Tag Archives: SIIM

Signify Premium Insight: The Trends in Imaging IT at SIIM 2022

This Insight is part of your subscription to Signify Premium Insights – Medical Imaging. This content is only available to individuals with an active account for this paid-for service and is the copyright of Signify Research. Content cannot be shared or distributed to non-subscribers or other third parties without express written consent from Signify Research. To view other recent Premium Insights that are part of the service please click here.

Co-written by Amy Thompson

Although the recent annual meeting of SIIM, the Society for Imaging Informatics in Medicine show is much smaller than many of the other events on the medical imaging calendar, its focus on imaging informatics and the emphasis on business-to-business relationships means that it is still an influential event and a good bellwether of the trends in the medical imaging IT market. As such, while there was only a limited number of product announcements, the debate was stimulating. This was particularly true given that some innovative smaller vendors, which risk being crowded out at larger shows, were able to enjoy the limelight.

Continue Don’t Counter

Despite its different composition, however, the trends witnessed at the show were very much a continuation of those seen at other recent events. Such was the case for cloud technologies, for example. Vendors were keen to promote their solutions and marketing boasting of cloud capability was ubiquitous. Despite this promotion, reluctance around cloud technology remains, with providers hesitant to commit themselves completely to a fully cloud-based platform deployment. This reticence is largely a result of providers’ scepticism that cloud solutions can match the performance of on-premise data centres, especially when evaluating performance of core diagnostic tools such as the PACS which plays a critical role in patient care. After all, any slight delay in accessing an image or report, even by a few seconds, for example, is not something a provider is willing to accept when it could impact patient care.

While providers are yet to be fully convinced of the merits of a transition to cloud, vendors also still have some work to do. Although almost all vendors offer some cloud capability, and some vendors can offer some cloud-based enterprise imaging, no vendor is yet ready to offer a complete cloud-based enterprise imaging platform at scale. While interest in cloud continues to grow, these challenges mean that providers will continue to favour hybrid cloud solutions as opposed to fully hosted public cloud. This approach could cause a provider to miss some of the more sophisticated opportunities offered by full cloud deployments, though in the near term retaining some core components, such as AV or the core diagnostic PACS on premise represents a sensible compromise. Further, in many cases it will allow providers to enjoy the benefits of the evolution towards a SaaS-based business model, without risking any detriment to performance that they fear from a full cloud deployment.

Cybersafety in the Cloud

Despite these reservations however, there are some reasons motivating providers to adopt cloud solutions. Among the most prominent of those at SIIM, especially compared to previous years, was cybersecurity. Although cybersecurity has been a concern for a number of years, the growth in the numbers of ransomware attacks has helped mean that cybersecurity has become one of the main catalysts driving the transition to cloud, especially at larger organisations.

Progress is slow, however. This is particularly true as the transition to cloud isn’t occurring in isolation, with the development coinciding with the broader adoption of enterprise imaging strategies. As such, vendors and providers aren’t merely focused on the technical challenges, they are also determining how departments other than radiology, such as cardiology and pathology will sit in the cloud. Digital pathology images, for example, offer unique challenges compared to radiological images, in part due to their sheer size and differing workflow. These are challenges that will be addressed, but at present, despite some vendors’ fanfare, adoption of cloud technology has been fastest at specific customer segments, such as smaller outpatient sites (looking to remove IT administration burden), and academic sites (leveraging cloud for research). The bulk of the market has still not made the jump, and in many cases, won’t do for some time.

Pathologically Minded

Another of the major technological changes that vendors and providers were discussing at the show was digital pathology. There has been more interest in digital pathology since the Covid-19 pandemic, which highlighted the possibilities of digitisation for a number of clinical areas and hospital departments. In the US this is being reflected by the easing of FDA rules on primary diagnostic use of digital pathology, which is offering the opportunity for imaging IT vendors to enter the market and create interoperability between two departments and their specialists.

While there are some imaging IT vendors that do already offer capability in digital pathology, the fact the topic was being very visibly discussed at a traditionally radiology-focused show, shows that it is becoming a more mainstream opportunity. The fruits of vendor investment in this sector will take time to be realised however. Significant questions need to be answered such as whether a specialist viewer is required, or if it is preferable to be able to view pathology images in the radiology viewer, and whether radiologists need access to slides or if access to the report is sufficient, must all be determined. Notably, some imaging IT vendors are circumventing this learning curve by partnering up with digital pathology specialists, with traction already evident in Europe. The same is expected in North America in the near-term future.

