A new federated learning (FL) challenge organized by the America College of Radiology® (ACR®), Harvard Medical Schools’ Mass General Brigham, University of Colorado, NVIDIA and the National Institutes of Health National Cancer Institute invites data scientists, informaticists and medical physicists to submit models for breast density estimation using distributed or federated learning. This effort will promote the development and fair evaluation of different FL algorithms, with the goal of creating generalizable models for breast density estimation that can be used across different systems.
Federated learning makes multi-institutional data sets available for analysis without the need for data sharing, protecting patient privacy and managing data security by keeping the data private. Algorithms that estimate breast density have the potential to play an important role in the diagnosis and early detection of breast cancer. Accurate estimation of breast density early in a treatment plan could help clinicians balance the benefits and risks of additional medical imaging procedures.
During the challenge, participants will develop, train and test models against digital mammographic imaging screening trial data from more than 33 institutions totaling over 100,000 images from more than 21,000 patients. The challenge training phase is live for participants to submit their models for training and ranking. Final ranking will be performed in the test phase starting on Aug. 15. Winners will share their proceedings at the 2022 Medical Image Computing and Computer Assisted Intervention Conference.
For more information on the competition, contact Kendall Schmidt.