Synthesizing Accurate and Realistic T1-weighted Contrast-Enhanced MR Images using Posterior-Mean Rectified Flow

2025-08-18

AI Summary

This study addresses the clinical, economic, temporal, and environmental burdens associated with gadolinium-based contrast agents (GBCAs) in T1-weighted brain MRI. To enable reliable contrast-free neuro-oncological diagnosis, we propose a two-stage posterior mean correction flow framework. First, a patch-based 3D U-Net estimates voxel-wise posterior means to preserve anatomical fidelity. Second, a time-conditioned 3D correction flow refines lesion boundaries and vascular textures, jointly optimizing perceptual realism and geometric accuracy. Evaluated on 360 test cases, our method achieves an axial Fréchet Inception Distance (FID) of 12.46—68.7% lower than the baseline—along with a Kernel Inception Distance (KID) of 0.007 and a volumetric mean squared error (MSE) of 0.057. These results significantly outperform existing approaches, establishing a highly reliable generative solution for contrast-free brain tumor assessment.

Technology Category

• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Machine Learning: Calibration & Uncertainty Quantification
• Natural Language Processing: Generation

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Large pretrained models with web data
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications

Abstract

Contrast-enhanced (CE) T1-weighted MRI is central to neuro-oncologic diagnosis but requires gadolinium-based agents, which add cost and scan time, raise environmental concerns, and may pose risks to patients. In this work, we propose a two-stage Posterior-Mean Rectified Flow (PMRF) pipeline for synthesizing volumetric CE brain MRI from non-contrast inputs. First, a patch-based 3D U-Net predicts the voxel-wise posterior mean (minimizing MSE). Then, this initial estimate is refined by a time-conditioned 3D rectified flow to incorporate realistic textures without compromising structural fidelity. We train this model on a multi-institutional collection of paired pre- and post-contrast T1w volumes (BraTS 2023-2025). On a held-out test set of 360 diverse volumes, our best refined outputs achieve an axial FID of $12.46$ and KID of $0.007$ ($sim 68.7%$ lower FID than the posterior mean) while maintaining low volumetric MSE of $0.057$ ($sim 27%$ higher than the posterior mean). Qualitative comparisons confirm that our method restores lesion margins and vascular details realistically, effectively navigating the perception-distortion trade-off for clinical deployment.

Problem

• Synthesize CE MRI without gadolinium contrast agents
• Improve image realism while preserving structural accuracy
• Reduce environmental and patient risks in neuro-oncology imaging

Innovation

• Two-stage Posterior-Mean Rectified Flow pipeline
• Patch-based 3D U-Net predicts voxel-wise mean
• Time-conditioned 3D rectified flow refines textures