GE, UCSF partner to build deep learning algorithm library

The GE Healthcare logo

GE Healthcare and UC San Francisco have teamed up to build a library of deep learning algorithms. The partners plan to stock the library with algorithms that cut the time it takes for users of GE’s cloud and imaging machines to diagnose conditions.

Deep learning algorithms learn to spot patterns in images and other data. By pairing such algorithms with powerful computers and growing repositories of medical images and data, GE and rivals such as IBM think they can give clinicians the information they need to quickly diagnose patients and choose the most appropriate intervention.

Partnering with UCSF’s Center for Digital Health Innovation gives GE access to expertise to move it toward the realization of this vision.

FREE DAILY NEWSLETTER

Like this story? Subscribe to FierceBiotech!

Biopharma is a fast-growing world where big ideas come along every day. Our subscribers rely on FierceBiotech as their must-read source for the latest news, analysis and data in the world of biotech and pharma R&D. Sign up today to get biotech news and updates delivered to your inbox and read on the go.

“With this collaboration, these technologies will be applied to our clinical data and images to provide clinicians with actionable information in near real-time,” Dr. Michael Blum, associate vice chancellor for informatics at UCSF, said in a statement. “Together, we will develop tools and algorithms that will allow clinicians and researchers to identify problems and ask questions that are only achievable with vast computing power and datasets.”

The partners’ first goal is to create algorithms to support “high-volume, high-impact” imaging. These algorithms are expected to help healthcare professionals differentiate between normal results and those that need further investigation or treatment. GE cites an algorithm designed to spot collapsed lungs as an early example. If the algorithm can teach machines to identify collapsed lungs, physicians can act sooner, potentially improving health outcomes.

Over time, GE and UCSF plan to branch out beyond this initial area of focus. Electronic health records will join imaging technologies on the list of data sources used by the partners. This broadening of the data pool is intended to enrich algorithm development and boost sensitivity.

Long term, GE sees algorithms supporting everything from the automation of triage in routine care to the development of personalized therapies. For now, the approach has a lot to prove. By teaming up with UCSF, GE has positioned itself to find out whether deep learning algorithms can live up to the hype in healthcare.

Suggested Articles

By employing heart rate signals, physical activity and sleep quality, common Fitbit trackers may be able to predict the spread of the flu.

Nanox has raised $26 million to help fuel the development and commercialization of its Star Trek-inspired digital X-ray bed.

Oncology is clearly a major medical and societal issue, but one that sees too much focus from biopharmas at the expense of other killers.