Dense Temporal Contrast Synthesis via Conditioned Latent Transport
AuthorsSmriti Joshi, Apostolia Tsirikoglou, Daniel M. Lang, Richard Osuala, Noah Márquez Varaa, Alejandro Guzman, Grzegorz Skorupko, Sebastian Ibarra Arregui, Lidia Garrucho, Akane Ohashi, Dimitra Ntoula, Eugen Divjak, Oğuz Lafcı, Jan C. Peeken, Julia A. Schnabel, Fredrik Strand, Oliver Diaz, Karim Lekadir
Resources
A single-pass generative model synthesizes realistic, time-resolved breast MRI contrast without gadolinium and shows promising robustness and clinical utility.
Key results
Patients in the multicenter training and validation dataset
Downsampling factor of the custom VAE
MSE reported at 10−2 scale on internal validation
Dice achieved with synthesized contrast versus 0.49 for pre-contrast baseline
Improvement from 0.49 baseline Dice to 0.60
Cases in which synthetic images supported the same management decision as real DCE-MRI
What the paper found
Researchers from the Universitat de Barcelona, Karolinska Institutet, Helmholtz Munich, and Technical University of Munich propose conditioned latent transport for gadolinium-free or reduced-contrast breast DCE-MRI. Their model anchors a continuous-time generative trajectory to each patient’s pre-contrast anatomy, predicts the residual enhancement in latent space, and reconstructs any acquisition phase in a single forward pass. A custom VAE with 4x spatial compression preserves microvascular detail, while sinusoidal time conditioning, a fixed patient-level noise map, LPIPS perceptual loss, and focal frequency loss jointly enforce spatial realism and temporal continuity. On MAMA-MIA, a 1,506-patient multicenter dataset, the method achieved an internal-validation MSE of 0.84×10−2, PSNR of 21.61, and DTW of 0.68×10−2, while remaining competitive on an independent Karolinska cohort despite scanner and protocol shifts. In downstream tumor segmentation, synthetic enhancement increased Dice from 0.49 on pre-contrast images to 0.60, a 22.4% relative improvement, with boundary error also reduced by more than 39%. In a reader study of 40 cases involving four breast radiologists, synthetic images supported the same management decision as real DCE-MRI in 70% of cases and were preferred over U-Net and TeNCA outputs in 83.8% of evaluations, although incorrect tumor localization remained a critical failure mode. The paper acknowledges Gemini 3.1 Pro, Claude Sonnet 4.6, and DeepSeek V4 Flash for assistance with text and code.
Original abstract
Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) is essential for breast cancer management, but reliance on gadolinium-based contrast agents (GBCAs) restricts use in contraindicated populations, prolongs scan protocols, and presents environmental toxicity concerns. Contrast synthesis offers a non-invasive alternative; however, existing approaches struggle to balance spatial realism with temporal continuity, suffer from slow iterative sampling, underutilize structural priors, and lack clinical validation. We propose a novel conditioned latent transport framework that predicts contrast enhancement in a single forward pass. By anchoring the latent trajectory to the pre-contrast anatomy and applying continuous time conditioning, the model synthesizes patient-specific contrast evolution at any acquisition time. The proposed approach outperforms baseline and the state-of-the-art models across spatial, perceptual, temporal, and distributional metrics. Evaluated on an independent external cohort, the method demonstrates robustness to domain shifts induced by scanner noise as well as differing acquisition protocol. Furthermore, our synthetic contrast enhancement significantly improved downstream tumor segmentation performance, yielding a 22.4% relative increase in Dice coefficient (0.60 vs. 0.49 baseline pre-contrast, p < 0.01), reducing boundary segmentation error by over 39%, while outperforming all other generative model baselines. Finally, a reader study involving four breast radiologists evaluated the image quality, kinetic fidelity, and diagnostic viability of our synthesized sequences across 40 randomly selected cases. The results demonstrated that in 70% of cases, synthesized images provided sufficient clinical information to support the same management decisions as real DCE-MRI, suggesting a path toward safer and faster contrast-free or contrast-reduced imaging workflows.
Read the original paperMore in Generative Models
Browse all 63 papers →RULER: Instance-aware Rubric Rewards for SVG Generation
Hangyu Ran, Yuhao Zheng, Yingying Zhang, Kevin Qinghong Lin, Han Peng
RULER uses instruction-specific visual rubrics as reinforcement-learning rewards to make SVG generation more faithful, stylish, and resistant to reward hacking.
Think Before You Score: Thinking Reward Model for Visual Generation
Xuehai Bai, Zhenchen Tang, Yang Shi, Dianyi Wang, Tengfei Liu, Wanshun Su, Xuanyu Zhu, Ruohui Wang, Haiwen Diao, Haotian Wang, Xiaoling Gu, Yuanxing Zhang
A visual reward model that first decides what matters in each image-generation case, then scores outputs with detailed rubrics to provide better training signals.
WanPE: Towards Cinematic Prompt Enhancement for Modern Text-to-Video Generation
Yubo Zhu, Yawen Shao, Ziyun Dai, Zixun Fang, Kai Zhu, Siyang Sun, Haolan Xue, Chuxin Wang, Tingyu Weng, Jingming Luo, Chen Shi, Lianghua Huang, Yufeng Ai, Yuzheng Wang, Wenyuan Zhang, Yu Shang, Yuxiang Bao, Zoubin Bi, Jie Xiao, Jinbo Xing, Jiaxing Zhao, Chongyang Zhong, Hengjian Chen, Chenwei Xie, Akide Liu, Zhehan Kan, Yu Liu, Wei Zhai, Sheng Zhong, Wei Tong
WanPE turns ordinary text prompts into director-level cinematic plans, substantially improving the quality and consistency of long-form AI-generated videos.