IBM Watson to bring AI to brain bleed detection
IBM Watson is to collaborate with Israel-based MedyMatch to use AI-based decision tools to help identify bleeds on the brain.
The eventual goal is to integrate MedyMatch's solution into Watson's imaging offerings.
MedyMatch's deep learning algorithm analyses head scans and combines its results with other patient data and clinical data insights to highlight which areas of the brain could be experiencing a bleed caused by head trauma or stroke. The entire process from head scan to bleed suggestions is intended to take under five minutes.
The company is looking to address image-based misdiagnoses which it says stands at 30% of all medical imaging decisions – 80% of which are perceptual errors missed by the human eye.
According to the American Heart Association and American Stroke Association, in the US alone stroke is the fourth leading cause of death and will be costing the health system around $240 billion per year by 2030. Worldwide, stroke is the second leading cause of death.
Developing a tool that could identify brain bleeds quickly can drastically improve its associated prognosis, preventing extensive brain and nerve damage and even death.
“The opportunity to license our deep vision application to IBM Watson Health creates a unique value proposition in healthcare,” said Gene Saragnese, chairman and CEO of MedyMatch. “Engaging closely with IBM allows for a near-zero footprint implementation at a customer location delivering ‘A.I. to the bedside’ where I believe the future of healthcare lies.“
Moving into the area of image analysis is a relatively new area for Watson which has so far only seen one acquisition of Chicago-based Merge Healthcare. Its biggest competitor in Google's DeepMind Health is using its capabilities to improve eye scanning for signs of diabetic retinopathy whilst a team of its researchers recently trained an algorithm to detect breast cancer from digital pathology slides.
For MedyMatch, the five-year collaboration with Watson is a first with a major AI company since its founding in 2013 and one of the first deals made under the reign of former Philips Imaging Systems CEO Gene Saragnese.
The company is currently investigating its algorithm in an intracranial bleed clinical trial and is working towards a PMA Class 3 US FDA approval.
