Cardiac MRI sequences generated from one frame without ECG signals
One Frame, Full Heartbeat: ECG-Free Cardiac Cine MRI Synthesis via Phase-Conditioned Flow Matching
Computer Vision and Pattern Recognition
Summary
Cardiac MRI scans usually need multiple images timed to a heartbeat, guided by an ECG signal, which can be tricky if the patient can't cooperate or the data is corrupted. The authors present PhaseFlow, a method that uses just one frame of heart MRI plus a pattern of heart motion to create a full heartbeat video without needing ECG data. It models the heart's motion accurately by learning the heart phase from the images themselves, producing clearer and more realistic cardiac movies. Tests show that PhaseFlow better captures heart function measures important for diagnosis.
What this means in practice
- •For hospital imaging teams: Generate full cardiac cycle MRI images from a single frame without ECG signals, improving scans for patients who cannot undergo standard ECG-gated procedures.
- •For medical device developers: Develop advanced MRI software that synthesizes dynamic heart images without dependency on ECG input, enabling more flexible cardiac imaging systems.$Commercial implications: Enables MRI scanner manufacturers to offer ECG-free cardiac imaging solutions, expanding usability in challenging clinical environments.
Authors
Shiyi Wang, Ruochen Sun, Peirong Liu, Xiang Li, Fangxu Xing
Abstract
Cine cardiovascular magnetic resonance (CMR) analysis relies on multi-frame sequences capturing the full cardiac cycle. However, standard multi-frame acquisition depends heavily on electrocardiogram (ECG) gating and repeated breath-holds, posing challenges in uncooperative populations, resource-limited settings, and temporally corrupted datasets. Existing methods that synthesize full cardiac sequences either rely on explicit ECG signals to parameterize myocardium function, or employ deformable registration without physiological constraints, failing to faithfully reproduce clinically relevant dynamic metrics such as ejection fraction (EF) and ventricular contraction magnitude. We present PhaseFlow, a unified generative framework that overcomes both limitations. PhaseFlow estimates a non-linear cardiac phase signal directly from the input sequence via a segmentation-derived left-ventricular (LV) area curve, capturing the asymmetric dynamics of systole and diastole without any ECG dependency. At inference, this phase signal is provided by a pathology-specific template, informing phase-specific frame generation. A rectified flow model conditioned on the phase and slice position synthesizes the full cardiac motion trajectory in the latent space, decoded into a diffeomorphic displacement field that warps end-diastole pixel intensities directly, eliminating the reconstruction blur often accompanying the variational autoencoder. On the ACDC benchmark, PhaseFlow achieves superior physiological fidelity and image realism, with best LV volume curve $R^2$, structural similarity (SSIM) and generative quality (FID) among all baselines. Ablation studies confirm that each proposed component contributes measurably to the overall performance.