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 Tools annual

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 monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

The remarkable generative capabilities of multimodal foundation models arecurrently being explored for a variety of applications. Generating radiologicalimpressions is a challenging task that could significantly reduce the workloadof radiologists. In our study we explored and analyzed the generative abilitiesof GPT-4 for Chest X-ray impression generation. To generate and evaluateimpressions of chest X-rays based on different input modalities (image, text,text and image), a blinded radiological report was written for 25-cases of thepublicly available NIH-dataset. GPT-4 was given image, finding section or bothsequentially to generate an input dependent impression. In a blind randomizedreading, 4-radiologists rated the impressions and were asked to classify theimpression origin (Human, AI), providing justification for their decision.Lastly text model evaluation metrics and their correlation with theradiological score (summation of the 4 dimensions) was assessed. According tothe radiological score, the human-written impression was rated highest,although not significantly different to text-based impressions. The automatedevaluation metrics showed moderate to substantial correlations to theradiological score for the image impressions, however individual scores werehighly divergent among inputs, indicating insufficient representation ofradiological quality. Detection of AI-generated impressions varied by input andwas 61% for text-based impressions. Impressions classified as AI-generated hadsignificantly worse radiological scores even when written by a radiologist,indicating potential bias. Our study revealed significant discrepancies betweena radiological assessment and common automatic evaluation metrics depending onthe model input. The detection of AI-generated findings is subject to bias thathighly rated impressions are perceived as human-written.

Evaluation of GPT-4 for chest X-ray impression generation: A reader study on performance and perception