GE HealthCare's Innovative Approach to Transforming All Facets of Patient Treatment Through AI
The international healthcare sector is grappling with two substantial challenges: a deluge of patient data that largely goes unused due to its unstructured nature, and widespread operational inefficiencies that strain resources and put undue pressure on medical professionals. GE HealthCare posits that artificial intelligence (AI) holds the solution to these dual conundrums, and they're wholeheartedly investing in AI deployment throughout their organization.
"It's revolutionary from my perspective," asserts Taha Kass-Hout, GE HealthCare's Global Chief Science and Technology Officer. Currently, nearly 97% of hospital system data remains untouched because it's mainly unstructured, consisting of videos, images, wave streams, and more.
GE HealthCare isn't merely dabbling in AI; they're fully immersing themselves. The company's commitment to AI innovation is evident in its statistics – for three consecutive years, they've led the FDA's list of AI-enabled device authorizations with over 80 approvals. Their ambition now is to more than double that number to over 200 authorizations within the next three years, while also expanding their cloud-enabled product offerings threefold.
Revolutionizing Cancer Care Using Multi-Agent AI
One of their most innovative projects is the Health Companion research initiative, which investigates how multiple AI agents can mimic the collective knowledge of a multidisciplinary cancer care team. Each AI agent specializes in a distinct medical domain – radiology, oncology, genetics, pathology, and beyond – independently analyzing patient data before a supervisory agent synthesizes their findings.
Dr. Taha Kass-Hout, GE HealthCare's Global Chief Science and Technology Officer, illustrates this concept in complex cases. "Just to assemble a tumor board can take weeks, sometimes even longer," he explains. The Health Companion initiative aims to significantly reduce this timeframe while broadening access to expert-level cancer care analysis.
The vision for the system's practical application extends beyond basic analysis. For instance, the AI agents can differentiate between symptoms associated with treatment and those indicating disease progression, such as cancer spread. Upon detecting concerning patterns, the system can alert the care team, suggesting subsequent actions, thus enabling more proactive and timely interventions. This functionality is designed with robust safety and explainability principles, ensuring healthcare providers can understand and authenticate the AI's decision-making process.
Aiding Radiologists with Sophisticated AI Solutions
With radiologist shortages worsening while imaging volumes increase annually by 3-5%, GE HealthCare is developing intricate AI solutions to assist specialists in working more efficiently. Their full-body X-ray foundation model research project, trained on 1.2 million anonymized images, represents a significant breakthrough in medical imaging AI.
In preliminary tests, the model excellently performs three critical tasks: segmentation (distinguishing images into distinct anatomical parts), classification (categorizing images based on medical conditions), and visual localization (precisely identifying features like fractures or tumors). Besides these fundamental capabilities, GE HealthCare is investigating the system to automatically generate medical reports, potentially eliminating hours of administrative work.
"Consider a single model capable of performing all these tasks across all body structures," Dr. Taha Kass-Hout, GE HealthCare's Global Chief Science and Technology Officer, suggests. Developed in collaboration with Amazon Web Services, this technology aims to lessen cognitive burden on healthcare professionals while equipping them with efficient and dependable diagnostic tools.
Furthermore, GE HealthCare is developing AI for specialized purposes to expedite mammography screenings. Radiologists can more quickly identify normal cases (which make up approximately 90% of all breast screening exams) using this tool, assisting them in focusing more attention on potentially suspicious cases.
Streamlining Cancer Care with CareIntellect
GE HealthCare has also unveiled CareIntellect for Oncology – a new cloud-based application that consolidates multi-modal patient data into a unified view. Initially planned for wide release in the US by 2025, with subsequent expansion to Canada, the UK, and Ireland, this platform uses generative AI to revolutionize how oncologists manage patient care.
Presently, physicians often spend hours familiarizing themselves with a patient's history and care status, especially for new patients. They must manually search through multiple databases and amalgamate information from various sources - including medical images, clinical notes, lab results, and device readings. CareIntellect aims to decrease this time-consuming process from several hours to mere minutes.
The platform includes practical features addressing daily clinical challenges. It can flag overlooked lab work that might postpone the next round of treatment, track adherence to treatment protocols, and help match patients to potential clinical trials by comparing their health records to trial criteria. Additionally, CareIntellect is compatible with providers' existing single sign-on systems, eliminating the need for multiple log-ins.
Tampa General Hospital and UT Southwestern Medical Center serve as early evaluators of this technology. "We were amazed by how promptly GE HealthCare developed the application to include breast cancer within weeks," observes Dr. Peter Chang, Senior Vice President at Tampa General Hospital. "We eagerly anticipate incorporating this AI-enabled functionality into our care teams to help clinicians invest time where it matters most – delivering exceptional patient care."
Addressing Global Healthcare Challenges
These advancements arrive at a critical juncture for global healthcare. Approximately 4.5 billion people globally – almost half of the world's population – require better access to healthcare services. Concurrently, the healthcare workforce encounters unprecedented difficulties, with over 50% of physicians reporting burnout, and hospitals encountering 100% staff turnover every five years.
GE HealthCare collaborates with organizations such as the Bill & Melinda Gates Foundation to expand AI-enabled portable ultrasound technology to underserved regions. By integrating AI insights with portable devices, they seek to empower non-specialist healthcare workers to deliver high-quality diagnostic services in remote locations.
Moving forward, GE Healthcare envisions AI becoming a fundamental component in every medical device and clinical process. They're equally thrilled about upcoming technologies like theranostics, which merges diagnostics and therapeutics to foster more customized cancer treatments.
As Dr. Kass-Hout stresses, AI holds immense promise. Nevertheless, its safety and efficacy hinge on responsible development and use. Above all, patient security is paramount.
The message is unmistakable: AI isn't phasing out healthcare personnel – instead, it's equipping them to work more proficiently and productively while widening the reach of top-notch medical care on a global scale. With these advancements, GE Healthcare is striving to fashion a future where healthcare knows no bounds.
GE HealthCare's commitment to AI in patient care is evident in their ambition to more than double their FDA-authorized AI-enabled device authorizations to over 200 within three years. This expansion aims to enhance their cloud-enabled product offerings threefold (ge healthcare, healthcare, ai, authorizations).
The Health Companion research initiative is one of GE HealthCare's innovative AI projects. It uses multiple AI agents to mimic a multidisciplinary cancer care team, analyzing patient data independently and synthesizing their findings to provide expert-level analysis, potentially reducing the time required for tumor board meetings (ge healthcare, patient care, cancer care, ai, tumor board, Health Companion).
GE HealthCare's full-body X-ray foundation model project, trained on 1.2 million anonymized images, represents a significant breakthrough in medical imaging AI. The model can perform segmentation, classification, and visual localization tasks, with the potential to automatically generate medical reports, reducing administrative work for radiologists (ge healthcare, radiologists, medical imaging, ai, X-ray, administrative work).
