Summary:
UCLA researchers have developed a software-based denoising method that makes cardiac MRI images clearer and more reliable, enabling more accurate heart function measurements without requiring longer scan times.
Background:
Cardiac MRI is a powerful imaging tool for assessing heart health because it can provide detailed pictures of how the heart muscle moves and pumps in both healthy and diseased states. Among its advanced techniques, DENSE MRI (Displacement Encoding with Stimulated Echoes) can provide very detailed measurements of myocardial strain (the stretching and contracting of heart muscle fibers). Strain analysis offers unique insights into heart function, and it is becoming increasingly important for diagnosing early-stage heart disease, monitoring progression, and guiding treatment decisions.
Despite its potential, DENSE MRI faces a major technical challenge: low signal-to-noise ratio (SNR). Because the technique relies on a very weak type of signal (the “stimulated echo”), the useful information is often buried under random noise. This results in images that are grainy, harder to interpret, and less reliable for quantitative analysis. In practice, low SNR limits the accuracy of strain measurements, reduces reproducibility, and can undermine clinical confidence. Traditional denoising methods—like smoothing filters—may remove some noise but often blur the images or eliminate subtle details that are essential for precise diagnosis.
To unlock the full clinical value of cardiac MRI and make DENSE a more practical tool for widespread adoption, there is a strong need for smarter methods that can selectively suppress noise while preserving fine anatomical and functional details.
Innovation:
UCLA researchers have developed a novel denoising technique that significantly improves the clarity of DENSE MRI images. This method can separate true signal from random noise by recognizing distinct statistical patterns, allowing noise to be suppressed without erasing important anatomical or functional details.
In testing with both healthy volunteers and patients with heart disease, the method improved both MRI magnitude and phase image quality, effectively removing background noise, and producing clearer, more interpretable data. Importantly, the denoised images remained consistent with the original data, meaning the measurements were more reliable but not distorted. Because this tool is software-driven, this denoising approach can be integrated into existing MRI systems, broadening access to advanced cardiac imaging for both clinical validation and eventual adoption in routine care.
Potential Applications:
- Cardiac MRI, such as clearer, more reliable strain analysis
- Heart disease diagnosis with improved detection rate and monitoring
- Standardized and higher-quality imaging data for clinical trials
- High-resolution research protocols without extending the scan times
- Integration into existing clinical MRI scanners as software add-on
- Broader imaging use and adaptable to other MRI sequences that suffer from noise
Advantages:
- Significantly improves MRI signal-to-noise ratio (SNR)
- Preserves fine anatomical and functional details
- Produces more consistent and reliable strain measurements
- Enhances image quality without extending scan time
- Software-based – compatible with current MRI hardware
- Enables higher-resolution imaging protocols previously limited by noise
- Applicable to both clinical and research settings
State of Development:
Successfully developed and demonstrated in vivo on both healthy volunteers and patient data.
Related Papers:
This technology is summarized as an conference abstract and presented at the annual meeting of the Society for Cardiovascular Magnetic Resonance (SCMR) on Jan 31, 2025
Reference:
UCLA Case No. 2025-204