MedDream improves chest X-ray diagnosis and evidence generation
A radiographic world model for clinical reasoning and evidence generation
Artificial Intelligence
Summary
Medical AI often treats X-rays and their reports separately, but MedDream learns a shared understanding of both together. This helps the AI better recognize diseases, assess severity, and create realistic X-ray images from text reports. The model was trained on millions of chest X-rays with their corresponding reports and outperformed other AI methods in tests. It can also create synthetic images that help improve diagnosis, especially for patient groups that need more focused study.
What this means in practice
- •For radiology departments: Enhance diagnostic accuracy and clinical assessments using AI that integrates X-ray images with textual reports.
- •For medical ai developers: Generate targeted synthetic chest X-rays to improve training data diversity and model performance for specific patient subgroups.
Authors
Suyang Xi, Songtao Hu, Shansong Wang, Mojtaba Safari, Luke del Balzo, Ehsan Ul Karim, Mingzhe Hu, Kuo Zhang, Tonghe Wang, Ralph R. Weichselbaum, Xiaofeng Yang
Abstract
Medical imaging artificial intelligence (AI) is commonly developed as separate mappings from radiographs to diagnostic outputs or from clinical descriptions to generated images, although both arise from the same underlying radiographic state. A world-model formulation instead seeks to learn an internal representation of this state that can support both clinical readout and conditional simulation of radiographic observations. Here we introduce MedDream, a radiographic world model that learns a shared continuous latent state from paired chest radiograph-text observations for diagnostic reasoning and report-conditioned evidence generation. MedDream was pretrained on 2.65 million leakage-controlled chest radiograph-text pairs curated from 4.40 million candidates. Across eight clinical datasets and two independent reader cohorts, MedDream outperformed leading diagnostic and generative comparators. For diagnostic reasoning, MedDream showed strong generalization across disease recognition, label-scarce adaptation, severity assessment, and localization, while MedDream-supported review increased mean resident concordance with independent radiologist consensus from 56.3% to 63.0%. For evidence generation, MedDream produced radiographs that preserved clinically relevant pathology and improved downstream performance on held-out real data, with synthetic augmentation increasing external VinDr-CXR macro-AUROC from 76.4% to 81.4%. More importantly, conditioning generation on prespecified subgroup performance gaps enabled targeted evidence construction, increasing weighted F1 by 3.1 percentage points in Asian patients, whereas matched-volume unguided augmentation decreased it by 2.3 points. These findings establish radiographic world models as a path toward medical AI that learns clinically meaningful internal states for interpreting, simulating, and constructing evidence for clinical use.