Once an isolated subspecialty of radiology, imaging informatics has transformed in the last few years into an enterprise asset that encompasses pathology, cardiology, dermatology, ophthalmology, critical care and more. This evolution calls for new forms of coordination and new strategies to succeed.
“How do we make sure we have the right players around the table that bring in the right perspective around fully capitalizing on this enterprise asset,” Dr. Rasu Shrestha, executive vice president and chief strategy and transformation officer for Atrium Health, told HCB News. “The conversation has to always be focused on what’s the benefit, overall, for the entire enterprise versus what are the specific challenges for one department or division.”
For this to happen, each specialty involved must understand one another’s needs and how together they can forge an imaging informatics dynamic that in the long-term translates to more efficient patient care.
Creating better workflows
The focus of imaging informatics is to create better imaging workflows by improving how medical images are collected, assessed and exchanged. The overall goal is to form ones that ensure more timely and cost-efficient care that results in more optimal outcomes.
"It's going to build workflows for people. As imaging informaticists, we need to be able to collaborate with the specialists who do the work, and understand the workflow that they want," said Dr. Alexander Towbin, associate chief in the department of radiology and the Neil D. Johnson Chair of Radiology Informatics at Cincinnati Children's Hospital. "Just like we wouldn't expect a pathologist to tell a radiologist how to do their work, a radiologist shouldn't tell another specialist how to do their work."
Driving this collaboration is the shift toward enterprise imaging, with providers turning in their traditional PACS systems for VNAs that allow multiple specialties to access, store, manage and share images in one place. “Things like point-of-care ultrasound have spread throughout our organizations. Imaging informatics creates workflows that enable us to bring those point-of-care ultrasound images, as well as photographs and other nontraditional imaging types, into our archives,” said Towbin.
The other driver is the development of AI algorithms for reading scans, triaging cases and automating other tasks to help physicians manage heavy workloads and use their time more efficiently. With imaging informatics, clinicians can create metadata standards that label different types of images with information that makes the data contained in them shareable, understood and accessible among different specialties for AI development and research. This metadata information can be incorporated into algorithms such as radiomics and radiogenomics that can be used to more accurately diagnose conditions, with clinicians then able to create personalized treatment plans.
“What type of scanner, what sequences were used, will feed into algorithms. All of that comes from the metadata that’s stored in an imaging file,” said Dr. Cheryl Petersilge, founder and CEO of Vidagos, an enterprise imaging consulting firm. “As we move beyond DICOM to things like digital photos, we have to ensure those photos are also labeled with the same metadata so they can be just as valuable in AI.”
Working as a team
The first question Petersilge asks people when they ask about setting up an enterprise imaging system is what is the “driving force” behind their decision. This gives a clear indication of not only what they are seeking to gain but also what their imaging informatics plan will be.
“The first thing they have to know is what is the current state of their organization,” said Petersilge. “They need to do a deep dive. An organization may think it has a good knowledge of its imaging activities. But when you start talking to people, lots of things come out of the woodwork.”
This beckons the need for collaborations with both medical and non-medical professionals, all of whom contribute to a successful imaging informatics strategy. The COVID-19 pandemic illustrated this over the last year by forcing clinicians to work from home. “We had to make sure we had workflows that could work from home, that workstations and monitors were there,” said Towbin. “Having ways to do that really fell on IT staff. People with good IT staff saw value in that, and people with limited IT staff saw where that caused them problems.”
Such workflows must accommodate for hospitals of any size in any scenario, he says, as “you have hospitals that send us tons of images a day, while others only one image ever,” and that having a radiologist or other professional who specializes in imaging informatics “is critical because that person can then drive workflow efficiency.”
Imaging informatics plans should also keep in mind limitations that various hospitals face. For instance, in nearly two-thirds of the world, a large segment of radiology facilities still relies on analog equipment and does not have access to digital imaging. In addition, the installation of a VNA is oftentimes for either the radiology or cardiology department but not at an enterprise level, where important figures such as the chief information officer can learn about and understand its advantages for operations.
Declining growth in reimbursement has also made hospitals and health systems in various countries financially stressed. As a result, some may lack the necessary infrastructure or resources to set up and maintain an imaging informatics system. While mergers and acquisitions can improve margins and unite operating systems and resources, they can also give rise to multiple different imaging systems that are rarely interoperable.
“Although many suppliers claim their solutions to be vendor agnostic, interoperability issues persist even today. Migration from one vendor’s solution to another presents significant challenges,” said Dr. Suresh Kuppuswamy, industry principal at Frost & Sullivan.
A more patient-centric world
With lots of different players moving into enterprise imaging, imaging sharing requires coordination among them. The end goal is to create a streamlined workflow for all. One that ensures patients receive timely and affordable care that guarantees the most optimal outcomes with superior experiences. At the same time, these improved and automated workflows give clinicians more time to spend with their patients directly — and for radiologists, clinicians and even patients to collaborate more intently and more effectively.
“With imaging informatics leading the charge, we’re moving from a reality that we’ve been stuck in, in the last couple of decades, said Shrestha. “This is a transformational change for our industry, where we’re moving from an image-centric practice to more of a patient-centric approach to care."
The incorporation of metadata standards is expected to help further develop many different types of AI algorithms, including those for deep learning and image reconstruction and manipulation, as well as features such as autopopulation of measurements and other technologies that help eliminate errors of human data entry.
"AI in imaging informatics can have the potential to substantially change a radiologist's day to day work and a lot of tasks can be automated," said Petersilge. "Eventually we may get to the point where AI can identify a normal X-ray or CT with sufficient accuracy that we feel comfortable not having a human eye laid on that study. Eliminating these tasks will free the radiologists up to engage in higher levels of thinking and become more integrated into the clinical care team.”
Technologies like the cloud will also play a role in the imaging informatics game and help practices in developing countries overcome the limitations they face. “In developed nations, migration from a well-established, secure on-premises server to the cloud could raise questions about data loss, privacy, and security. But in less developed countries, this could be an opportunity to leapfrog, as they do not need to deal with the issues of legacy servers, communication systems and human expertise,” said Kuppuswamy.
Towbin says that by creating one source for images, imaging informatics forms a “holistic” electronic health record from an imaging perspective, that enables various specialties to access and use each other’s images to better assess patients and form more precise and personalized care plans. “Just like the electronic medical record did for all the text data, the enterprise imaging archive will do that for all the imaging data.”