English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Recent advances in training deep learning models have demonstrated thepotential to provide accurate chest X-ray interpretation and increase access toradiology expertise. However, poor generalization due to data distributionshifts in clinical settings is a key barrier to implementation. In this study,we measured the diagnostic performance for 8 different chest X-ray models whenapplied to (1) smartphone photos of chest X-rays and (2) external datasetswithout any finetuning. All models were developed by different groups andsubmitted to the CheXpert challenge, and re-applied to test datasets withoutfurther tuning. We found that (1) on photos of chest X-rays, all 8 modelsexperienced a statistically significant drop in task performance, but only 3performed significantly worse than radiologists on average, and (2) on theexternal set, none of the models performed statistically significantly worsethan radiologists, and five models performed statistically significantly betterthan radiologists. Our results demonstrate that some chest X-ray models, underclinically relevant distribution shifts, were comparable to radiologists whileother models were not. Future work should investigate aspects of model trainingprocedures and dataset collection that influence generalization in the presenceof data distribution shifts.