Abstract #1472

Fibroglandular tissue segmentation and background parenchymal enhancement quantification in breast MRI using an anatomy-aware loss function

Ran Yan^1,2, Haoxin Zheng^1,3, Alex Ling Yu Hung^1,3, Tiffany Yu¹, Stephanie Lee-Felker¹, and Kyunghyun Sung^1,2

¹Department of Radiological Sciences, University of California, Los Angeles, Los Angeles, CA, United States, ²Department of Bioengineering, University of California, Los Angeles, Los Angeles, CA, United States, ³Department of Computer Science, University of California, Los Angeles, Los Angeles, CA, United States

Synopsis

Keywords: Segmentation, Machine Learning/Artificial Intelligence, Fibroglandular tissue; Background parenchymal enhancement; Breast cancer

Motivation: Fully automatic segmentation of fibroglandular tissue (FGT) and background parenchymal enhancement (BPE) quantification methods with high generalizability for different FGT levels are still lacking.

Goal(s): We aimed to improve the segmentation accuracy and generalizability across various FGT levels that accurately quantify FGT density and BPE.

Approach: A novel anatomy-aware loss function based on the variations in FGT level was applied in a fully automatic segmentation model training on breast MRIs.

Results: The accuracy of breast tissue segmentation, FGT density estimation, and BPE quantification were improved at various FGT levels.

Impact: The anatomy-aware loss function can help improve the generalization of the breast tissue segmentation model on patients with different breast densities, thereby enabling the model to be more widely used in fibroglandular tissue density estimation and background parenchymal enhancement quantification.

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.