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.
In February, Sirona Medical announced that it was acquiring the AI capabilities of Nines, a young vendor known for its teleradiology offering. Included in the deal are Nines’ FDA-cleared algorithms as well as its clinical data pipeline, machine learning engines and the vendors’ AI-powered radiology workflow management and analytics offerings. In addition, several key personnel will join Sirona Medical.
The Signify View
While AI will transform medical imaging over the coming years, its widespread adoption is, at present still very much in its infancy. Almost all medical imaging IT vendors offer some AI capability, whether it’s through AI partnerships, AI marketplaces or native development, but the as-yet limited adoption, as well as the fierce rate of change in the industry, means that much of this capability is based on temporary arrangements. Imaging IT vendors are yet to acquire AI capability, due to the crowded market of 200+ AI vendors to choose from, creating a reluctance to hedge their bets and acquire an asset that may not become a market leader longer term. Instead, for minimal outlay, a vendor can partner with an AI-specialist to carry their solutions, safe in the knowledge that if the market changes, the product performs poorly commercially, or a better alternative comes along, they can, simply walk away without incurring additional risk.
Sirona Medical, a vendor that only emerged from stealth some six months ago, however, has shrugged off this cautious approach and spent an undisclosed portion of its $22.5m Series A funding on acquiring AI capability. This is a move that is, if nothing else, bold.
One explanation for this approach is that Sirona Medical is looking to establish its own AI division, with a strategy that somewhat mirrors that of RadNet. While that company is primarily an imaging service provider, the vendor identified the opportunity that AI offered, and made a string of acquisitions, starting with Nulogix in 2019 and DeepHealth in 2020. These acquisitions formed the core of RadNet’s own AI division, and were determined to represent a better investment than developing an AI division from scratch, allowing the firm to actively target the high-volume screening market.
An Acquisitive Advantage?
Sirona could likewise have seen the merit in acquiring a fully functional AI division, with proven personnel and several US-FDA approved products, giving it, at a stroke, a competitive differentiator and an in-house team that can meet its own strategic needs. Partnerships are cheap and flexible, but they are reliant on finding a partner which can offer the capability a vendor needs. Over time, having a development team in-house can bring increased flexibility and more specialised solutions.
This option could have been too tempting for Sirona to resist, especially if Nines was available at an attractive price. This seems likely. Its relentless focus on efficiency makes teleradiology one of the markets that is most suited for AI incorporation, with any marginal time saving offering sizeable returns over the longer term. As such, it is surprising that Nines was willing to let its radiology AI tools go. This divestiture suggests that its AI tools were not having the impact Nines had originally hoped. What’s more, with the vendor having received its last funding round in December 2019, free cash could be in short supply and Nines could have been looking to streamline, removing parts of the business that distract from its core teleradiology offering.
Sirona is, on the face of it, a good home for these tools. Its offering is entirely cloud-based, which lends itself to the deployment of AI. What’s more, Sirona, like other “challenger” imaging IT vendors such as Visage and Sectra, isn’t burdened by the legacy architectures that many larger imaging IT vendors, which amassed their capability through acquisitions, will have to face. While more traditional vendors will have to reengineer and consolidate large parts of their portfolios to be able to capitalise on new technology trends, Sirona should have a cohesive and unified platform from the start.
Algorithmic Appropriateness
Conversely, the relevance of one of the acquired algorithms is questionable. Sirona already has a partnership with Reveal DX, which offers a tool for the quantification and classification of lung nodules, capability which the acquired NinesMeasure pulmonary nodule quantification tool appears to overlap. Sirona could potentially expand the capabilities of that tool into something which more obviously aligns with its own strengths such as a detection tool, which would open up lung screening pathways, although this would likely prove resource intensive.
Nines’ FDA-cleared tools for intracranial haemorrhage and mass effect CT diagnosis more clearly complement Sirona’s offering. These tools could, with further development or partnership for vascular occlusion and perfusion analysis tools for example, form the basis of an end-to-end stroke solution. Such a move would be bold, but adding such a high-value AI capability within a lightweight workflow layer would certainly represent a differentiator, even accounting for the challenges in flexibility and scalability compared to a platform-based approach, not to mention the stiff competition for such a tool.
One risk of this integrated approach, however, is in the commodification of AI. Although interest in AI is gathering momentum, an imaging IT vendor’s challenge remains converting customers interest into willingness to pay for AI capability. With Sirona owning the AI capability, the vendor has flexibility in its commercial offering to support adoption, without the need to overcome partners’ unrealistic price points for AI algorithms. Commoditising AI will remain a challenge, but one that needs to be done to derive revenues from its AI products.
Exhibit A
Whether the vendor can achieve this in the longer term remains to be seen. In the shorter term, the firm faces a more mundane risk; trying to accomplish too much too quickly. It is true that young vendors need to progress at a rapid pace, to begin bringing in revenues and build their customer bases. However, healthcare tends to move slowly. As well as factors like clinical validation and regulation slowing an AI product’s adoption, contract lengths, budgeting periods and renewal cycles for imaging IT solutions can also frustrate young hungry innovators looking to make their mark. The impact of these delays can be compounded if a vendor is trying to do too much too quickly, a trap that Sirona, a vendor that is at once developing its product, commercialising its product, integrating an acquisition and establishing a new division, risks falling into.
Should it be able to manage this range of interests, however, the vendor’s bold approach could pay off. The medical imaging IT market is very consolidated and doesn’t typically offer the growth prospects that a young vendor seeks. As such, Sirona’s unique approach, in creating a lightweight, vendor neutral workflow layer with fully built-in AI functionality, which could lead to a broader imaging IT system is audacious, but it does bring useful technology, personnel and experience on board. Ultimately, to outperform a market, a vendor must take a different tact to the established competitors in that market. The acquisition of Nines is evidence of Sirona’s first step down this new path.
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