Present and Future Innovations in AI and Cardiac MRI

Manuel A. Morales,Warren J. Manning,Reza Nezafat

DOI: https://doi.org/10.1148/radiol.231269

IF: 19.7

2024-01-10

Radiology

Abstract:Cardiac MRI is used to diagnose and treat patients with a multitude of cardiovascular diseases. Despite the growth of clinical cardiac MRI, complicated image prescriptions and long acquisition protocols limit the specialty and restrain its impact on the practice of medicine. Artificial intelligence (AI)-the ability to mimic human intelligence in learning and performing tasks-will impact nearly all aspects of MRI. Deep learning (DL) primarily uses an artificial neural network to learn a specific...

radiology, nuclear medicine & medical imaging

What problem does this paper attempt to address?

This paper aims to explore the current applications and future innovations of artificial intelligence (AI) in cardiac magnetic resonance imaging (Cardiac MRI). Specifically, the paper attempts to address the following aspects:

1. **Improving the efficiency of cardiac MRI**:
- **Patient selection and protocol development**: Optimize patient selection criteria and scanning protocols through AI technology to make cardiac MRI more efficient.
- **Accelerating data acquisition and reconstruction**: Use AI technology to reduce scanning time and improve the spatial and temporal resolution of images, thereby improving patient compliance.
- **Automated image analysis**: Automatically complete image segmentation, measurement, and preliminary analysis through AI, reducing the workload of doctors.

2. **Enhancing the clinical value of cardiac MRI**:
- **Image reconstruction**: Improve image reconstruction methods using AI technology to reduce artifacts caused by undersampling and improve image quality.
- **Image enhancement and denoising**: Improve the resolution and signal - to - noise ratio of images through AI technology, making low - resolution or low - signal - to - noise - ratio images of higher quality.
- **Quantitative myocardial tissue feature analysis**: Estimate myocardial tissue parameters from fewer images using AI technology, shortening the total scanning time.

3. **Integrating multimodal data**:
- **Combining clinical, genetic, "omics", and biomarker data**: Integrate these data through AI technology to provide more comprehensive patient management plans, thereby improving patient prognosis.

4. **Addressing data challenges**:
- **Data availability and validation**: Emphasize the importance of high - quality training data and discuss how to conduct independent validation to ensure the performance and robustness of the model.
- **Handling bias and uncertainty**: Discuss how to handle ethnic, gender, body type, and referral biases, as well as inconsistent expert labels in the case of limited data.

5. **Promoting personalized medicine**:
- **Extracting new imaging markers**: Extract information that is under - utilized in traditional analysis processes through AI technology to achieve in - depth phenotyping of cardiovascular diseases.

In summary, through a detailed review of the applications of AI in cardiac MRI, this paper aims to show how AI technology can improve the efficiency and clinical value of cardiac MRI at multiple levels, thereby promoting the diagnosis and treatment of cardiovascular diseases.