 Exchange of Information

The technology to underpin image exchange is very attainable for a large imaging IT vendor, but whether it is worth attaining is still unknown, with vendors keen to understand the market opportunity for such solutions. This is in part because, at least in private markets, it represents something of a double-edged sword, with providers interested in the ability to share images between their own sites (intra-network exchange), and the ability to use the technology as a differentiator to entice patients in. However, making images easier to share also means patients will find it easier to leave one provider network (inter-network exchange), a loss of business that hospitals will not wish to facilitate.

Despite these questions, from a market perspective the additional interest from providers does make sense. Spurred on by the need for digitisation during the Covid-19 pandemic, image exchange is becoming a dedicated budget item for providers, giving imaging IT vendors more of a reason to target the segment for upsell opportunities. However, particularly in comparison to the core imaging IT platform, it is still a very small market. Moreover, it is also likely to be a market that will increasingly become integrated into enterprise imaging platform competency as providers continue to consolidate contracts and streamline technology.

Putting Words to Good Use

Such discussions surrounding upcoming technological shifts will have been useful for vendors. The show highlighted that vendors must be able to demonstrate considered, deliverable strategies for several technologies instrumental in the development of the imaging IT market. Providers might not be looking to make the switch to new technologies and systems today, but they need to know that when it comes to cloud capability, AI integration platforms and enterprise imaging and interoperability tools, their chosen vendor must be ready when they are.

About Signify Premium Insights

This Insight is part of your subscription to Signify Premium Insights – Medical Imaging. This content is only available to individuals with an active account for this paid-for service and is the copyright of Signify Research. Content cannot be shared or distributed to non-subscribers or other third parties without express written consent from Signify Research. To view other recent Premium Insights that are part of the service please click here

Signify Premium Insight: The Trends in Medical Imaging AI at SIIM 2022

This Insight is part of your subscription to Signify Premium Insights – Medical Imaging. This content is only available to individuals with an active account for this paid-for service and is the copyright of Signify Research. Content cannot be shared or distributed to non-subscribers or other third parties without express written consent from Signify Research. To view other recent Premium Insights that are part of the service please click here.

Co-written by Dr. Sanjay Parekh.

Although not one of the biggest medical imaging shows on the calendar, the annual meeting of the Society for Imaging Informatics in Medicine (SIIM) is still a significant and influential event thanks to its emphasis on informatics and focus on business-to-business relationships.

At this year’s event, which took place in Kissimmee, Florida earlier this month, there was a relatively strong turnout, especially considering the lingering effects of the coronavirus pandemic in some parts of the world and the decision to concurrently host the conference virtually. However, one area that was less well represented than might otherwise have been expected was AI.

A Conspicuous Absence

This absence, is, on the face of it, surprising, given that the show should have represented a good opportunity for AI developers to enter in conversations with informatic partners and increase their exposure; an attractive proposition for those smaller vendors that may be overlooked at some of the bigger shows such as RSNA. Particularly surprising given the trends in funding for medical imaging AI vendors was the acute absence of Asian vendors, arguably representing something of a missed opportunity.

There were however good reasons of this absence, particularly for some of the larger scale-ups. Chief among them is the fact that these, more established AI vendors are already well recognised in radiology. These vendors are looking to build on this broader awareness by more nuanced targeting of potential customers across other hospital departments, and are looking to achieve this by focusing their efforts on other more specific clinical shows.

This is particularly true for those vendors who have extended their product’s capability into the workflow beyond the imaging itself. These vendors, which increasingly address entire workflows for some specific conditions, are looking to tap into larger operational budgets and secure bigger, hospital-wide contracts. As such, they are looking to promote their products directly to professionals working on those clinical areas. Thus, specialist conferences on neurology, cardiology, or oncology appear to be a key part of building brand and providing education to potential new users.

AI’s Rising Tide

Those vendors that did attend the show, however, will no doubt have got a sense of the growing importance of AI to both vendors and providers alike. One such area of focus was AI platforms. Despite the growing capability of AI tools, it is still sometimes a challenge for vendors to facilitate their adoption by providers. Vendors are increasingly looking to AI platforms to meet the needs of deploying and orchestrating multiple AI tools. SIIM epitomised this shift, with several vendors demonstrating their platforms, and several more having discreet conversations about their plans to launch a platform or partner with an existing platform, for example. Other vendors, however, have pursued an approach for tighter integrations of AI tools within other components of the imaging workflow, albeit at the expense of a dedicated, branded deployment engine.

