AI Model Synthesizes DCE-MRI Contrast Without Gadolinium

A novel artificial intelligence framework has been introduced in a preprint on arXiv (2607.29394), offering a technique for synthesizing contrast enhancement in dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) without relying on gadolinium-based contrast agents (GBCAs). This method, known as conditioned latent transport, enables the prediction of contrast enhancement through a single forward pass by linking the latent trajectory to pre-contrast anatomy and utilizing continuous time conditioning. This innovation allows for the creation of patient-specific contrast evolution at any acquisition time. It overcomes the challenges faced by current contrast synthesis methods, which often struggle with spatial realism, temporal continuity, slow sampling, and lack of clinical validation. The model demonstrates superior performance compared to baseline and advanced models in various metrics. This research holds significance for breast cancer management, particularly as GBCAs are limited in certain populations, extend scan protocols, and raise environmental toxicity issues. The preprint was shared as a cross-type submission on arXiv.