By Don Dennison, CIIP, FSIIM

Medical imaging is a dynamic field, evolving with new technology and changes in the way healthcare organizations deliver care.

There are several areas — such as cloud-based hosting of IT systems, applications that use Artificial Intelligence (AI), multimedia reporting, and others — where the impact of technology and industry can be seen.

Healthcare professionals working in imaging IT have also seen changes in their careers, driven not only by new technologies, but from evolving expectations from their organizations.

Cloud
The idea of cloud-based hosting has been around for some time now, but the technical capabilities are rapidly advancing. Within medical imaging IT, there can be several different drivers for an organization to host their imaging IT solution, like their PACS, in the cloud.

Some organizations may have a directive to reduce their data center usage — for example, if the data center is located on land better suited for patient care facilities or is outdated. For others, it may be simply a matter of cost comparison.

For many enterprises, it is difficult to develop a realistic model to reflect their costs in comparison to a cloud-hosted model. While cloud hosting fees cover all kinds of IT infrastructure build and maintenance, as well as facilities costs, many enterprises do not have their current (or projected future) internal costs clearly defined. Collecting these costs can often be challenging, depending on how costs are accounted for; or sensitive, as labor rates reflect employee compensation.

It is not uncommon for different perspectives on the use of cloud within an enterprise. For example, staff directly responsible for IT infrastructure may prefer to keep applications hosted on-premises, as this is their work responsibility (outsourcing may mean a reduction in their group’s staff). Other roles, often more senior, like a VP or the CIO, may approach the decision more strategically, seeking cost savings. Vendors would do well to try to understand the roles of stakeholders and their perspective on cloud before promoting its use — depending on the audience, the information may be welcome, or not.

Finance department staff, who will be interested in both capital and recurring operating expenses, may also have strong opinions on the impact of cloud-hosted solution costs on how budgets are managed, along with the bottom line. Not all enterprises are set up to shift much of their accounting or budget planning and approval processes from one-time capital to ongoing operating expenses.

At the RSNA (Chicago, Illinois) in 2022, several PACS vendors announced new, or promoted existing, cloud-hosted PACS solutions. One term that was so heavily used that it likely lost some differentiation among buyers is “cloud-native”. Almost every vendor uses the term, with little explanation as to the benefit of this (other than the implication that it is better somehow) or how their solution specifically provides this. Vendors should back up these claims with some clear examples of how their solution is cloud-native and how that affects the customer.

Most cloud-hosted PACS solutions are offered as a private cloud, meaning that each cloud-based instance is built and hosted for each specific customer. Some solutions are multi-tenant, with the software developed to run a single logical instance, but use access control and data management features to segregate each customer's access to their data.

There are several opportunities within the medical imaging domain that could be realized in a multi-tenant solution deployment, including secure and managed research data and teaching file sharing across enterprises, among others.

Artificial intelligence (AI)
AI, in its many flavors, continues to capture the attention and imagination of both vendors and buyers. It remains in peak hype mode, as evidenced by the number of AI-focused vendors and the booth signage at RSNA 2022.

Despite of a number of published governance and scoring models in journals, many healthcare organizations are still in the process of defining and applying formal or ad hoc policies and models for assessing and prioritizing where to invest in AI-based applications. Governance and surveillance tools and methods are absent for many enterprises, while Radiologists in different subspecialties continue to advocate for the purchase of an AI-based application for use in their section. Business-case models to justify investment in new AI-based solutions are often immature or incomplete.

At RSNA 2022, some solutions to help develop and/or validate AI-based applications have started to appear. Some enterprises will look to these types of tools to test and monitor the performance of AI-based solutions before and during clinical use.

The types of suppliers of AI-based applications are also varied. From startups, aggregators, incumbent imaging IT vendors like PACS and reporting solution providers, even to imaging center chain and reading group organizations — all are offering AI-based applications (or pay-per-use services). Buyers seeking to minimize the number of partnerships they need to manage are overwhelmed by the amount of choice available.

Buyers looking to evaluate and select AI-based solutions often encounter a lack of both maturity and consensus of requirements and process. Procurement department staff may not fully understand the complexity of some AI-based solution options, or what terms and conditions to negotiate for during contracting.

Where some Radiologists are eager to use AI-based solutions to perform common, repetitive, and time-consuming tasks, with the hope that they will be given more time to focus on diagnostic reasoning, collaboration, education, research, and other activities they value, it is likely that the healthcare system will look to fill any reading capacity created with more exam volumes. AI may not be the cure to physician burnout that some Rads hope it is.

Interactive multimedia reporting (IMR)
Another topic that is once again trending is that of IMR (called Multimedia Enhanced Radiology Report or MERR, among other names, in prior incarnations).

While IMR has its advocates within some Radiology organizations that promote its added value to the clinicians (and perhaps patients) that receive the reports — because of the structure and formatting of data, inclusion of graphics and images, and even potentially, interactivity — widespread adoption has been elusive. Others are eager to exploit the benefits of structured, semantically labelled discrete data results for population-level surveillance and research.

