Improving Performance, Robustness, and Fairness of Radiographic AI Models with Finely-Controllable Synthetic Data

Achieving robust performance and fairness across diverse patient populations remains a central challenge in developing clinically deployable deep learning models for diagnostic imaging. Synthetic data generation has emerged as a promising strategy to address current limitations in dataset scale and diversity. In this study, we introduce RoentGen-v2, a state-of-the-art text-to-image diffusion model for chest radiographs that enables fine-grained control over both radiographic findings and patient demographic attributes, including sex, age, and race/ethnicity. RoentGen-v2 is the first model to generate clinically plausible chest radiographs with explicit demographic conditioning, facilitating the creation of a large, demographically balanced synthetic dataset comprising over 565,000 images. We use this large synthetic dataset to evaluate optimal training pipelines for downstream disease classification models. In contrast to prior work that combines real and synthetic data naively, we propose an improved training strategy that leverages synthetic data for supervised pretraining, followed by fine-tuning on real data. Through extensive evaluation on over 137,000 held-out chest radiographs from five institutions, we demonstrate that synthetic pretraining consistently improves model performance, generalization to out-of-distribution settings, and fairness across demographic subgroups defined across varying fairness metrics. Across datasets, synthetic pretraining led to a 6.5% accuracy increase in the performance of downstream classification models, compared to a modest 2.7% increase when naively combining real and synthetic data. We observe this performance improvement simultaneously with the reduction of the underdiagnosis fairness gap by 19.3%, with marked improvements across intersectional subgroups of sex, age, and race/ethnicity. Our proposed data-centric training approach that combines high-fidelity synthetic training data with multi-stage training pipelines is label-efficient, reducing reliance on large quantities of annotated real data. These results highlight the potential of demographically controllable synthetic imaging to advance equitable and generalizable medical deep learning under real-world data constraints. We open source our code, trained models, and synthetic dataset.