Abstract #3840

Denoising 7T Structural MRI with Conditional Generative Diffusion Models

Binxu Li^1,2, Yixin Wang^2,3, Yihao Liang⁴, Mackenzie Carlson⁵, Phillip DiGiacomo⁶, Hossein Moein Taghavi⁶, Julian Maclaren⁶, Meghan Bell⁶, Elizabeth Mormino⁵, Victor W. Henderson⁵, Greg Zaharchuk⁶, Brian Rutt⁶, Wei Shao⁷, Marios Georgiadis⁶, and Michael Zeineh⁶

¹Department of Electrical Engineering, Stanford University, Palo Alto, CA, United States, ²Equal Contribution, Stanford University, Palo Alto, CA, United States, ³Department of Bioengineering, Stanford University, Palo Alto, CA, United States, ⁴Department of Electrical and Computer Engineering, Princeton University, Princeton, NJ, United States, ⁵Department of Neurology and Neurological Sciences, Stanford School of Medicine, Palo Alto, CA, United States, ⁶Department of Radiology, Stanford School of Medicine, Palo Alto, CA, United States, ⁷Department of Electrical & Computer Engineering, University of Florida, Gainesville, FL, United States

Synopsis

Keywords: AI Diffusion Models, High-Field MRI, Denoising structural 7T MRI, Generative AI

Motivation: 7T MRI offers ultra-high resolution, but the commonly used long acquisitions are challenging, especially for elderly subjects.

Goal(s): Efficiently denoising 7T MRI images from a short acquisition without sacrificing image quality.

Approach: We introduced the 7T Conditional Diffusion Model (7TCDM), a conditional diffusion model derived from generative AI that is trained on raw acquisitions and references high-quality low-noise images to guide the denoising process and reconstruct high-quality images.

Results: 7TCDM significantly reduced noise and artifacts, improving image quality over each acquisition and outperforming the Convolutional Neural Network (CNN)-based model in maintaining image details.

Impact: Our newly introduced 7T Conditional Diffusion Model (7TCDM) enables faster MRI acquisition by providing high-quality denoised images from shorter scans, increasing the feasibility of scanning patients in shorter times while preserving essential anatomical details.

How to access this content:

For one year after publication, abstracts and videos are only open to registrants of this annual meeting. Registrants should use their existing login information. Non-registrant access can be purchased via the ISMRM E-Library.

After one year, current ISMRM & ISMRT members get free access to both the abstracts and videos. Non-members and non-registrants must purchase access via the ISMRM E-Library.

After two years, the meeting proceedings (abstracts) are opened to the public and require no login information. Videos remain behind password for access by members, registrants and E-Library customers.

Click here for more information on becoming a member.