Medical imaging accounts for 90% of total healthcare data
Intel AI Implements Deep Learning Solutions on Existing Infrastructure
GE Healthcare Provides Fast and Effective Treatment

Medical imaging is the largest and fastest-growing data source in the healthcare industry. While it accounts for 90% of all healthcare data, more than 97% of the information remains unanalyzed or unused.

Until now, processing such a vast amount of medical imaging data has required a long time from image acquisition to diagnosis and treatment. Furthermore, rapid analysis and delivery of images are essential, as there is a possibility that a patient's condition may deteriorate while waiting for a diagnosis.

Intel and GE Healthcare announced on November 30 that they have entered into a collaboration to provide AI solutions across various medical imaging formats to prioritize and streamline the process of treating patients.

Intel implements deep learning solutions on existing infrastructure with AI software solutions, and GE Healthcare provides more efficient and effective treatment. In addition, better value can be delivered to patients and healthcare providers by improving the decision-making process.
GE Healthcare is accelerating deep learning in point-of-care medical imaging by leveraging the Intel Distribution of OpenVINO Toolkit on Intel processor-based X-ray systems. Using these systems, X-ray technicians, critical care teams, and radiologists can immediately review critical clues to rapidly process patient diagnoses.

A key application of this technology is the early detection of pneumothorax, a fatal lung disease. Radiologists can utilize predictive algorithms optimized for scanning and detecting pneumothorax within seconds during point-of-care, enabling rapid response for clinical diagnosis and the reallocation of X-ray priorities.

In initial tests, when GE Healthcare applied the OpenVINO toolkit to its pneumothorax model, the Intel processor-based X-ray system recorded a detection speed 3.3 times faster compared to the unapplied model. Performance improved across all models through optimization, with the pneumothorax model showing the best results. Overall, the speed of pneumothorax detection decreased from over 3 seconds to within a few seconds.

Keith Bigelow, Senior Vice President of Portfolio Strategy at GE Healthcare Edison, said, “GE Healthcare achieved a 3.3x improvement in deep learning optimization during the initial testing phase with the OpenVINO toolkit running on existing Intel processors,” adding, “This enables early prioritization and communication regarding critical conditions, allowing us to provide faster treatment to our patients.”

He also added, "GE Healthcare can transform the work of radiologists and patient care by extending AI solutions across various imaging equipment through Intel technology."