Nuance maps out platform to support use of AI in radiology
Plans for the Precision Imaging Network could offer a way to better manage algorithms, enabling multiple entities to apply them to patient images.
While much hope and hype surrounds the use of artificial intelligence in healthcare, a company that provides supportive report writing services to radiologists and other connection solutions is developing a platform that will offer a cloud-based approach to improve the application of AI in imaging.
Nuance Communications expects the platform, which it calls the Precision Imaging Network, will enable healthcare organizations to use AI-based algorithms of their choice that can be easily accessed by radiologists and other imaging professionals in the course of their workflow.
Nuance expects to use connections that it already has with imaging professionals – it says that 80 percent of the nation’s radiologists already use its solutions, one of which (PowerScribe) enables the creation of image reports through dictation, while PowerShare supports the distribution of standards-based images.
Executives say the approach also will enable the secure sharing of AI-enhanced imaging data with other key healthcare stakeholders, such as health plans and life science organizations, to use their own algorithms in support of image assessment – for example, for facilitating prior authorization efforts or identifying potential candidates for clinical trials.
The network concept was introduced at the recent annual conference of the Radiological Society of North America conference in Chicago. While still in the early stages of deployment, the cloud-based approach that already has a footprint within radiological suites of healthcare systems could offer a solution for several vexing issues that have held back AI adoption in healthcare.
Nuance says the new platform can bring patient-specific data and insights from diagnostic imaging into existing clinical and administrative workflows. Nuance expects to have sufficient bandwidth to scale its offerings, because it runs on Microsoft’s Azure cloud platform. Nuance is in the process of being acquired by Microsoft, which announced a $16 billion acquisition in April.
In recent years, researchers have developed several algorithms related to diagnosis and patient care, but the challenge has been enabling them to be deployed at scale, within workflow and in a way that they can be easily managed, said Peter Durlach, Nuance’s chief strategy officer. Bringing advanced computing intelligence to imaging is important, because 80 percent of all hospital and health system visits include at least one imaging study.
“While we’re really focused on use by radiologists, it’s become clear that (using AI in imaging) is not the only use case,” Durlach said. “It’s not just a question of how to scale AI for imaging to drive meaningful impact – it’s the same problem for a variety of use cases. For example, intensivists could benefit from AI downstream from when the images were taken.”
The platform also can offer these benefits, both within healthcare organizations as well as in other entities. For example, about half of all follow-up recommendations derived from imaging exams end up in electronic medical records but are never acted upon, and Nuance believes its platform can help to aid patient care.
Additionally, health plans increasingly are looking at images or imaging reports to help support early detection of medical problems or to run their own AI algorithms against them, for example, to aid in prior authorizations of approved medical treatment. Or life science companies might be able to make earlier determinations about whether patients in the early stages of a disease might be identified through algorithms as likely to benefit from a clinical trial.
Nuance’s solutions already are making use of forms of AI, for example to better structure dictated reports. The company hopes a unified solution, such as the Precision Imaging Network, can help migrate clinicians to accept AI to assist with some clinical purposes.
The goal is for a range of providers to be able to use it, from facilities operating a vendor-neutral archive to systems that have facilities operating multiple picture archiving and communications solutions (PACS), Durlach said.