measurements combined machine-learned software output with two blinded human reads. This is a selected-cohort association, not an AI-only mortality predictor or demonstrated clinical benefit.
Technology: Ultromics machine-learned left-ventricular measurement algorithms
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Previously learned commercial algorithms generated ventricular volumes, ejection fraction and strain used with human reads in the association analysis. Conventional regression estimated the association.
- Test and result
- 870 hospitalized patients at13 centers;722 had usable left-ventricular analysis. Final measurements averaged one automated and two blinded human reads. Adjusted strain association OR1.179,95%CI1.045–1.358.
- Limitation / still unknown
- Retrospective first-wave population selected for clinically indicated echocardiograms; variable missing views and same-cohort stepwise adjustment. 89 missing two-chamber measurements imputed. No locked external mortality evaluation; individual strain discrimination limited. Right-ventricular measures used a separate semiautomated package. Software architecture/training supplement uninspected.
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/PMC8137346/ · https://pmc.ncbi.nlm.nih.gov/articles/PMC8418014/ · https://pmc.ncbi.nlm.nih.gov/articles/PMC8817760/
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 primary Methods/Results/limitations and both linked letters read in supported Chrome. Letters debate interpretation and utility; response clarifies odds ratios and all-cause in-hospital death, not PPV or COVID-specific death..
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.