Brain tumor detection can be improved by a third, federated learning study shows

Largest federated learning study from Intel Labs and Penn Medicine uses privacy-preserving AI to advance medicine
Amsterdam, December 6, 2022 – Intel Labs and the Perelman School of Medicine at the University of Pennsylvania (Penn Medicine) have conducted a joint study using federated learning – a distributed machine learning (ML) artificial intelligence (AI) approach – to develop global help health and research institutions identify malignant brain tumors. It is the largest medical federated learning study to date, examining an unprecedented global data set from 71 institutions on six continents. The Dutch Erasmus MC in Rotterdam is also involved. The project showed that it is possible to improve the detection of brain tumors by 33%.

“Federated learning has tremendous potential in many areas, especially in healthcare, as our research with Penn Medicine shows. The ability to protect sensitive information and data opens the door to future research and collaboration, especially in cases where data sets would otherwise be inaccessible . Our work with Penn Medicine has the potential to positively impact patients around the world, and we look forward to further exploring the promise of federated learning.”
– Jason Martin, Chief Engineer, Intel Labs

Making data available
Data availability has long been an issue in healthcare due to national data protection laws, including the General Data Protection Regulation (GDPR). This made it almost impossible to realize large-scale medical research and data sharing without compromising patients’ health information. Intel’s federated learning hardware and software meet data protection requirements and protect data integrity, privacy and security through confidential data processing.

The Penn Medicine-Intel result was achieved by processing large amounts of data in a decentralized system. This was done using Intel federated learning technology in combination with Intel® Software Guard Extensions (SGX). This technology removes barriers to data sharing that have previously stood in the way of collaboration in similar cancer and disease research. The system addresses many data privacy concerns by keeping the raw data within its own hospital network and only allowing model updates calculated from that data to be sent to a central server or aggregator, not the raw data.

Radiologist Prof. Dr. Smits and biomedical researcher Dr. Van der Voort from Erasmus MC: “We at Erasmus MC were able to contribute to improving automatic tumor detection through this federated learning study without having to send patient data. Automatic tumor detection is an important step to personalize and monitor a treatment, and to develop this methodology it is important to use data from many different institutions. With this collaboration, we were able to make it easy while maintaining control over our data.”

“Federated learning offers a breakthrough in ensuring secure multi-institutional collaborations. It provides access to the largest and most diverse data set ever seen in the literature. All data are kept at each institution at all times,” said senior author Spyridon Bakas, PhD, assistant professor of pathology and laboratory medicine and radiology at the University of Pennsylvania Perelman School of Medicine. “The more data we can feed into machine learning models, the more accurate they will become. This will in turn improve our ability to understand and treat even rare diseases, such as glioblastoma.”

To improve the treatment of disease, researchers need access to large amounts of medical data – in most cases data sets that exceed the threshold that one institution can produce. The research demonstrates the effectiveness of federated learning at scale and the potential benefits healthcare can achieve when multisite data silos are opened up. Benefits include early detection of disease, which can improve quality of life or extend a patient’s lifespan.

The results of the Penn Medicine-Intel Labs study have been published in the peer-reviewed journal, Nature Communications.

###

About the survey
In 2020, Intel and Penn Medicine announced the agreement to collaborate and use federated learning to improve tumor detection and improve treatment outcomes for a rare cancer called glioblastoma (GBM). Glioblastoma is the most common and fatal brain tumor in adults, with a median survival of only 14 months after standard treatment. Although treatment options have expanded over the past 20 years, overall survival has not improved. The research was funded by the Informatics Technology for Cancer Research program of the National Cancer Institute of the National Institutes of Health.

Penn Medicine and 71 international healthcare/research institutions used Intel’s federated learning hardware and software to improve detection of rare cancer frontiers. A new advanced AI software platform called Federated Tumor Segmentation (FeTS) was used by radiologists to determine the boundary of a tumor and improve the identification of the “operable region” of tumors or the “tumor core”. Radiologists annotated their data and used open federated learning (OpenFL), an open source framework for training machine learning algorithms, to perform the federated training. The platform was trained on 3.7 million images from 6,314 GBM patients from six continents, the largest brain tumor dataset to date.

What’s next: With this project, Intel Labs and Penn Medicine have created a proof of concept for using federated learning to extract knowledge from data. The solution can have a major impact on healthcare and other fields of study, especially within other types of cancer research. In particular, Intel developed the OpenFL open source project to enable customers to apply realistic cross-silo connected learning and confidently deploy Intel SGX. In addition, the new FeTS initiative was created as a collaborative network to provide a platform for continued development and encourage collaboration with the FeTS platform and Intel’s OpenFL open source toolkit, both available on GitHub.

More info: Intel partners with University of Pennsylvania to use privacy-protecting AI to identify brain tumors | Link to paper | Press kit | Technical briefing | Videos

About Intel
Intel (Nasdaq: INTC) is an industry leader creating world-changing technology that enables global progress and enriches lives. Inspired by Moore’s Law, Intel is constantly working to design and manufacture semiconductors to meet its customers’ toughest challenges. By building intelligence into the cloud, networks, peripherals and every type of computer, Intel unlocks the potential of data to change business and society for the better. For more information on Intel’s innovations, visit newsroom.intel.com and intel.com.

© Intel Corporation. Intel, the Intel logo and other Intel marks are trademarks of Intel Corporation or its subsidiaries. Other names and brands may be owned by others.

This article is a submission and is not the responsibility of the editors.

Leave a Comment