There have historically been several fundamental technical barriers to effective IMR creation across applications. While creating IMR reports is much simpler within a single application, like a PACS that includes a full reporting module with voice-recognition, the majority of enterprises view images in one application, view patient chart information in another application, and create reports in a third application, all integrated through APIs on the reading workstation desktop.

Emerging standards like FHIRcast provide new opportunities, using modern web-like technical methods, that allow discrete structured data elements — such as a measurement, table, URL, or thumbnail image — to be passed from one application to another. Development of the APIs necessary to use these methods among PACS and reporting solutions is ongoing, but given the varied data and semantic context of each discrete data element, no one should expect this to be a plug-and-play integration in the near term. Early adopters may set some de facto integration and data element definition standards for others.

Integrating the Healthcare Enterprise (IHE) has recently published a new integration profile aimed at addressing IMR (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9485305/).

Yet, even with new technical method and integration guidance, challenges to adoption remain. Results generated using IMR generate no additional reimbursement, but many Radiologists have the perception that IMR reports take longer to create (more work for less money). Not all clinicians, and certainly not patients, fully understand the benefits, and are not generating strong demand. Not all procedures, like simple ankle X-rays, benefit from the effort put into IMR, so only a subset of procedures (like complex CT exams) will likely see significant benefits. Many EMR integration teams are more comfortable with the exchange of traditional HL7 ORU results, which provide none of the benefits of IMR, but are well-understood and have proven interoperable between systems. Finally, the display of IMR reports is more complex than traditional unformatted results, limiting the systems that can present the data consistently, safely, and effectively.

RIS
In early 2020 and again in 2022, I wrote articles for HealthCare Business News on the two prevailing types of RIS: stand-alone and as part of the EMR. Not much has changed in 2023. Large enterprises are commonly adopting the RIS module provided by their EMR vendor, which provides some system efficiencies and reduced interfacing. Reading groups, imaging centers, and smaller health systems adopt stand-alone RIS, or ones available from their PACS vendor, and establish interfaces with one or more HIS/EMR systems.

The number of mature, stand-alone RIS is limited in many markets, especially the U.S. Many of the innovations coming from this class of RIS are focused on functions for improving the patient experience, including through the use of text-messaging notifications and self-serve patient portals.

Healthcare provider operations
Over the past several years, diagnostic imaging departments have changed how they operate. Many changes are driven by policies or simply habits established as part of a consolidation of departmental IT staff to a central enterprise team and the implementation of an enterprisewide EMR system, which leads to a focus on performance measures, analytics, and adoption of workflow best practices that cross departmental lines.

One practice observed among imaging IT professionals in recent years is the need to estimate costs and expected savings for individual projects prior to approval and initiation. Likely driven by practices established by EMR teams and vendors, many imaging IT staff either lack the internal labor costing data or (sometimes) the financial analysis skills to prepare these projections, leading to too many assumptions.

Tying into the above, as these projections are developed, the imaging IT staff often turn to imaging IT vendors to provide historical effort estimates for healthcare provider IT staff from past projects. Often, the imaging IT vendors do not have access to this aggregated, statistical data for the provider staff, only their own staff. Healthcare provider staff sometimes think that imaging IT vendors collect this type of data, when it is not normally provided to them by their healthcare provider enterprise customers.

Related to the above, healthcare provider staff are increasingly asked to provide projections on several imaging IT project aspects prior to the project plan being developed along with the system’s vendor.

For example, before an imaging IT system implementation project kickoff, contract signing, or even budget approval, some healthcare provider enterprises want projections for when specific project expenses will be realized (typically down to a given quarter of the fiscal year). They may also want to know in which month each role required to perform work for the project will be asked to invest their time, along with an estimate of how much of their time will be needed.

Often the vendor is ready to work through these details once they have a contract and the project is kicked off, but given the expectations of such enterprises, this needs to be completed much earlier. Vendors that can provide these insights during the evaluation/sales process, complete with all of the necessary stated assumptions, will provide increased value to their potential customers.

The future
There is little indication that the trend toward consolidation among healthcare provider and vendor organizations will cease. Larger orgs will acquire smaller orgs to fill gaps in their strategies. New technologies, like AI and IMR, will continue to receive attention, but will likely suffer some setbacks before achieving widespread adoption. And healthcare provider organizations will continue to look to align their staff reporting structure, as well as planning and accounting practices, to establish more enterprise-wide standards of operating, performance measures, and workflows. The common theme is efficiency — getting more out of what is invested.

About the author: Don Dennison, CIIP, FSIIM, is a consultant who has worked in the medical imaging informatics industry for over 15 years. A fellow of the College of the Society of Imaging Informatics in Medicine (SIIM), he is a frequent speaker and panelist at SIIM, RSNA, and other conferences on topics ranging from medical imaging record interoperability, integration of imaging data within the EMR, multifacility integration, and others.