This momentum towards delivering AI platforms was even displayed by some vendors which, at present, only have a relatively minor presence in the imaging IT markets (e.g., Bayer, DeepC). For such vendors the development of their own platform solution allows them to target a segment that is set to be very significant in the coming years, as well as capitalising on its growth. Of equal significance is the fact that no vendor wants to be left behind. These vendors may not expect to derive significant revenues from them soon, but by releasing their own platforms, they at least can offer the capability and look to leverage it to drive revenues in future.

At SIIM there were also many conversations among vendors which were still seeking answers to some very fundamental questions. For example, large imaging IT vendors must decide whether they will enable the use of third-party platforms on their imaging IT systems or will they instead try to lock their customers into their own ecosystem of developer partners, or even consider the possibility of combining both approaches.

The Customer Gets what the Customer Wants

As is most often the case, customers with the necessary money, and a strong enough will are likely to get what they want, even if not officially supported by vendors. Large imaging vendors prefer customers using their native AI platforms rather than third-party AI platforms, even if they remain nascent in development. However, although they may appear reluctant, these vendors may also ‘permit’ the use of third-party AI platforms on its imaging IT systems. There are some advantages to this. It does, after all allow them to cater to a wider range of use cases, including offering several best-of-breed options for niche use cases. Another advantage for large imaging vendors is that they can give customers the opportunity to use these niche products, without having to undertake costly, or time-consuming integrations themselves. These vendors can focus their efforts on ensuring the tools that are most useful to the greatest numbers of their customers work seamlessly, and leave other, niche capability to third parties.

Alternatively, some imaging IT vendors will integrate specific third-party applications on a customised basis, even if these applications fall outside of the vendor’s own orchestration framework. This work can be costly, with either the application developer (less common) or imaging IT vendor (more common) footing the bill. For vendors, sometimes this is a worthwhile investment, as the API developed can be rolled out across other customer sites in its installed base. Alternatively, and increasingly, customers will foot the bill.

Specialist third-party platforms do not currently represent a threat to a large imaging IT vendor, but as they grow in competency and breadth, competition will intensify. Today, they may take a small percentage of a large vendor’s AI spend, but their impact will be minor. A bigger risk for imaging IT is that third-party platforms, or the performance of  third-party AI vendors hosted, will be poor, a factor that could reflect badly on the larger imaging IT platforms. Ultimately, it’s likely that large imaging IT vendors will develop and integrate a select few algorithms directly into their software near-term, but allow third-party platforms to operate, unsupported in the background to meet customers’ more niche requirements, at least until market consolidation intensifies.

A Matter of Choice

Regardless of how customers can select AI tools, vendors still need to work out how best they can get the fruits of AI into imaging workflows. This hurdle not only entails giving radiologists the autonomy and the control to choose which algorithms should be applied to images, but also how they access and interact with the results. Making this aspect of the AI imaging workflow even more challenging is the lack of standardisation in algorithm outputs.

Commonly algorithms return results in a DICOM secondary capture format, which as a static image is widely usable but lacks capability compared to more sophisticatedly integrated solutions. Similarly with some structured reporting tools. They may interact directly with the workflow, but simply outputting a PDF into the report leaves the radiologist somewhat on the sidelines. Instead, vendors are converging around a principle which sees the AI engine run in the background, but the radiologist can actively interact with the results, overriding them and disregarding some findings, for example, or making amendments or comments before committing an image to the PACS.

While vendors see this as a noble aim, there is no consensus on how to reach it. These are the sorts of issues that will have animated off-the-record discussions at SIIM. No vendor will draw a line in the sand and lead the vanguard towards greater standardisation, but many vendors are considering it and trying to elicit soft support from their peers and partners.

Questions Yet to Answer

It’s a similar story for other key discussions which are yet to be settled. Whether the nuance of reimbursement, and how vendors can best encourage the issuance and uptake of new and existing AI reimbursement codes, or the strategy for assessing post-deployment performance. These are significant topics for vendors, but the fact they are now being discussed in earnest highlights the traction that AI has already shown.

The vendors present at SIIM, as well as those that were conspicuous by their absence have already achieved a lot. They have highlighted the need for AI, and at shows like RSNA and ECR in particular, have convinced the radiology market as to the opportunity the technology presents.

Now, in smaller and more focused conversations at smaller and more focused shows, AI developers and large specialist vendors can get started on the detail.

About Signify Premium Insights

This Insight is part of your subscription to Signify Premium Insights – Medical Imaging. This content is only available to individuals with an active account for this paid-for service and is the copyright of Signify Research. Content cannot be shared or distributed to non-subscribers or other third parties without express written consent from Signify Research. To view other recent Premium Insights that are part of the service please click here