Abstract #1198

Training Diffusion Probabilistic Models with Limited Data for Accelerated MRI Reconstruction with Application to Stroke MRI

Sidharth Kumar¹, Yamin Arefeen¹, Hamidreza Saber^2,3, and Jonathan I. Tamir^1,4,5

¹Chandra Family Department of Electrical and Computer Engineering, The University of Texas at Austin, Austin, TX, United States, ²Dell Medical School Department of Neurology, The University of Texas at Austin, Austin, TX, United States, ³Dell Medical School Department of Neurosurgery, The University of Texas at Austin, Austin, TX, United States, ⁴Dell Medical School Department of Diagnostic Medicine, The University of Texas at Austin, Austin, TX, United States, ⁵Oden Institute for Computational Engineering and Sciences, The University of Texas at Austin, Austin, TX, United States

Synopsis

Keywords: AI Diffusion Models, AI/ML Image Reconstruction, Diffusion probabilistic models. generative AI, machine learning

Motivation: Faster MRI reduces motion artifacts, costs, and time to treatment, crucial for stroke diagnosis; however, current protocols are lengthy, often making CT the preferred option.

Goal(s): To accelerate the stroke MRI protocol using diffusion probabilistic models, reducing scan time by half.

Approach: A foundation model is initially trained on a large public dataset, followed by fine-tuning on a smaller, contrast-specific dataset with a decayed learning rate and a short training period.

Results: The proposed method applied to retrospective data supports a 50% reduction in scan time, with improvements validated through both numerical error metrics and a qualitative neurologist assessment.

Impact: Training diffusion probabilistic models with limited data across various MRI contrasts holds substantial potential to accelerate diverse MRI protocols, addressing a critical unmet need in time-sensitive care scenarios, such as stroke diagnosis.

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.