Abstract #0056

SGDNet: Synthetic Data Guided Supervised Deep Learning Network for Multi-Component T1$$$\rho$$$ Mapping in the Knee Joint

Dilbag Singh^1,2, Ravinder R. Regatte^3,4, and Marcelo V. W. Zibetti^1,5

¹Bernard and Irene Schwartz Center for Biomedical Imaging, Department of Radiology, New York University Grossman School of Medicine, New York, NY, United States, ²Center for Advanced Imaging Innovation and Research (CAI2R), Department of Radiology, New York University Grossman School of Medicine, New York, NY, United States, ³Bernard and Irene Schwartz Center for Biomedical Imaging, Department of Radiology, New York University Grossman School of Medicine, NEW YORK, NY, United States, ⁴Center for Advanced Imaging Innovation and Research (CAI2R), Department of Radiology, New York University Grossman School of Medicine, NEW YORK, NY, United States, ⁵Center for Advanced Imaging Innovation and Research (CAI2R), Department of Radiology, New York University Grossman School of Medicine, New York, India

Synopsis

Keywords: Cartilage, Machine Learning/Artificial Intelligence

Motivation: The nonlinear least squares (NLS)-based estimation of multi-component T1$$$\rho$$$ maps in the knee joint is highly time-consuming.

Goal(s): The main goal is to propose Synthetic Data-Guided Supervised Deep Learning Network (SGDNet), which uses synthetic target data for training, eliminating the need for extensive reference data.

Approach: A customized loss function is designed to ensure consistency between actual and predicted T1$$$\rho$$$ values and between the measured and predicted MR data. A self-attention module is also integrated into SGDNet.

Results: SGDNet produces T1$$$\rho$$$ maps similar to those obtained by NLS and regularized NLS (RNLS) while being, on average, 66 times faster in fitting time.

Impact: Our results indicate that SGDNet is faster for whole knee joint T1$$$\rho$$$ mapping than the NLS-based methods, with comparable errors. Thus, SGDNets is an alternative to replace NLS and RNLS for T1$$$\rho$$$ mapping when computational time is an issue.

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.