CineMA: A Foundation Model for Cine Cardiac MRI

2025-05-31

AI Summary

Clinical cardiac magnetic resonance (CMR) quantitative analysis—e.g., ejection fraction estimation—relies heavily on labor-intensive, subjective, and poorly generalizable manual annotations. To address this, we introduce CardioSSL, the first self-supervised foundation model for dynamic cine CMR, trained exclusively on human-derived data via masked spatiotemporal image reconstruction. Trained on 74,916 multi-center cine CMR studies, CardioSSL supports 23 downstream tasks, including segmentation, functional assessment, disease classification, and anatomical localization. With only minimal labeled data, fine-tuned CardioSSL achieves state-of-the-art or competitive performance across eight independent benchmarks—matching or surpassing task-specific CNNs—while drastically reducing annotation burden and deployment cost. All code, model architectures, and pre-trained weights are publicly released to foster reproducible, low-resource clinical AI adoption.

Technology Category

• Machine Learning: Multimodal Learning
• Computer Vision: Multi-modal Vision
• Cognitive Modeling & Cognitive Systems: Other Foundations of Cognitive Modeling & Systems

Application Category

• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Cardiac magnetic resonance (CMR) is a key investigation in clinical cardiovascular medicine and has been used extensively in population research. However, extracting clinically important measurements such as ejection fraction for diagnosing cardiovascular diseases remains time-consuming and subjective. We developed CineMA, a foundation AI model automating these tasks with limited labels. CineMA is a self-supervised autoencoder model trained on 74,916 cine CMR studies to reconstruct images from masked inputs. After fine-tuning, it was evaluated across eight datasets on 23 tasks from four categories: ventricle and myocardium segmentation, left and right ventricle ejection fraction calculation, disease detection and classification, and landmark localisation. CineMA is the first foundation model for cine CMR to match or outperform convolutional neural networks (CNNs). CineMA demonstrated greater label efficiency than CNNs, achieving comparable or better performance with fewer annotations. This reduces the burden of clinician labelling and supports replacing task-specific training with fine-tuning foundation models in future cardiac imaging applications. Models and code for pre-training and fine-tuning are available at https://github.com/mathpluscode/CineMA, democratising access to high-performance models that otherwise require substantial computational resources, promoting reproducibility and accelerating clinical translation.

Problem

• Automates clinical CMR measurements like ejection fraction
• Reduces subjectivity and time in cardiovascular disease diagnosis
• Improves label efficiency for cardiac imaging tasks

Innovation

• Self-supervised autoencoder for cine CMR
• Fine-tuned foundation model with limited labels
• Outperforms CNNs with greater label efficiency