Tag Archives: Radiology

Signify Premium Insight: Where AI and IT Belongs in Radiology at ECR 2023

With the European Congress of Radiology (ECR) now squarely behind us, it is time to reflect on last week’s Vienna event.

Last year’s meeting, which took place in July, was somewhat subdued, with a limited presence from many Asian vendors thanks to lingering Covid restrictions, while other would-be attendees from Europe were instead on their summer holidays. As such, this year’s conference had a something of a make-or-break atmosphere, with some vendors ready to re-evaluate their commitment to the show, should the 2023 event also fail to excite. Fortunately for the European Society of Radiology, vendors, for the most part, went home satisfied. While the events’ opening day was a little shy of footfall in some quarters, the overall number of visitors and the presence of some new vendors brought positive energy to the exhibition’s halls.

Signify Premium Insight: Enlitic Adopts Standard Approach in GE Partnership

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.

Imaging AI vendor Enlitic recently announced it has signed a deal with GE HealthCare, which will see the two firms integrate each other’s technologies.

The agreement means that GE will embed Enlitic’s platform into GE’s PACS software in order to help improve radiologist workflow and efficiency. GE’s adoption of the platform, dubbed Curie, seeks to improve data standardisation and increase efficiency by reducing the requirement for radiologists to spend time on tedious administrative tasks like correcting broken hanging protocols, for example.

While the partnership marks a significant milestone for Enlitic, it also represents a growing interest in AI tools being utilised for tasks other than image analysis.

The Signify View

As highlighted in numerous past Insights and evidenced by the challenges faced by vendors such as MaxQ AI, it can be difficult for a medical imaging AI vendor to survive from AI image analysis. Many markets are well served with lots of start-ups and scale-ups all targeting some of the most common use cases, while in other areas, market leaders have already emerged, making it difficult for smaller vendors to succeed.

Instead of trying to compete in these already well served markets, some vendors such as Enlitic are focusing on other areas of the medical imaging AI ecosystem. This wasn’t the original specialism of Enlitic, which previously focused on image analysis, with its latest funding round in 2019 taking the vendor’s total funding to $55m, a figure that was at the time very significant. But, the vendor was also working on a tool internally to standardise medical imaging data in order to make it easier for the developer to create algorithms.

This tool filled an uncatered for niche in medical imaging AI, and so was commercialised by Enlitic, becoming the company’s primary focus rather than the image analysis solutions it was originally developing.

GE’s decision to partner with Enlitic somewhat validates this decision, and, through the potential installed base and opportunities that GE offers, can help drive forward the commercialisation of the Curie platform. Moves such as this will be essential if Enlitic, and other vendors which offer AI tools which aren’t focused on image analysis, are to begin to generate significant revenues from such tools.

Where is the Value?

Such an ambition is attainable, but vendors such as Enlitic must highlight the value that their tools can bring. While the value of a product which automatically identifies pathologies on a medical image is self-evident, the utility of tools like Enlitic’s is harder to convey. The vendor must, for instance, illustrate the downstream benefit that can be gained from processes such as fixing hanging protocols and standardising nomenclature, all tasks which radiologists would have to complete manually.

Enlitic estimates that using AI to automate such tasks will save radiologists between 30 and 90 seconds per study, representing a significant improvement to efficiency. This hints at the opportunity GE’s partnership offers and explains the reason AI developers may look to focus on tasks before the reading of a medical image. However, the concept of leveraging AI to support workflow is not new per se, with many diagnostic viewers marketing “AI-enabled” support for workflow optimisation within the reading environment, achieved via self-development and partnerships with white-label applications.

For Enlitic, there is greater opportunity to have a more significant impact by addressing all elements of the reading process before the radiologist actually conducts the read, than by focusing on reducing the reading time for the radiologist. That, after all, is a small component of the whole care pathway. Moreover, it enables the radiologist to focus on high value tasks such as the diagnosis.

Although providers will likely consider the benefits of such tools to help with regards to efficiency, by reducing the need for radiologists to undertake tasks not directly related to the read, providers will also hope to improve patient outcomes. Time sensitive conditions, such as stroke or tension pneumothorax, for example could be treated more quickly, if radiologists can get to diagnosing more quickly rather than spending valuable time performing non-diagnostic tasks such as fixing broken hanging protocols.

Dealing with Data

The advantage of this standardisation becomes more significant in larger healthcare networks in which there is a greater range of disparate sources of data, sometimes with different naming conventions and varying protocols. As individual practitioners or departments send and request studies from across the network, the issues with inconsistencies are exacerbated. As different departments become more closely linked, these inconsistencies will have a more significant impact and the need for greater standardisation increases.

This is also a consideration for providers looking to leverage AI tools across hospital sites. AI tools which can improve standardisation of medical images and their associated metadata will facilitate the use of image analysis algorithms, helping to ensure that orchestration is conducted correctly, and the right scans are routed to the right places.

