Cine Cardiac MRI Motion Artifact Reduction Using a Recurrent Neural Network.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Cardiovascular Disease Diagnosis

Background:

• Cine cardiac magnetic resonance imaging (MRI) offers excellent contrast for diagnosing heart conditions.
• However, MRI scans are lengthy, leading to motion artifacts and patient discomfort, necessitating faster imaging techniques.
• Deep learning shows promise for improving MRI quality and reducing scan times.

Purpose of the Study:

• To develop a novel deep learning model for reducing motion artifacts in cine cardiac MRI.
• To enhance image quality and temporal resolution in cardiac MRI sequences.
• To improve the clinical utility of cardiac MRI by addressing motion-related artifacts and scan duration.

Main Methods:

• A novel recurrent generative adversarial network (GAN) model was proposed for cardiac MRI motion artifact reduction.
• The model incorporates bi-directional convolutional long short-term memory (ConvLSTM) for temporal feature handling.
• Multi-scale convolutions were employed to capture both local and global image features.

Main Results:

• The proposed deep learning method demonstrated superior performance in reducing motion artifacts compared to state-of-the-art techniques.
• The model achieved improved image quality for cine cardiac MRI.
• The method successfully generated reliable intermediate frames, enhancing temporal resolution.

Conclusions:

• The novel recurrent GAN effectively reduces motion artifacts and improves image quality in cine cardiac MRI.
• The architecture's ability to capture cardiovascular dynamics ensures good generalizability.
• This approach offers a significant advancement for cardiac MRI, potentially leading to more accurate and comfortable diagnoses.