Medical AI Imaging: Looking Ahead with Cloud-Based Applications

Breakthroughs in Patient Care with Cloud-Powered Medical Imaging

Providers such as Mass General Brigham have deployed Nuance’s AI-powered Precision Imaging Network cloud platform in conjunction with the NVIDIA-backed Medical Open Network for AI, an open-source framework for developing and training AI models in medical imaging.

AI algorithms can analyze medical images much more quickly and accurately than human radiologists, leading to faster and more accurate diagnoses, says Dr. Keith Dreyer, chief data science officer and chief imaging information officer at Mass General Brigham.

“There’s an opportunity for optimization and efficiency gains through the use of AI,” he says. “It can also see things that humans cannot, which is a breakthrough in our domain as well.”

EXPLORE: How medical imaging is pushing clinical IT to the cloud.

AI can flag anomalous findings and provide the radiologist with past examinations that have yielded similar results, says Tom Schultz, senior director of enterprise medical imaging at the Mass General Brigham Data Science Office. “AI can help them find what the outcome in those cases were, which can then inform them on the current case they’re looking at,” he says.

Tools powered by AI and machine learning (ML) can enable interoperability, integrate information across different electronic health record systems, and combine imaging data with genomic and EHR data, says Dr. Rowland Illing, chief medical officer and director of international public sector health at Amazon Web Services. “Natural language processing can be used to identify and present data that’s relevant to the context of the study,” he says.

Scaling Up Medical Imaging with Cloud Technology

Nuance Chief Strategy Officer Peter Durlach points to the advantages cloud technology adds to AI-aided medical imaging programs: graphics processing unit-intensive processes can be scaled up and down again without the enormous expense of on-premises infrastructure. 

“You have this burst capability and an inferencing capacity in the cloud, and IT staff doesn’t need to worry about deploying these constantly changing systems inside the hospital, which they could never keep up with or afford,” he says.

The cloud also allows for faster experimentation and implementation, Illing says, pointing to the example of Aidoc’s always-on, AI-based decision support software, which analyzes CT scans to flag abnormalities and prioritize life-threatening cases.

“They’re regularly looking to release new algorithms targeting new pathologies, and the cloud provides the necessary development velocity for this,” he says.

As AI evolves and machine learning models become more advanced, Durlach explains, the potential for medical imaging to improve early detection capabilities for cancers and other diseases is cause for hope in healthcare.

“If you could do AI for cancer screening at scale, using these models to help pick out patients that are likely to have something suspicious, the effect on care would be astronomical,” he says. “Finding patients that you might have missed or finding disease early is probably the biggest clinical value that’s going to come from this by far.”

UP NEXT: Learn how to improve management in complex healthcare cloud environments.