What question did this study set out to answer?

This research aims to evaluate the ability of generative AI models to create realistic knee radiographs, addressing data scarcity in osteoarthritis studies.

March 5, 2026BMC Medical Imaging0 citationsOpen Access

Comparative evaluation of generative artificial intelligence models for synthetic knee radiograph augmentation in clinical research

KCKwangho ChungSeverance Hospital JNJi-Hoon NamSK Group (South Korea)ADArailym DossetKorea Pharma (South Korea)

Key Points

This research aims to evaluate the ability of generative AI models to create realistic knee radiographs, addressing data scarcity in osteoarthritis studies.
Trained three generative models (StyleGAN3, CycleGAN, DCGAN) on 10,042 knee X-rays.
Assessed image quality using Fréchet Inception Distance (FID) and conducted visual evaluations by orthopedic experts.
Measured Joint Line Convergence Angle (JLCA) for anatomical fidelity and calculated observer reliability with intraclass correlation coefficients (ICC).
StyleGAN3 achieved the best performance with an FID of 10.84, showing high visual and anatomical fidelity.
Both CycleGAN and DCGAN showed lower FID scores, indicating varying quality in synthetic images.
No significant differences in JLCA measurements between real and synthetic images across all models, suggesting high anatomical similarity.

Abstract

In this study, the capability of state-of-the-art generative models to synthesize realistic knee radiographs was evaluated to address dataset scarcity in osteoarthritis (OA) research. Three generative frameworks—Style Generative Adversarial Network3 (StyleGAN3), a stable diffusion + Cycle-consistent Generative Adversarial Network (CycleGAN) pipeline, and Deep Convolutional Generative Adversarial Network (DCGAN)—were trained on 10,042 real knee X-rays. Image quality was assessed using Fréchet Inception Distance (FID) while visual fidelity was evaluated via a Visual Turing Test conducted by two orthopedic surgeons and a musculoskeletal radiologist. Joint Line Convergence Angle (JLCA) was compared between real and synthetic images for anatomical fidelity. Inter- and intra-observer reliability for JLCA was measured using intraclass correlation coefficients (ICC). StyleGAN3 achieved the best performance (FID 10.84), showing high visual and anatomical fidelity. Integrating Stable Diffusion with CycleGAN showed a moderate FID of 39.79, suggesting that adversarial enhancements improved the diffusion-based synthesis. DCGAN showed lower quality, achieving an FID of 74.15. Expert accuracy in distinguishing real from synthetic images ranged between 36% and 88%, confirming difficulty in visual differentiation. Furthermore, JLCA measurements showed no significant difference between real (4.19 ± 3.07°) and synthetic (3.36 ± 2.19°) images generated by DCGAN (p = 0.12). Similarly, Diffusion + CycleGAN (3.91 ± 2.59° vs. 3.72 ± 2.52°, p = 1.00) and StyleGAN3 (4.27 ± 3.01° vs. 3.60 ± 2.37°, p = 0.25) showed no statistically significant differences. These results indicate that all elevated generative models maintained high anatomical fidelity relative to real radiographs. Inter-observer agreement was strong, with ICC values ranging between 0.83 and 0.97. Intra-observer reliability was also excellent. StyleGAN3 generated the most realistic knee radiographs. Diffusion-based pipelines showed promising results when enhanced with adversarial networks. These findings underscore the potential of generative AI to mitigate data limitations in orthopedic research.

问 AI

Bookmark

View Full Paper