The question for GE is where this capability is layered into its PACS alongside GE’s own Edison Open AI Orchestrator. Does GE intend to use Curie to carry out the standardisation, before an image enters the AI platform and is then routed to the correct algorithm? Or is it layered in after GE’s own orchestrator? The latter may allow GE more specific control, but for Enlitic, the former would place it in a far stronger position – if its Curie platform was utilised between the PACS and AI platform its oversight of the flow of information and its role in standardisation is far more impactful, while also streamlining and support more effective use of AI tools within the orchestrator platform.

Making a Mark

Such opportunity means that transitioning to this or similar fields could be an attractive opportunity for other vendors. Many AI vendors that are presently associated with image analysis could be facing difficulty in the market and finding it hard to gain commercial traction. For these vendors, transitioning to standardisation and protocolling tools, could be a realistic alternative.

The same is also true for another less well-established vendors that are developing tools to facilitate AI development. Such vendors are focused on assembling datasets and creating toolkits or development “sandboxes” for vendors to utilise in the development of machine learning algorithms. These vendors may, quite naturally, pivot to standardisation platforms given that they have a repository of data as well as a significant amount of expertise. Offering such expertise up via partnerships, or even as the result of acquisitions could provide these typically niche vendors with an opportunity to gain greater commercial traction in a quickly consolidating market.

More broadly, such developments highlight the opportunity that AI offers away from imaging analysis. While that may be the most obvious use case for AI, there are a number of equally significant tasks that AI can be charged with accomplishing with fewer hurdles to commercialisation. Particularly in the near term such solutions will likely provide the bulk of commercial opportunities for vendors. Moreover, partners such as GE will also need to leverage this technology to improve their own portfolio offerings and ensure users have more timely and effective access to new AI-based tools, a clear area of growth opportunity. Further, as PACS and AI platforms blur, data standardisation and reading workflow performance will become a greater aspect of user decision making for purchasing PACS or sustaining existing installed base.

Enlitic is an accomplished vendor in this growing space with its $55m in funding, and the analysis algorithm, which has regulatory approval in Japan, evidencing its potential. However, the partnership with GE HealthCare for its Curie platform offers a potentially lucrative commercial route forward, opening one of the largest installed customer bases of imaging IT users worldwide, ultimately allowing the vendor a big opportunity to realise that potential.

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 Insights: Structured Reporting – Radiology’s Most Overlooked and Undervalued Opportunity

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 and Steve Holloway

Limited radiologist resource has been a long-standing industry concern. While there are an increasing array of products and technologies that are available to offer significant efficiency and workflow improvements to the radiology workflow, some have enjoyed far less commercial traction than their capability warrants. As detailed in Signify Research’s latest report from its Imaging IT Market Intelligence service, structured reporting is one such innovation.

Structured reporting offers radiologists and healthcare providers numerous benefits. By using structured reports, radiology departments can improve the efficiency of their reading and standardisation of report content. The removal of ambiguity in reports supports collaboration across teams and supports improvements in care quality, an ever-greater priority as providers form multi-disciplinary teams to facilitate better patient care and diagnosis.

Beyond improving reports the adoption of structured reporting will also help bestow healthcare providers with a wealth of clean, standardised healthcare data that can be used to improve and develop new diagnostic techniques. As this quantity of data available to providers grows, there will also be opportunities to support the use of precision medicine or further develop and refine AI algorithm development. These benefits will become greater as providers adopt structured reporting at scale, with full enterprise-wide deployments necessary for the maximal advantages to be realised.

Using structured reporting can also have a positive impact on providers’ finances too. Adoption of the technology can, for instance, reduce the risk of missed revenue associated with incomplete reports. More simply, the adoption of structured reporting will also help providers to operate more efficiently across diagnosis and treatment, and as such, save providers’ resources. The potential of federated diagnostic data, enabled with structured reporting, also has substantive commercial value to providers, with the pharmaceuticals and clinical trial sectors especially interested in obtaining high quality, annotated radiology data.

Slow to Bite

Despite these advantages and the potential of the technology, adoption has so far been sluggish. This is, in short, because customers have been slow to demand structured reporting, making it a lower priority for imaging IT vendors to advance past voice recognition or basic reporting templates. There are also some more specific barriers such as the lack of national frameworks for reporting standards, overcoming traditional free-form reporting culture within radiology and the relatively complex integrations required to embed structured reporting across all radiology reporting workflows.

