The abstract is read at the publisher; the summary is JClub's.
A deep learning model detected any stage of chronic kidney disease from parasternal long-axis echocardiography videos with AUC 0.756 (0.749–0.763) in a held-out CSMC test cohort.
- Why it matters: Undiagnosed chronic kidney disease is common, and current screening is limited; leveraging links between kidney and cardiovascular disease could enable noninvasive detection.
- What they did: The model was trained on 325,377 PLAX videos from 62,818 patients at Cedars-Sinai Medical Center and externally validated on cohorts of 2,762 patients (Stanford Healthcare) and 41,611 patients (Kaiser-Permanente Northern California).
- The result: Performance remained robust externally, with AUC 0.718 (0.714–0.723) at KPNC and 0.719 (0.704–0.735) at SHC, including across subgroups with and without metabolic comorbidities.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.