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
Aidoc has continued to forge partnerships to expand its coverage of radiological specialities, and last week added ScreenPoint Medical to its roster. The collaboration will see the mammography AI vendor’s toolset added into Aidoc’s platform, allowing breast imaging specialists to access its capability within the existing workflow.
Aidoc has been increasingly utilising partnerships to complement its own AI tools and expand the capability it offers its customers. The addition of ScreenPoint to Aidoc’s stable will mean that the Israeli vendor’s customers will now be able to utilise ScreenPoint’s Transpara breast decision support solution, alongside solutions from its other partners, which include Imbio, icometrix, and Riverain Technologies for pulmonary embolism, stroke care, and lung cancer respectively.
The Signify View
AI will, over the coming years, transform medical imaging. Tools to supplement radiologists in their diagnoses, streamline radiology workflows and positively impact clinical outcomes will become ubiquitous. However, in order to reap any of these benefits, these tools must first be incorporated into existing radiology workflows.
One of the ways this challenge is being addressed is through the use of AI platforms. Such platforms facilitate the deployment and integration of AI tools into the PACS, making it easier for radiologists to use and manage multiple AI tools within the imaging workflow. Vendors of third-party AI marketplaces, which facilitate the selection and purchasing of algorithms, such as Terarecon and Blackford Analysis were among the first companies to offer these platforms as a way for providers to efficiently deploy the algorithms purchased through their marketplaces.
As AI developers expanded their portfolios and began offering a greater range of algorithms, challenges around integration remained. Some vendors, such as Aidoc, have also chosen to go down this route (and develop native platform capabilities), enabling multiple AI solutions to be deployed through a single platform rather than having to be brought into radiologists’ workflows individually. Initially these platforms from algorithm developers started as a way to deploy their own algorithms into providers’ workflows, however, the market has evolved, and providers are seeking more complete solutions than Aidoc alone can offer. This has prompted Aidoc to form partnerships to enable it to quickly expand beyond its original brief of detection and triage for certain CT exams. It now offers several complete end-to-end solutions for a number of high-value clinical use cases, including pulmonary embolism care, and stroke care.
Functional Focus
The flurry of recent partnerships made by Aidoc, as well as its continued development of native algorithms has been integral in its ability to offer these end-to-end solutions. Its partnership with Imbio in December 2020 provided quantification capabilities as part of its pulmonary embolism care solution, and its partnership with icometrix in June added perfusion capabilities to its stroke care coordination solution. In September, Aidoc partnered with Subtle Medical, to bring image enhancement capabilities for image acquisition. Last month, the vendor partnered with Riverain Technologies to take advantage of its chest AI solutions (which improve efficiency and improve nodule detection rates), while most recently, the collaboration with ScreenPoint Medical aids its breast imaging play.
These partnerships have enabled Aidoc to bundle tools that address specific clinical use cases. One challenge in this approach, however, will be addressing gaps beyond its existing capabilities. Unlike some vendors that promise to offer ‘comprehensive’ solutions, which ‘solve’ a particular body area/modality combination, Aidoc’s focus on specific use cases mean that for anything outside of that specific scenario, radiologists will have to still read the image themselves, or seek out other algorithms to supplement the offerings available on Aidoc’s platform. While this situation will improve over time as Aidoc adds more partners and develops more native algorithms, it is, at present, a limitation.
The impact of this limitation will vary by provider. One key metric for radiology reading groups, for example, is their efficiency and turnaround times. This is particularly true for emergency cases and night reads. As such, despite the gaps in its capability, if Aidoc can help improve turnaround times for these reading groups and thereby give them a competitive advantage, its tools will still be sought after. This competitive advantage will also enable radiology groups to compete with their peers (other radiology groups) more aggressively as well as with hospitals themselves, thereby generating further business.
This perhaps, is one of the reasons Aidoc felt another of its partnerships made sense, that with one of the leading radiology groups in the US, Radiology Partners. At the time of the partnership (April 2021), it was touted as the largest clinical deployment of AI in healthcare, affecting as many as 40 million annual exams.
Standing Out
Aidoc’s focus on use cases is one factor which differentiates the vendor from its closest competitors, such as Arterys, which instead bundles algorithms by body area and modality. Arterys arrived at this approach after setting out to compete with other marketplaces. The vendor’s initial focus was establishing a wealth of third-party AI developers to host in its marketplace and on its platform. The bundled offerings came later, as it sought to add greater value for its customers. This is in contrast to Aidoc, which has prioritised providing solutions which address an entire clinical workflow for a given use case; a method which potentially harbours more commercial potential given that a provider may now be more willing to pay for a fully-fledged solution, at a higher price, compared to a single tool.
Another difference is that as well as offering a platform which hosts a number of both native and third-party AI solutions, Arterys’ platform also includes a web-based viewer, which the vendor says negates the need for a PACS viewer from an imaging informatics vendor. Users of Aidoc’s platform, meanwhile, will still require a traditional PACS from another vendor. This has potential to be a challenge for Aidoc, given that many of the large imaging IT vendors also offer AI platforms which could offer similar functionality to Aidoc’s. As such, why would a provider use a third-party AI platform when it already has one as part of its imaging IT system?
To convince these providers to use its own platform, Aidoc will need to demonstrate the additional value it brings to a provider. The fact that, through its own in-house development and its partnership with third parties, it has been able to curate solutions that address entire use cases adds value beyond what an imaging IT vendor’s platform might readily offer. Aidoc’s direct-to-customer sales strategy and platform play enables its customers to more easily scale up their AI offerings, rather than the provider having to do the extra legwork of integrating a number of different AI tools from different vendors to achieve the same result.
Furthermore, should providers be more willing to use Aidoc’s platform given the value it provides, regardless of their PACS vendor, larger imaging IT companies could also be motivated to integrate the AI developer’s offering into their systems. This would, after all, immediately add considerable functionality to an IT vendor’s system, rather than it having to forge numerous separate partnerships with a whole host of individual AI developers.
A further argument Aidoc will have in its favour is its range of native algorithms. If Aidoc can establish the reputation of its own solutions and position the tools it gets from its partners as a supplement to these native tools, highlighting that they have all been validated for its own platform, then the fact it has already completed some of the heavy lifting could also fall in its favour.
Aidoc’s Calling
Aidoc’s success with this strategy remains to be seen, but its continuing drive for broader solutions comprised of both native and third-party algorithms and deployed via its own platform does highlight the evolution in the medical imaging AI market.
It has become apparent that it is not viable in the long term to be a company whose only products are narrow, single purpose algorithms. While this may have been feasible several years ago due to the novelty and hype surrounding deep learning, vendors now need to offer more complete solutions that solve providers’ problems. Partnerships are for many vendors such as Aidoc, a preferred way to achieve this, as it enables the vendor to leverage another vendor’s expertise and investment for the benefit of its own customers, and scale its solutions much more rapidly. The progression of some AI developers, from being a provider of AI tools to one with a portfolio of AI solutions and maybe a platform, is, in this light, a response to the challenge of getting these complete solutions to customers, and as such it is increasingly necessary to consider.
Aidoc and its ilk’s biggest challenge may be retaining relevance in the face of an increasingly saturated market for AI platforms. However, by giving providers carefully curated AI solutions that address specific use cases, while also enabling them to take advantage of the innovative, best of breed solutions from other AI vendors, Aidoc may yet be able to carve out its own segment. It will need to both continue to develop its own capabilities (native AI solutions and platform) and continue to carefully curate a band of partners which will enable it to comprehensively address certain cases, but in doing so Aidoc may have found its calling. This vendor’s course has been set.
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