Best of breed vendors and some imaging IT vendors have started developing advanced structured reporting tools using integrated AI and NLP technology to support “no-click” report generation to  overcome these barriers. However, from an imaging IT vendor’s perspective, the customisability and complexity of such tools mean they are relatively demanding of resource to develop, implement and deploy. This is exacerbated by the lack of a standard framework of how reports should be structured, therefore each customer requires the product to be reconfigured. For a midsize vendor, this requires R&D technical investment that it can ill afford to spend, especially when broader focus and resources are increasingly geared towards bidding for larger enterprise imaging or PACS contracts in order to maintain presence in the increasingly competitive and consolidated market. Consequently, the cost of developing a structured reporting add-on cannot be justified in many markets, as the return on that outlay will not be felt for several years with mainstream adoption in mature markets not expected for 5-7 years. Furthermore, structured reporting integration  is rarely a customer “deal-breaker” for major contracts.

The prioritisation of R&D for imaging IT vendors is complex, with many competing investments needed across the core imaging platforms, as outlined in figure 1 below. Vendors are being led by customer needs, which at present are centred around the transition to cloud, enterprise imaging strategies or the integration of AI, all of which are deemed to offer a greater return.

Figure 1: Structured reporting alongside other R&D areas of focus for imaging IT vendors, with our view of when, based on mainstream adoption, renewal cycles and product readiness, vendors can expect ROI from their investment.

A Base on Which to Build

The irony of this lack of prioritisation is that structured reporting could contribute significantly to the other, higher priorities that are on vendors’ roadmaps. Structured reporting for example, provides a centralised access point for AI integration directly into the workflow with the key findings being pre-populated within the radiologists’ report. For operational workflow and BI, it helps automate and streamline the workflow with the findings enabling deeper insights across both clinical and operational metrics. The value of structured reporting, is not only in the tool itself but its ability to advance, connect and support the wider enterprise.

The development of structured reporting tools will accelerate as providers begin to recognise the value outside of niche departmental applications. Although structured reporting is readily used in departments such as neurology, surgery and invasive cardiology, these are typically only partial rather than comprehensive solutions closely aligned to procedures as opposed to broader diagnosis. The wider adoption of enterprise imaging strategies at healthcare providers will help heighten awareness of the benefits of structured reporting.

Locally Led

There will also be regional differences in the rate of adoption of structured reporting. Some regions such as the UK and DACH are relatively advanced in terms of uptake. This has been catalysed by technical considerations, such as the availability of structured reporting tools and integration into best-of-breed RIS  workflow components, as well as expanding capability to digest AI and AV findings.

Adoption also varies by provider type. Although some hospitals with stretched budgets will be keen to take advantage of the solutions, particularly on the back of the economic disruption caused by COVID-19, most customers so far have been large academic hospitals. This is primarily because of additional research capabilities that structured reporting will facilitate, with the vast amounts of data it brings offering opportunities in everything from clinical practice to AI development.

Other types of providers, such as outpatient centres have been more reluctant, with the tools, at present, offering them less upside and limited use cases demonstrating the efficiency these products create. To date there has also been little evidence to suggest public or private payer-driven markets will adopt more quickly; while the population health benefits of structured reporting could support public health initiatives, most publicly driven imaging IT deals are focused more heavily on cost and reducing complexity, threatening the potential for widespread structured reporting adoption without widespread case studies outlining the ROI. In the private sector, there is arguably more potential near-term given the care quality benefits and growing commercial value of richer imaging datasets. However, until vendors can provide evidence of the operational, monetary savings or care quality benefits, adoption will be limited to the top end of most markets.

Building Momentum

Adoption will increase, but there is still a long way to go before widespread adoption is common. Vendors should take stock and focus on structured reporting competency as a clear differentiator competitively, even swallowing the near-term cost of development in order to lay a foundation for longer-term ROI and deal success. Case studies of successful deployments will also be crucial to prove the advantages of their solutions and allow them to categorically quantify the impact their tools had on their existent customers. Over time, providers will increasingly stipulate structured reporting capability in their tenders, so those that invest in the technology earlier will benefit as customer demand catches up.

However, it is the broader impact of structured reporting that should be most compelling to vendors and providers alike. As AI adoption gathers pace, structured reporting will be at the intersection of ensuring AI-driven results are embedded into the radiologist workflow without slowing reading. Furthermore, broader research, care quality and population health initiatives will increasingly rely on real-world evidence from the federated output of structured reporting.

Healthcare providers may have limited capacity to undertake substantial changes to radiology reporting workflow and care practices currently as they battle the tail-end of the COVID-19 pandemic and resources are stretched. Before long however, the penny will drop that structured reporting deployment can have substantiative benefits in radiology and across the care continuum. Some vendors are gambling this will happen far into the future. Based on Signify’s recent research, however, we believe the realisation of the importance of structured reporting might occur much more quickly given its foundational impact on the success of next-generation imaging IT. In a market where competitors are clawing for differentiation in a commoditised market, structured reporting is today perhaps the most overlooked differentiator of all.

 

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