HPE GreenLake Scores AI-As-A-Service Pact With AI X-ray Imaging Pioneer Carestream

‘HPE is a known leader in as a service, but their solutions were also more vendor agnostic from an applications stack perspective,’ says Carestream Health CTO Dharmendu Damany. ‘Also HPE had solutions that supported seamless deployment and management of the devices at the micro-edge, edge, private cloud and public cloud, so the scale was there.’

Carestream Health, a pioneer of AI-based X-ray imaging systems, has inked one of the first Hewlett Packard Enterprise GreenLake AI-as-a-service pacts.

The three-year, multimillion dollar GreenLake on-premise cloud services deal is a “game changer” that opens the door to a new era of increased adoption of AI-based systems in healthcare, said Dharmendu Damany, chief technology officer at Rochester, New York-headquartered Carestream.

“We have seen AI being so pervasive and ubiquitous in many different industries, healthcare is poised for increased adoption with the demands being placed on the healthcare system,” said Damany. “We went through that last year with huge demand (for our mobile AI-based software systems). I think the adoption curve of AI in healthcare is going to grow very rapidly.”

The HPE GreenLake pact comes with Carestream AI-based mobile systems playing a critical role in preventing coronavirus infections of healthcare professionals during the global pandemic.

There is “huge potential” for AI to improve patient care, said Damany. “The adoption curve is just picking up,” he said. “With all the pressures that we have on our healthcare system whether it is as a technologist or a radiologist, every second and minute we save, a patient going through triage, is a big difference.”

Carestream, which has been supplying AI based imaging software for a decade, has already seen a dramatic impact on software development as a result of the partnership with HPE, said Damany, The training for a smart noise cancellation X-ray imaging technology feature decreased from 60 hours to 16 hours, he said.

“The time and effort that HPE spends in understanding our business model and understanding our future and growth vectors and the experience that HPE brings from the solutions they provided in this space was very important for us to learn from,” said Damany. “We have already started to see a reduction in cycle time for AI model creation and optimization.”

HPE helped Carestream drive a “more comprehensive AI platform and architecture” to support its AI based systems, said Damany. “HPE is a known leader in as a service, but their solutions were also more vendor agnostic from an applications stack perspective,” he said. “Also HPE had solutions that supported seamless deployment and management of the devices at the micro-edge, edge, private cloud and public cloud, so the scale was there…Our customers want that flexibility. The key thing for us is to be flexible enough to provide our customers with what ever solution fits them. HPE was one of the top ones that was able to cover the entire spectrum.”

Critical to the as-a-service pact is the ability for Carestream- which has 100,000 systems in use across the globe- to scale its AI-based model as the business grows, said Damany. “That scalability was what we were looking for so we don’t get too far ahead in expenses as the business continues to grow,” he said.

Damany singled out HPE’s support for “good machine learning practices” as laid out by the Food and Drug Administration’s Artificial Intelligence/Machine Learning Based Software as a Medical Device Action Plan released earlier this year. “Complying with that gives us the ability to bring these AI features (to our customers) and deploy and manage them (with the HPE infrastructure),” he said.

Carestream is hailed as one of the market leaders in the mobile-based X-ray system solutions market with its ImageView software and Eclipse AI technology. Carestream applies AI-based technology solutions in four key areas: image intelligence, workflow intelligence, analytics intelligence and decision support systems.

Among the areas Carestream is using AI is with “smart patient positioning” for X- Rays and AI based X-ray “smart noise cancellation to cancel out “noise” to improve image quality.

The “solid” HPE GreenLake infrastructure “foundation” is critical in bringing out more AI based features for Carestream customers, said Damany, “There is a learning and adoption curve we are pushing,” he said.

Carestream’s AI-based X-ray platforms support both full X-ray room solutions and mobile technology. The mobile systems were critical to preventing infections of medical professionals during the coronavirus pandemic. “Last year we were fortunate to provide a number of our mobile DRX-Revolution systems to many of our customers,” said Damany. “We ramped up 2x to 3x times our normal output.”

Those Carestream Mobile DRX-Revolution systems provided critical chest X-rays during the global pandemic. “It is more difficult to take a patient to an X-ray room as opposed to having the mobile system come to the patient bedside,” said Damany.

In the future, Carestream is interested in breaking beyond the device centric AI-based X-ray imaging model into an as a service model in the future, said Damany.

Mike Strohl, CEO of Entisys360, a California-based HPE Platinum partner that is seeing dramatic everything as a service sales growth, said he was heartened to see HPE expanding its GreenLake footprint into AI as a service.

“As an early adopter of GreenLake we are excited to see HPE expanding the GreenLake offering into AI, Citrix, SAP and other use cases,” he said. “HPE continues to create a standard for utilizing GreenLake for many use cases. When you tie GreenLake into application optimized scenarios it really brings bigger depth to what HPE is doing in as a service. A real differentiator in the everything as-a-service market is channel commitment. That is an HPE hallmark.”

Arwa Kaddoura, vice president worldwide sales and go to market leader GreenLake for HPE, said the GreenLake model allows companies like Carestream to focus their “best and brightest” on building their core products and capabilities.

“The (GreenLake) ML Ops solution we came up with just a few months ago is really a step in helping our customers develop the types of AI algorithms that then power their business,” she said. “We are providing the open platform so that Dharmendu and his team can then build the business outcomes they need for their customers and then not to have to worry about extending all the way down into that infrastructure layer.”

GreenLake effectively delivers the “true cloud promise” by eliminating infrastructure complexity with a “fully managed, as a service experience” so that customers can build their own “software stack” on top of that HPE infrastructure, Kaddoura said. “That is what is really huge here,” she said.

A critical differentiator in such solutions is HPE advisory and professional services to cater to the specific use case for Carestream, said Kaddoura. That means making sure that the Carestream infrastructure stack was “tailor built from the infrastructure stack all the way up to the integration layer to be highly optimized for their use case,” she said.

The HPE Ezmeral data fabric – the heart of the ML Ops as a service platform- is “really starting to pay off massively” for HPE GreenLake customers, said Kaddoura. “They need the that data fabric layer and the ability to containerize these applications and have that consistent approach,” she said. “We have everything from infrastructure to the platform and managed services covered. We are super excited about continuing to provide these innovations to our customers so that they can go build their own solutions seamlessly.”

Damany, for his part, said he is optimistic about the future AI potential given the Carestream-HPE partnership. “We’ve already seen the benefits of working with HPE and I am looking forward to the journey we are taking together,” he said. “There is a lot we can do.”