Researchers at Intel Labs and the Perelman School of Medicine at the University of Pennsylvania (Penn Medicine) are using privacy-preserving technique called federated learning to train AI models that identify brain tumors. With federated learning, research institutions can collaborate on deep learning projects without sharing patient data.
AI shows great promise for the early detection of brain tumors, but it will require more data than any single medical center holds to reach its full potential,” said Jason Martin, principal engineer at Intel Labs. “Using Intel software and hardware and support from some of Intel Labs’ brightest minds, we are working with the University of Pennsylvania and a federation of 29 collaborating medical centers to advance the identification of brain tumors while protecting sensitive patient data.”
Penn Medicine and Intel Labs were the first to publish a paper on federated learning in the medical imaging domain, particularly demonstrating that the federated learning method could train a model to over 99% of the accuracy of a model trained in the traditional, non-private method. This paper was originally presented at the International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI) 2018 in Granada, Spain. The new work will leverage Intel software and hardware to implement federated learning in a manner that provides additional privacy protection to both the model and the data.
It is widely accepted by our scientific community that machine learning training requires ample and diverse data that no single institution can hold,” Bakas said. “We are coordinating a federation of 29 collaborating international healthcare and research institutions, which will be able to train state-of-the-art AI models for healthcare, using privacy-preserving machine learning technologies, including federated learning. This year, the federation will begin developing algorithms that identify brain tumors from a greatly expanded version of the dataset. This federation will allow medical researchers access to vastly greater amounts of healthcare data while protecting the security of that data.”
According to the American Brain Tumor Association (ABTA), nearly 80,000 people will be diagnosed with a brain tumor this year, with more than 4,600 of them being children. In order to train and build a model to detect a brain tumor that could aid in early detection and better outcomes, researchers need access to large amounts of relevant medical data. However, it is essential that the data remain private and protected, which is where federated learning with Intel technology comes in. By utilizing this approach, researchers from all partner organizations will be able to work together on building and training an algorithm to detect a brain tumor while protecting sensitive medical data.
Federated Workflow
In 2020, Penn and the 29 international healthcare and research institutions will use Intel’s federated learning hardware and software to produce a new state-of-the-art AI model that is trained on the largest brain tumor dataset to date — all without sensitive patient data leaving the individual collaborators. The subset of collaborating institutions expected to participate in initiating the first phase of this federation includes the Hospital of the University of Pennsylvania, Washington University in St. Louis, the University of Pittsburgh Medical Center, Vanderbilt University, Queen’s University, Technical University of Munich, University of Bern, King’s College London and Tata Memorial Hospital.
