Cloud

NetApp Technology Experts Explain How Medical Imaging Is Pushing Clinical IT to the Cloud

Current and projected medical imaging workloads are increasingly pointing healthcare organizations to the cloud. Those looking to make the transition should consider how cloud services have been built to ensure solutions meet their needs.

Brian Eastwood

Twitter

Brian Eastwood is a freelance writer with more than 15 years of experience covering healthcare IT, healthcare delivery, enterprise IT, consumer technology, IT leadership and higher education.

Medical imaging has often been at the forefront of new technology adoption in the clinical setting. First, there was the push for upgraded networks that could support the transmission of large files. Then, there was the use of artificial intelligence to assist with analysis of breast and lung images.

Today, the size and scale of imaging is testing the limits of on-premises storage servers and networks for today’s hospitals and ambulatory clinics. Coupled with a shift in technology and operational proprieties because of COVID-19, medical imaging is pushing more organizations to consider the cloud for clinical applications.

“The pandemic shined a light on the need to be able to manage resources effectively and efficiently, and to be able to access information from anywhere,” says Kim Garriott, chief innovation officer at NetApp Healthcare. “Imaging is leading the way — again — and it’s being driven by a demand for turnkey managed service offerings.”

DISCOVER: How NetApp simplifies the cloud and supports medical imaging.

Long-Term Medical Imaging Plans Call for a Move to the Cloud

Medical imaging certainly has enough clout to push healthcare to the cloud. Garriott notes that imaging typically accounts for more than 80 percent of an organization’s total volume of clinical content.

As imaging becomes more sophisticated, that number will only increase. A traditional chest X-ray — the most-ordered radiology exam — is a single 15 megabyte image, about the size of a few JPG images. Breast tomosynthesis, a type of 3D mammography, creates files that are about 300MB. The emerging field of digital pathology, which turns microscopic images from glass slides into digital slides viewable on a computer monitor, generates images as large as 3 gigabytes.

Organizations may not yet be adopting digital pathology, but IT leaders need to consider its impact as they develop their three-to-five-year infrastructure strategies. While it’s true that the cost of physical storage is decreasing, especially given the benefits of leveraging object-based storage for unstructured data such as images, organizations will likely find it difficult to scale onsite.

“The size of images that will be produced by more advanced imaging technologies is going to dwarf what we see happening in all other aspects of medical imaging,” Garriott says. “Organizations are also worried about performance, because they want to be able to pull up images whenever and wherever someone is seeing a patient.”

“The challenges that imaging presents makes it harder to keep up with procurement. That’s where the value and attractiveness of the cloud comes into play,” says Tony Turner, strategic partner manager for NetApp Healthcare.

Click the banner below to learn more about optimizing your cloud environment.

Comparing Native Cloud to Cloud-Enabled Services in Healthcare

The transition to the cloud for imaging is gaining steam. According to Signify Research, investment in on-premises imaging IT architecture will decrease slightly from 2020 to 2024, with a compound annual growth rate of -0.2 percent. Spending on cloud architecture, on the other hand, will increase at a CAGR of 19.5 percent.

It’s worth noting that all cloud services for enterprise medical imaging are built the same way. Those that are cloud-enabled tend to trace their origins to on-premises imaging archives that have been adapted to the cloud.

Making this work can be a tall order, as it requires building cloud architecture from the ground up. It also represents a significant shift in the approach to research and development. Instead of focusing on building new clinical software features that may generate sales and renewals, a vendor needs to allocate resources to infrastructure development.

EXPLORE: How hybrid cloud unlocks the power of clinical data management.

A cloud-native service, meanwhile, has been purpose-built for a hybrid cloud environment.

“There’s a first-mover advantage and an understanding of how to do business in the cloud,” Garriott says. These services, for example, integrate out of the box with major public cloud services such as Amazon Web Services, Google Cloud and Microsoft Azure, without the need for custom development.

With a cloud-native service, organizations are better positioned to monitor workloads and use patterns, implement data management and governance strategies, and detect potential security threats across their entire enterprise medical imaging environment (rather than one application or archive at a time). This offers organizations the flexibility to use resources as needed, and it relieves medical imaging and IT teams of the need to maintain hardware and software onsite.

An Opportunity for Medical Imaging to Support Research

The transition from on-premises to native-cloud enterprise medical imaging helps healthcare organizations cut infrastructure costs and boost efficiency. It also presents a business opportunity that can advance research and technology development.

Just as medical device manufacturers must receive Food and Drug Administration clearance before bringing a product to market, analytics companies need FDA clearance for AI algorithms. For the growing number of companies that are training AI models for diagnostic and decision support in medical imaging, healthcare providers’ imaging data — once deidentified — is extremely valuable, Garriott says.

Given the security and compliance considerations, providing a vendor with access to information stored on-premises is typically a nonstarter. However, enabling secure access to deidentified data hosted in the cloud may be an attractive option for some organizations, she says.

“The cloud makes it easier for organizations to commercialize their data so that it can aid in the development of AI models, and hosting in the cloud means you don’t need to give third parties firewall access to your internal network,” Garriott says.

Brought to you by: