Scaling Generative Foundation Models for Chest Radiography with Rectified Flow Transformers

2026-06-17

AI Summary

This work addresses the limited clinical deployment of existing chest X-ray AI models due to poor generalization. We propose the first billion-parameter generative foundation model for chest radiographs, trained from scratch on 1.2 million heterogeneous X-ray images accompanied by expert-annotated metadata. Built upon a Rectified Flow Transformer architecture and trained on 1.6 trillion tokens, the model enables high-fidelity, controllable generation and editing conditioned on demographic attributes, imaging views, and over ten pathological findings. The synthesized images achieve a new state-of-the-art in visual fidelity, with clinical experts unable to distinguish them from real radiographs in blinded evaluations. This capability demonstrates significant potential for data augmentation and robustness evaluation of diagnostic AI systems.

Abstract

We introduce the first generative foundation model for chest radiograph synthesis trained from scratch at the billion-parameter scale. Existing radiographic AI models often suffer from poor generalisation across patient subpopulations, institutions, and acquisition settings, resulting in limited real-world clinical utility. Controlled, high-fidelity synthesis of chest radiographs is a promising path toward diversifying clinical datasets and evaluating the robustness of diagnostic models. Therefore, we present the largest specialist generative foundation model for chest radiographs to date, with over 1.3B parameters, trained for 1.6T tokens on a curated, heterogeneous dataset comprising 1.2M radiographs and clinical expert-guided metadata. Our model supports controllable radiograph generation and editing across multiple demographic subgroups, acquisition views, and a dozen pathologies. Moreover, we significantly advance the state of the art in radiograph synthesis fidelity, producing images that are indistinguishable from real radiographs to clinical experts.

Problem

• chest radiography
• generalisation
• clinical utility
• medical imaging
• AI robustness

Innovation

• generative foundation model
• chest radiography synthesis
• rectified flow transformers
• controllable medical image generation
• billion-parameter scale