A clinically accessible small multimodal radiology model and evaluation metric for chest X-ray findings

Large foundation models show promise in biomedicine but face challenges in clinical use due to performance gaps, accessibility, cost, and lack of scalable evaluation. Here we show that open-source small multimodal models can bridge these gaps in radiology by generating free-text findings from chest X-ray images. Our data-centric approach leverages 697K curated radiology image-text pairs to train a specialized, domain-adapted chest X-ray encoder. We integrate this encoder with pre-trained language models via a lightweight adapter that aligns image and text modalities. To enable robust, clinically relevant evaluation, we develop and validate CheXprompt, a GPT-4-based metric for assessing factual accuracy aligned with radiologists' evaluations. Benchmarked with CheXprompt and other standard factuality metrics, LLaVA-Rad (7B) achieves state-of-the-art performance, outperforming much larger models like GPT-4V and Med-PaLM M (84B). While not immediately ready for real-time clinical deployment, LLaVA-Rad is a scalable, privacy-preserving and cost-effective step towards clinically adaptable multimodal AI for radiology.

Authors

Institutions

Publication Details

Journal
Nature Communications
Published
2025-04-01
DOI
https://doi.org/10.1038/s41467-025-58344-x
Citations
22
Primary Topic
Topic Modeling
Type
article
Field-Weighted Citation Impact
54.58

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

A clinically accessible small multimodal radiology model and evaluation metric for chest X-ray findings

Yujia Xie, Yanbo Xu, Naoto Usuyama, Jianwei Yang et al.
22 citations
Nature Communications
Topic Modeling
54.58
article

A clinically accessible small multimodal radiology model and evaluation metric for chest X-ray findings

Yujia Xie, Yanbo Xu, Naoto Usuyama, Jianwei Yang, Mu Wei, Juan Manuel Zambrano Chaves, Hoifung Poon, Shih-Cheng Huang, Tristan Naumann, Hanwen Xu, Ziyi Yang, Cliff Wong, Julia Gong, Sheng Zhang, Jianfeng Gao, Sheng Wang, Akshay Sanjay Chaudhari, Serena Yeung, Hany Awadalla, Curtis P. Langlotz, Fei Wang, Yu Gu, Houdong Hu, Chunyuan Li, Mahmoud Khademi, Muhao Chen, Matthew P. Lungren
article en
22 citations

Abstract

Large foundation models show promise in biomedicine but face challenges in clinical use due to performance gaps, accessibility, cost, and lack of scalable evaluation. Here we show that open-source small multimodal models can bridge these gaps in radiology by generating free-text findings from chest X-ray images. Our data-centric approach leverages 697K curated radiology image-text pairs to train a specialized, domain-adapted chest X-ray encoder. We integrate this encoder with pre-trained language models via a lightweight adapter that aligns image and text modalities. To enable robust, clinically relevant evaluation, we develop and validate CheXprompt, a GPT-4-based metric for assessing factual accuracy aligned with radiologists' evaluations. Benchmarked with CheXprompt and other standard factuality metrics, LLaVA-Rad (7B) achieves state-of-the-art performance, outperforming much larger models like GPT-4V and Med-PaLM M (84B). While not immediately ready for real-time clinical deployment, LLaVA-Rad is a scalable, privacy-preserving and cost-effective step towards clinically adaptable multimodal AI for radiology.

Nature CommunicationsVol. 16(1)
University of Southern California (US), Microsoft (United States) (US), University of California, San Francisco (US), University of Washington (US), Microsoft Research (United Kingdom) (GB), University of California, Davis (US), Stanford University (US)
Microsoft Research
Openalex Percentile: Top 0%
Topic Modeling
54.58
DOIOpen Access PDFOpenAlex
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.