Perelman School of Medicine at the University of Pennsylvania is working with Intel to advance a new distributed machine learning technique – one that allows for researchers at disparate organizations around the world to work together to develop new deep learning initiatives without having to share data.
WHY IT MATTERS
That approach, which aims for big advances in collaborative AI research while still maintaining patient privacy, is called federated learning.
Researchers at Penn Medicine have published findings that a federated-learning approach to medical-imaging AI could train a model to more that 99% of the accuracy of a model trained in the traditional, nonprivate method.
So Penn is working with 29 research institutions from the U.S. – as well as in Canada, the U.K., Germany, the Netherlands, Switzerland and India – using Intel technology to deploy a federated-learning approach to develop new deep learning models for identifying brain tumors.
In order to train and build a model to detect a brain tumor that could aid in early detection and better outcomes, these researchers need access to large amounts of relevant medical data. By using federated learning, they will be able to work together on building and training an algorithm to detect a brain tumor while protecting sensitive medical data.
Penn Medicine and Intel say the research will be trained on the largest brain tumor dataset to date, without identifiable patient data leaving the individual collaborators.
THE LARGER TREND
Even if more evidence is needed, AI and machine learning are proving their worth in the field of medical imaging. In April, for instance, a study on AI-augmented diabetic retinopathy screening indicated that such programs are cheaper than human grading – that a deep learning system would save on the roughly two minutes of human labor required to grade each case.
The Penn Medicine and Intel initiative is funded by a three-year, $1.2 million grant from the Informatics Technology for Cancer Research program of the National Institutes of Health.
Some of the collaborating institutions planning to participate in the first phase of this federation are 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.
ON THE RECORD
“It is widely accepted by our scientific community that machine learning training requires ample and diverse data that no single institution can hold,” said Dr. Spyridon Bakas at Penn’s Center for Biomedical Image Computing and Analytics, and principal investigator on this project.
“This year, the federation will begin developing algorithms that identify brain tumors from a greatly expanded version of the International Brain Tumor Segmentation challenge dataset. This federation will allow medical researchers access to vastly greater amounts of healthcare data while protecting the security of that 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, in a statement. “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.”
Email the writer: [email protected]
Healthcare IT News is a publication of HIMSS Media
Source: Read Full Article