Rad AI, a startup that developed an AI platform for radiology, is teaming up with Google to use its cloud and large language models to streamline workflows and reduce radiologists' administrative burdens.
Founder Dr. Jeff Chang, a radiologist, started the company because he was troubled by high error rates, radiologist burnout and rising imaging demand despite a worsening shortage of U.S. radiologists. Rad AI uses state-of-the-art machine learning and AI to automate repetitive tasks for radiologists and automate workflow for health systems.
Rad AI's software is in use at 8 of the ten largest private radiology practices in the U.S.
As part of the partnership with Google, Rad AI will use the tech giant's cloud platform and AI tools, including MedLM, a family of foundation models fine-tuned for healthcare industry use cases, including Gemini-based models in the future.
"This partnership represents an exciting leap forward in our commitment to transforming the radiology reporting landscape,” said Doktor Gurson, co-founder and CEO of Rad AI, in a statement. “Through this unique collaboration with Google, we can dramatically accelerate our mission of reducing radiologists' burnout, streamlining workflow, and ultimately improving the quality of patient care."
Medical imaging is critical for diagnosing patients, and billions of imaging examinations are performed globally each year. Imaging data accounts for about 90% of all healthcare data, according to one study, and the number of images continues to grow, increasing the workload for radiologists.
Radiologists spend most of their time dictating reports based on these images. Rad AI is addressing this increase with its AI-enabled reporting platform, Rad AI Reporting, which can reduce words dictated by up to 90%, according to the company.
The startup says 30% of U.S. radiology practices and health systems use its software, collectively caring for more than 50 million patients annually.
The collaboration with Google Cloud will result in platform enhancements with the latest gen AI technology, according to Rad AI executives.
Rad AI will be enhancing its platforms—Rad AI Omni Impressions and Rad AI Reporting—with additional domain-aligned gen AI models from Google, like MedLM. This will improve Rad AI’s ability to automatically generate more of the radiology report, customized to each radiologist’s preferred language and style, further saving radiologists time while improving report quality and consistency.
Rad AI also will rapidly increase the size and complexity of its Rad AI Omni Impressions and Rad AI Reporting gen AI models. Currently, Rad AI’s models reduce clinical error rates in reporting by nearly half compared to radiologist baseline, for the most complex computerized tomography (CT) angiogram reports, the company said.
“Radiology is a field that stands to see immediate high-value impact from advancements in generative AI, and radiology reporting is an area where this technology can have a meaningful impact,” said Aashima Gupta, global director of healthcare strategy and solutions, Google Cloud, in a statement. “As the number of medical images continues to grow, our goal is to enable the ecosystem and help our customers equip radiologists with the latest generative AI capabilities not only to help manage workflows but also to expedite patient treatment through faster and more accurate diagnoses.”
Rad AI pulled in $25 million in series A funding in 2021 to drive further development and commercialization of its AI platform. The company is backed by ARTIS Ventures, OCV Partners, Kickstart Fund and Gradient Ventures, which is Google's AI-focused fund.