The promise of AI in healthcare and life sciences is profound. It can help physicians and researchers prevent disease, speed recovery, conduct complex genomics processing, analyze medical images – and, quite simply, save lives.
In this post we’ll go through some of the real-world applications that Intel® AI technologies have enabled – and the impressive results they’re delivering.
Detecting cancer in real time: A pair of software developers, Mike Borozdin and Peter Ma, created an AI-based app called Doctor Hazel that detects skin cancer in real time by using a high-powered endoscope camera and the Intel® Movidius™ Neural Compute Stick (NCS) to analyze photos of moles. According to the Skin Cancer Foundation, half of the population of the U.S. will be diagnosed with some form of skin cancer by the time they reach age 65. If detected early, however, the survival rate is extremely high – that’s what Doctor Hazel aims to accomplish.
Reducing ER visits, saving big money: The Chilean chronic care management company AccuHealth is using AI to catch health problems before they send patients to the emergency room. Its system has reduced ER visits by 42% while saving six of the largest insurers an average of 50% per patient. It works by using biometric sensors to track patients’ health in their homes and sending the data to a data center. AI applications powered by Intel® Xeon® processors then rapidly analyze biometric, demographic, historical, and other data to identify warning signals and alert healthcare providers.
Finding patterns in big data: The University of California, San Francisco (UCSF) developed a deep learning analytics platform that aggregates data from disparate sources, including genomic sequencing information and medical devices. Dubbed the SmarterHealth Initiative, it was built using the Intel BigDL framework, which enables it to analyze vast amounts of data and find patterns that help doctors address patient illnesses.
Predicting patient problems: Predicting when a patient is about to take a turn for the worse is the goal behind an AI-based platform called PALM (Patient-centered Analytics and Learning Machine). Created by the academic medical center and hospital group Montefiore Health System, the system is powered by Intel® Xeon® Scalable processors, which help address the large amounts of system memory needed for training. PALM mines data in near real time to identify patterns that indicate a problem is emerging. It is being used to monitor patients in the intensive care unit, for example, to identify when a patient is at risk of respiratory failure.
Dramatic productivity improvements for radiologists: Intel worked with Philips to develop tools to help radiologists identify issues faster and more accurately. The team created deep learning inference models for two healthcare applications: X-rays for bone-age-prediction modeling, and CT scans for lung segmentation. Philips’s goal is to offer AI to customers on hardware they already have deployed in the field. Using the Intel® OpenVINO™ toolkit and other optimizations, Philips achieved a speed improvement of 188 times for the bone-age-prediction model, and 37 times for the CT scan model. (Learn more in this white paper).
Perhaps you noticed that the one common denominator for each of these projects is Intel, an indication of the depth and breadth of Intel AI solutions. Intel delivers a versatile mix of hardware, software, optimizations for popular deep learning frameworks, and other tools to bring AI out of the theoretical realm and into practice across a range of vertical industries, with healthcare being one prime example.
As Dr. Xavier Urtubey, founder and CEO of AccuHealth, put it, “Working with Intel means we have access not just to healthcare innovators, but broad technology innovators … to see the next thing coming and understand how we can do what we do better.”