Intel Labs and Penn Medicine use privacy-preserving AI to identify brain tumors

What you should know:

– Intel Labs and the Perelman School of Medicine at the University of Pennsylvania (Penn Medicine) have completed a joint research study using federated learning – a distributed machine learning (ML) artificial intelligence (AI) approach – to help international health and research institutions identify malignant brain tumors.

– The largest medical federated learning study to date with an unprecedented global dataset examined from 71 institutions on six continents, the project demonstrated the ability to improve brain tumor detection by 33%.

Using Federated Learning to Improve Brain Tumor Detection

In 2020, Intel and Penn Medicine announced the agreement to cooperate and use federated learning to improve tumor detection and improve treatment outcomes for a rare form of cancer called glioblastoma (GBM), the tumor most common and deadliest adult stroke with a median survival of just 14 months after standard therapy. Although treatment options have expanded over the past 20 years, there has been no improvement in overall survival rates. The research was funded by the National Cancer Institute’s Computing Technology for Cancer Research Program of the National Institutes of Health.

Penn Medicine and 71 international healthcare/research institutions have used Intel Federated Learning hardware and software to improve detection of rare cancer boundaries. A new state-of-the-art AI software platform called Federated Tumor Segmentation (FeTS) has been used by radiologists to determine the boundary of a tumor and improve the identification of the “operable region” of tumors or “heart of the tumor”. The radiologists annotated their data and used Open Federated Learning (OpenFL), an open source framework for training machine learning algorithms, to run the federated training. The platform was trained on 3.7 million images of 6,314 GBM patients across six continents, the largest brain tumor dataset to date.

Data accessibility has long been an issue in healthcare due to state and national data privacy laws, including the Health Insurance Portability and Accountability Act (HIPAA). For this reason, large-scale medical research and data sharing has been nearly impossible to achieve without compromising patient health information. Intel’s Federated Learning hardware and software respects data privacy concerns and maintains data integrity, privacy, and security with confidential computing.

The Penn Medicine-Intel result was achieved by processing large volumes of data in a decentralized system using Intel’s Federated Learning technology combined with Intel® Software Guard (SGX) extensions, which remove barriers sharing data that has historically prevented collaboration on cancers and similar diseases. to research. The system addresses many data privacy concerns by keeping the raw data in the data holders’ compute infrastructure and only allowing model updates computed from that data to be sent to a central server or an aggregator, not the data itself.

“All the computing power in the world can’t do much without enough data to analyze,” said Rob Enderle, principal analyst, Enderle Group. “This inability to analyze data that has already been captured has significantly delayed the massive medical breakthroughs promised by AI. This study of federated learning presents a viable path for AI to progress and realize its potential as a most powerful tool to combat our most difficult ailments.

“Federated learning has enormous potential in many fields, especially in healthcare, as our research with Penn Medicine shows. Its ability to protect sensitive information and data opens the door for future studies and collaborations, especially in cases where datasets would otherwise be inaccessible Our work with Penn Medicine has the potential to positively impact patients around the world and we look forward to continuing to explore the promise of learning Federated — Jason Martin, Principal Engineer, Intel Labs

To advance disease treatment, researchers need access to vast amounts of medical data – in most cases, datasets that exceed the threshold that an institution can produce. Research demonstrates the effectiveness of federated learning at scale and the potential benefits the healthcare industry can realize when multisite data silos are unlocked. Benefits include early detection of disease, which could improve quality of life or increase a patient’s lifespan.

Through this project, Intel Labs and Penn Medicine created a proof of concept for using federated learning to gain insights from data. The solution may significantly affect health care and other fields of study, especially among other types of cancer research. Specifically, Intel developed the OpenFL open-source project to enable customers to take cross-silo federated learning into the real world and confidently deploy it on Intel SGX. Additionally, the new FeTS initiative was established as a collaborative network to provide a platform for continued development and to encourage collaboration with the FeTS platform and Intel’s OpenFL open source toolkit, both available on GitHub.