Comparing Conditional Diffusion Models for Synthesizing Contrast-Enhanced Breast MRI from Pre-Contrast Images

2025-08-19

AI Summary

To address safety risks, contraindications, and clinical workflow burdens associated with contrast agent–dependent dynamic contrast-enhanced MRI (DCE-MRI), this work proposes a contrast-free dynamic enhancement image synthesis method. Leveraging denoising diffusion probabilistic models, we design 22 conditional diffusion variants, introducing a tumor-aware loss function and explicit tumor segmentation masks as strong conditional constraints; training further employs subtraction-based supervision and multimodal pre-contrast inputs (T2-weighted, DWI, and ADC). Evaluated on a public multicenter dataset, our method achieves state-of-the-art performance across five quantitative metrics—PSNR, SSIM, LPIPS, FID, and lesion Dice—outperforming all baselines. Blinded radiologist assessment confirms high visual fidelity and clinical interpretability of the synthesized images. This work establishes a novel paradigm for contrast-free breast cancer screening and diagnosis.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Cognitive Modeling & Cognitive Systems: Computational Creativity
• Search and Optimization: Mixed Discrete/Continuous Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Dynamic contrast-enhanced (DCE) MRI is essential for breast cancer diagnosis and treatment. However, its reliance on contrast agents introduces safety concerns, contraindications, increased cost, and workflow complexity. To this end, we present pre-contrast conditioned denoising diffusion probabilistic models to synthesize DCE-MRI, introducing, evaluating, and comparing a total of 22 generative model variants in both single-breast and full breast settings. Towards enhancing lesion fidelity, we introduce both tumor-aware loss functions and explicit tumor segmentation mask conditioning. Using a public multicenter dataset and comparing to respective pre-contrast baselines, we observe that subtraction image-based models consistently outperform post-contrast-based models across five complementary evaluation metrics. Apart from assessing the entire image, we also separately evaluate the region of interest, where both tumor-aware losses and segmentation mask inputs improve evaluation metrics. The latter notably enhance qualitative results capturing contrast uptake, albeit assuming access to tumor localization inputs that are not guaranteed to be available in screening settings. A reader study involving 2 radiologists and 4 MRI technologists confirms the high realism of the synthetic images, indicating an emerging clinical potential of generative contrast-enhancement. We share our codebase at https://github.com/sebastibar/conditional-diffusion-breast-MRI.

Problem

• Synthesizing contrast-enhanced breast MRI without contrast agents
• Evaluating 22 generative models for lesion fidelity enhancement
• Assessing clinical realism through radiologist and technologist study

Innovation

• Conditional diffusion models synthesize contrast-enhanced MRI
• Tumor-aware loss functions enhance lesion fidelity
• Segmentation mask conditioning improves contrast uptake realism