GE Healthcare offers radiology departments access to innovative AI tools

GE Healthcare has developed a next- generation, cloud-based Picture Archiving and Communication System (PACS) that aims to help healthcare organizations keep current with rapidly evolving technology, relieve the pressure on capital budgets and free up IT resources. The subscription-based diagnostic imaging solution called Edison True PACS¹ encompasses diagnostic reading, exam workflow, AI Applications, 3D post processing, enterprise visualization and archiving in a single platform.

Currently available in the United States, GE Healthcare is preparing for a Canadian launch in 2022.

What distinguishes Edison True PACS from traditional PACS systems is how it offers radiology departments access to innovative artificial intelligence tools, claims Peter Eggleston, GE Healthcare’s Global Product Marketing director for radiology. “We see the cloud as a way to democratize access to AI because the reality now is that the AI market is fragmented. There are a lot of different vendors focusing on very narrow areas and the cost to an organization to work with all these vendors is very high.”

GE Healthcare simplifies the adoption of AI through what it calls the Edison ecosystem. “We take away a lot of the work of finding, acquiring and standing up the AI,” said Eggleston. “In the cloud, it’s very easy for us to provision new services, so we can greatly shorten the amount of time it takes for an organization to adopt AI. And, it’s only one vendor, GE, that you have to work with because we serve as a distributor of AI tools through resell and distribution agreements we have with third parties.”

Currently, GE Healthcare offers AI algorithms that are available through Edison True PACS, but it’s working with more and more companies that are developing algorithms for a whole suite of AI use cases, noted Matthew Collingridge, General Manager of GE Healthcare’s Digital Solutions Group. They include for example, algorithms being developed to help identify and prioritize likely cases of pneumothorax, stroke and cancer, speed workflow and help enhance diagnostic confidence.

“AI is gaining confidence among users,” said Eggleston. Aside from prioritizing studies that are truly critical, “it’s good at doing things like measuring and quantifying physiological changes compared to previous readings and identifying things that aren’t visually perceptible to humans because the comparative data is so low or it’s so small. “Acceptance of AI is pretty high because radiologists are always concerned that they’re going to miss something,” he added. “They’ll tell us, for example, that they feel like they’ve got a second set of eyes and that somebody is watching their back.”

A study by the American College of Radiology found that clinical adoption of AI by radiologists went from virtually zero to 30 per cent from 2015 to 2020. Other studies cited by GE Healthcare have found that read times using AI applications are 26 per cent² faster for evaluating lung nodules and up to 60 per cent faster for evaluating multiple sclerosis cases³.

Radiologists see the value of AI, “but haven’t had an easy way of using it. That’s what the Edison ecosystem does,” said Eggleston.

“We’re excited about the next 24 months,” commented Collingridge. “We see a number of new technologies emerging that have had early clinical validation and we now need to push into broader access that we think will benefit not only the large institutions, but also many of the rural and regional areas. We’re embarking on an initiative in Quebec, for example, to bring AI access to radiologists serving rural areas.”

GE Healthcare’s Edison True PACS is deployed on the Amazon Web Services in the United States to ensure a high level of security and data protection, so hospital IT departments don’t have to worry about maintaining all of the security and intrusion protocols themselves. It also takes care of software upgrades, freeing IT staff to spend their time on other enterprise challenges.

Neuro Imaging Winter Park, an imaging centre in Florida, is one GE Healthcare customer that saw the value of a cloud-based PACS during their evaluation of the cloud based solution.

“With storage getting so expensive and constantly having to switch out drives, maintain them and make sure the RAID arrays are in good shape, it just makes sense to have it hosted offsite,” said Richard Duemm- ling, Neuro Imaging’s Chief of Business Operations. “If we have a fire or there’s a cyber attack on our network, we are likely in a better position with our data on the cloud-based PACS. It’s also easier and less expensive to deploy a new PACS because we don’t have to buy new server hardware. We don’t have to maintain it, or worry about operating software updates and patches. There are a lot of hours involved in maintaining an onsite server and when you go with a cloud-based solution, IT infrastructure is effectively managed as a Service. If you don’t have a big IT team that’s constantly dedicated to making sure everything is up to date, then you’re at risk.” A cloud-based PACS is also ideal for hospitals forced to leverage outsourced or distributed reading due to increasing workloads and the shrinking radiologist workforce, said Eggleston because “if you’re initiating a relationship with a third party and everything is in the cloud, it’s much easier to connect and scale that infrastructure to another organization.”

Edison True PACS is available through software as a service (SaaS) or on premise in three subscription-based models: Edison True PACS Technologist, Edison True Essentials and Edison True Professional. Hospitals opting for the on premise solution look after their own storage. For the SaaS solution, “we manage the servers and everything that goes along with that such as the third-party software, and since we manage them, we’re also responsible for keeping them updated, so the customer doesn’t have to do that,” said Eggleston. “When you’re on premise, software updates often incur charges or disruption. Usually, you have to find a server to stand it up and test it because you don’t want to just take it and put it into production, so there’s time and cost there. With the SaaS version, you don’t need to do any of that.” The on premise offering “is a toe in the water to help healthcare providers and hospitals move to a more native cloud environment where things can be more seamless and secure, particularly from the threat of cyber risks, and enjoy the advantage of continuous improvement and upgrading of functionality and features,” said Collingridge.

Edison True PACS Technologist accommodates the QA or technologist workflow and is designed for organizations that just acquire the images and outsource the reading. Edison True Essentials includes a diagnostic viewer and the basic requirements for diagnostic reading. Edison True Professional is for organizations that want all the latest technology, including GE Healthcare’s Open AI Orchestrator that can integrate AI tools into the radiologist’s workflow.

Cloud technology, interoperability and AI have all reached a state of maturity that was required for the launch of a cloud-based picture archiving and communication system, said Collingridge.

“Edison True PACS,” he predicted, “lays the groundwork for what we think will be an exciting two or three years in Canada for machine learning.”

1.Edison True PACS Solution consists of the following products: Universal Viewer, Enterprise Archive, Centricity Universal Viewer Zero Footprint, Edison Open AI Orchestrator and 3rd party AI applications

2.Source: Riverain Technologies Medical DeltaView FDA 510(K) Reader Study Results, 2011

3.Source: Lo, S. B., Freedman, M. T., Gillis, L. B., White, C. S., & Mun, S. K. (2018). JOURNAL CLUB: Computer-Aided Detection of Lung Nodules on CT With a Computerized Pulmonary Vessel Suppressed Function. American Journal of Roentgenology, 210(3), 4 80–488. doi: 10.2214/ajr.17.18718