Cloud-native AI offers advantages in radiology efficiency

Kottler said the distinction between “cloud-based” and “cloud-native” technology is important. Many vendors have merely transferred traditional on-premise server applications into cloud environments without redesigning them to take advantage of the scalability, flexibility and integrated services available through modern cloud platforms. She said this is why today, there is a lot of partnering by radiology IT vendors with cloud IT vendors Amazon Web Services (AWS) and Microsoft Azure.

“You want to ask, are you just lifting and shifting what you did on premise and moving that into the cloud?” she said. “Or are you actually taking advantage of the capabilities ... within the cloud-native system?”

She said the next step in radiology IT systems is “AI-native” technology, which she described as software architectures fundamentally designed around artificial intelligence rather than attempting to bolt AI applications onto decades-old radiology infrastructure.

Much of healthcare IT infrastructure, including radiology systems, was originally developed in the 1990s or earlier, Kottler noted. These legacy systems often operate in silos and were never designed to support advanced AI applications or seamless interoperability.

“When you get to AI tools, if you want to plug them into your own workflow, you’re trying to tag them on as a peripheral to technology that just wasn’t built to support it,” she said.

Instead, Kottler said radiology needs foundational platforms that integrate AI directly into the operating environment and workflow. This would allow AI applications to function as core components of image interpretation and workflow management rather than secondary add-ons.

“Cloud-native and AI-native technology enables us to rethink from first principles what we want to do and create to be not only faster, but better,” she said.

Kottler argued that workflow redesign is essential as radiology faces ongoing staffing shortages and increasing imaging volumes. Rather than optimizing isolated portions of the workflow, she said healthcare organizations should reconsider the entire radiology process. She noted that a lot of how current imaging IT systems operate was originally built in the 1990s and early 2000s, prior to cloud computing and the rapid roll-out of AI technologies

“Our workflow is terrible in radiology,” she said. “If you were going to recreate the radiology workflow, you would not create what we have today. You’d create something fundamentally different.”

She also emphasized that cloud-native infrastructure offers cybersecurity advantages and is increasingly becoming a prerequisite for adopting next-generation AI.

As more enterprise imaging and healthcare IT vendors transition to cloud-based platforms, Kottler warned that healthcare organizations delaying migration risk falling behind.

“For anyone that’s slowly thinking about going to the cloud, you must do it now,” she said. “You don’t want to be left behind because the technology that we’re seeing that’s available to us today is something that is totally transforming how we can practice.”