By: Yaron Blinder PhD, Senior Product Manager for X-ray products at Zebra-Med.
Radical AI was founded in 2017 to develop high impact, deep learning-based solutions for radiology workflow. When analyzing unmet needs in the field, a few key points quickly became clear – first and foremost, that radiology as a specialty is undergoing a real crisis. With an estimated 5 billion exams (and rising) currently performed annually worldwide, less than 10 radiologists per 100,000 patients (first-world average!), and the long, expensive process required to train new radiologists, the need for AI assistance is urgent. This is especially true considering half the world’s population has no access to radiology services at all.
Second, one type of exam is responsible for about 40% of all exams (2 billion annually) – the chest X-ray. The reason is that over the last couple of decades the chest X-ray has taken its place as standard procedure in many common protocols – from an outpatient coming in with a bad cough, to almost anyone undergoing surgery, and enormous public health screening programs for conditions such as tuberculosis. In many of these scenarios, the workflow burden is so severe that chest X-ray exams are hardly ever read by radiologists. Instead – physicians without proper training will take a brief look and make a quick, too-often incomplete and even wrong assessment.
With these insights in mind, we formed a strategic partnership with a large clinical data provider and developed an initial AI technology with promising results.
Earlier this year, a research team within Zebra led by Jonathan Laserson published a fantastic paper on a project called TextRay , in which they extracted a dataset of nearly 1M labeled chest X-ray studies with an ontology of the 40 most prevalent findings and proceeded to train classifiers for the top 12 findings with state-of-the-art performance. This was by far the biggest advancement in AI-powered understanding of chest X-ray to date, and the work was presented with great acclaim at the MICCAI 2018 conference.
It was around this time that the connection was made with Radical AI, and the potential for collaboration quickly became clear to both sides. By merging Radical AI’s efforts into Zebra, we have the opportunity to bring to market a platform of premier AI-powered chest X-ray products. I am tremendously excited and proud to join the Zebra team on this mission to create truly global impact through AI-powered radiology.
Yaron Blinder PhD
Senior Product Manager for X-ray products at Zebra-Med.