The Critical Care Suite: A Case Study in Collaborative Development
Medical history was made on September 12, 2019, when Imaging Technology News announced the first FDA cleared, on-device use of artificial intelligence (AI) algorithms. GE Healthcare’s Critical Care Suite, co-developed with UCSF’s Center for Digital Health Innovation (CDHI), embeds AI algorithms into the GE Edison platform, helping to reduce the turnaround time for radiologists to review suspected pneumothorax, a type of collapsed lung that can result in death if not found quickly.
Why does this matter for patients?
It can sometimes take up to eight hours or more before a radiologist can review a chest x-ray, but every minute matters in life-threatening situations. The Critical Care Suite can be used right at the patient’s bedside, using AI to automatically analyze the image to search for pneumothorax. If the system suspects the presence of a collapsed lung, GE’s Edison platform sends an alert to both the technologist operating the x-ray equipment and to a radiologist to expedite the review and ensure a faster diagnosis confirmation.
How does this benefit hospitals?
The initial collaboration between GE and CDHI was focused on the imaging space to recognize clinical findings and to:
-
Improve productivity for the radiology workflow
-
Lower the risk of “negative” clinical consequences associated with the delays in radiologist read and report timeframes
-
Empower the care team to easily view radiology work product, accelerate clinical decision making, and streamline workflows resulting in improved patient experience and patient outcomes
CDHI’s role in the collaboration
CDHI provided the data scientists to develop the scientific proof-of-concept algorithm models as well as insights into the clinical context, relevance, and workflow of pneumothorax detection. As a result of this work, a collaborative development model was established to align the scientific discovery work at CDHI with the product development efforts within GE Healthcare resulting in an end-to-end product development pathway that was supported by shared terminology.
The scientific efforts at CDHI were enabled by the development of advanced technology infrastructure and data sets. CDHI’s technology team created a secure clinical image repository along with de-identification and annotation tools and pipelines, which were built within UCSF’s IT environment using standards to optimize the security profile and maximize scale across the UCSF campus. This discovery platform has grown over the course of the collaboration and created the foundation for what is now UCSF’s de-identified image repository.
For further reading:
-
Artificial Intelligence That Reads Chest X-Rays Is Approved by FDA (UCSF News)
-
FDA clears GE’s AI system for flagging collapsed lung (AI in Healthcare)
-
FDA Clears GE Healthcare's AI Triage Algorithm on X-Ray Device (Medscape)
-
First AI-powered X-ray System: GE Healthcare bags USFDA approval (Business Medical Dialogues)
-
GE Healthcare receives FDA approval for AI algorithms that prioritize chest x-rays (Health Imaging)
-
FDA clears GE Healthcare's X-ray AI Suite (Inside Digital Health)
-
FDA approves GE Healthcare's AI algorithms for chest x-rays (Radiology Business)
-
New algorithms speed time to review collapsed lungs (Health Data Management)
-
FDA clears GE Healthcare's AI platform for X-ray scans (Healthcare IT News)