Jeff Chang, co-founder of Rad AI, on VatorNews podcast

Rad AI uses machine learning and artificial intelligence to make radiologists more efficient

Steven Loeb and Bambi Francisco Roizen speak with Jeff Chang, co-founder of Rad AI, a company that uses machine learning and artificial intelligence to make radiologists more efficient.

Rad AI's first product, called Omni, automates the impression section of a radiologist's report, in which they summarize and synthesize the conclusions of what they've seen. The technology can write in each radiologist’s preferred language, meaning that it automatically generates language that sounds like that person. The radiologist can then review the impression section, and then sign off. Since launching in 2019, the company has also released a second product, called Continuity, a coordination product for health systems.

The company recently raised a $25 million round of funding, bringing its total capital to $34 million.

Our overall goal with these podcasts is to understand how technology is radically changing healthcare: the way we screen, treat and measure progress and outcomes. How we’re empowering the consumer. Whether we’re creating productivity that drives economic costs down? And how tech advancements change the role of the doctor.

Highlights from the interview:

• Radiology has always been ahead of much of the rest of medicine in terms of adoption of technology, so they've been using voice recognition in radiology for the better part of 20 years, and digital display of images for much longer. 
• Radiologists typically looks at hundreds of studies a day, each one of which involves anywhere from one to several thousand images, and they're also dictating reports non-stop. There have also been more requirement that they make sure to include more patient history, potential additional data streams from AI, while also ensuring the correct recommendations are being made. This is leading to burn out.  
• Rad AI's main product works with the radiologist to automate the impression of the report, which is the last section where the radiologist summarizes the findings, draws conclusions, and has follow-up recommendations. The company automatically generates this in less than one second, and it's customized to the radiologist's language by pre-training on all of their prior reports.
• The combination of the radiologist plus Rad AI results in catching more errors that the radiologist otherwise would have missed and helps make reports more accurate overall. About 5% of the time the radiologist will go back and fix an error that they made in the findings as a result of something they saw in Rad AI's impression. 
• On average, the company saves radiologists about an hour of work per nine hour shift, which increases productivity and reduces burn out.
• The company's customers are health systems and radiology groups, and while it doesn't currently work with health insurers, that is something that it was discussing for its third product, which will involve flagging discrepancies between imaging and reports, ideally in real-time. This will help the radiologist catch mistakes before they're made, and improving the overall quality and accuracy the reports.

Thanks to our sponsors:

Clinically proven to help increase motivation to quit while improving quit rates, Pivot is a holistic approach that combines an FDA-cleared carbon monoxide sensor with personalized content, community and live 1:1 coaching via mobile app that can be accessed at any point during a participant's journey to quit.  

Subscribe to our podcasts to get our interviews and shows as soon as they're published!