Deep-learning chest-radiograph pressure estimates associate with heart-failure prognosis
Source check: Source checked. Not independent scientific confirmation.
Finding: A post hoc analysis of 534 acute-heart-failure admissions applied a previously developed VGG16-based model to estimate pulmonary arterial wedge pressure from chest radiographs. Higher discharge estimates and smaller decreases during hospitalization were associated with death or heart-failure readmission after adjustment; 223 patients reached the composite endpoint. This single-centre observational analysis does not establish a causal treatment target or improved care. Imaging-view differences and exclusion of in-hospital deaths limit interpretation.
Technology: VGG16-based regression convolutional neural network
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Previously trained VGG16 regression CNN with global-average-pooling output estimates pressure; model trained on 748 and tested on 188 data in earlier study. Present post hoc 534-patient registry links estimated pressure to outcomes; original prospective registry collection does not establish frozen-model prospective validation.
- Test and result
- Retrospective admission/discharge imaging and adjusted outcome association over median 289 days follow-up.
- Limitation / still unknown
- Prospective clinical utility, multicentre transportability and equivalence across radiograph views remain unestablished.
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/PMC10053257/
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 full article Methods, Results and limitations plus Europe PMC bibliographic metadata; PMC full article..
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.