TeraRecon Acquired by SymphonyAI – The Signify View

Published: March 6, 2020

On the 4th March 2020, TeraRecon announced it had signed an agreement to be acquired by the SymphonyAI Group. Here are the facts of the deal, followed by our take on the potential impact on the medical imaging market:

  • The terms of the deal were not disclosed and TeraRecon will continue to operate business as usual. However, it will benefit from greater financial backing and the synergistic technologies from other companies in the Group’s portfolio.
  • TeraRecon is an Advanced Visualization (AV) and Artificial Intelligence (AI) software developer with three main solutions: iNtuition, its AV platform; EnvoyAI, its AI marketplace for radiology; and NorthStar AI Explorer, the platform interface between its marketplace and the end user.
  • The SymphonyAI Group is a portfolio of companies developing new generation AI solutions across multiple industries. Its founder, Romesh Wadhwani, has committed over $1 billion to build the Group’s portfolio.
  • TeraRecon is the Group’s seventh company in its portfolio, and the second in healthcare; the other, Concerto Health AI is an oncology AI software developer. A third portfolio company, Ayadsi, is also active in healthcare, alongside financial services and the public sector, and offers clinical variation management, population health management and denials management solutions to support value based care providers.

 The Signify View

TeraRecon reports it will continue business as usual; however, the company will benefit from a much greater financial backing, which will facilitate the growth and development of its product offerings. We have identified four areas of this deal that may have a significant influence on the medical imaging market:

  • The synergies between TeraRecon and other companies in the SymphonyAI Group, including Concerto HealthAI.
  • Development of new products, such as ‘AI Suite’ solutions.
  • Protecting and enforcing its intellectual property rights, following the recent award of two patents to TeraRecon relating to how healthcare providers will consume, integrate and adopt AI algorithms into their workflows.
  • Strengthening its core AV business.

 Group Synergies

Alongside the capital injection, one of the main benefits of the deal for TeraRecon will be synergy opportunities with other SymphonyAI companies, including Concerto HealthAI and AyadsiAI. For example, Concerto’s eurekaHealth platform, which incorporates data from the American Society of Clinical Oncology’s CancerLinQ and other real-world sources, could become a common platform across Symphony’s healthcare activities, which will have many benefits.

By using a common platform, population data from Concerto HealthAI can be combined with imaging data from TeraRecon, for example to improve patient stratification for oncology clinical trials. TeraRecon’s AI solutions could then be used to provide quantitative image assessments throughout the trials, improving their efficiency and efficacy. For health providers, a common platform could lead to the development of a more integrated care management solution that combines AI-analysed imaging data from TeraRecon with other data sources, for example, electronic health records and real-world evidence from Concerto. This would provide clinicians with richer information for diagnosis and treatment planning and plays well to the current focus on precision medicine.

A common platform will also enable patient records and clinical outcomes to be matched with images to create highly targeted training data to develop new AI algorithms. The additional financial resources and access to the SymphonyAI Group’s machine learning expertise and clinical data will enable TeraRecon to accelerate its in-house development of AI algorithms and solutions.

New Product Development – AI Suites

At RSNA 2019, TeraRecon announced AI Sync, a software enhancement to the EnvoyAI platform that automatically prepares AI results and presents them for verification before, during or after the radiologists’ interpretations are completed. The company also announced three AI Key-Workflow subscription offerings for Stroke and Trauma, Auto TAVR and Cardiac MR. These leverage AI Sync to deliver AI results in the form of dashboard views that radiologists can engage with and verify as they generate reports. AI Key-Workflows are optimized applications designed to deliver clinical data and iNtuition workflows accelerated with AI functionality and automation.

TeraRecon has signalled that its AI Key-Workflow solutions will evolve to AI Suites. These will combine native and third-party algorithms to create workflow solutions for a variety of specific use-cases. This is a natural evolution for the AI market, from today’s multiple disparate point applications, to more complete solutions that address specific bottlenecks in the radiology workflow. With algorithm developers facing increasing competition for point solutions as the number of regulatory clearances rises, complete workflow solutions are a key strategy for differentiation. And potentially one with greater financial rewards.

We have identified four ways that complete workflow solutions are likely to make a market impact, namely: (1) simplifying the selection of AI algorithms; (2) streamlining the deployment and integration of AI algorithms; (3) using AI to solve a clinical need; and (4) more cost-effective, bundled pricing.

  1. AI marketplaces were the first step in simplifying the selection of AI tools by providing radiologists with a curated offering from a variety of algorithm developers. However, these are mostly narrow point solutions for specific tasks and radiologists are still required to make multiple algorithm selections. Complete workflow solutions, such as AI Suites, aim to simplify the selection process one step further by combining algorithms for a specific use-case into a single offering. For example, combining multiple vessel analysis algorithms into a single solution for TAVR planning. This solutions-based approach reduces the hassle of evaluating, selecting and deploying multiple applications, and applies AI to a wider aspect of the radiologist’s workflow.
  2. Complete workflow solutions help to streamline the deployment and integration of AI, when compared with a multiple stand-alone AI algorithms approach. The results can be combined and presented in a single dashboard view, unifying the user’s experience and providing faster and more easily accessible results.
  3. Complete workflow solutions combine multiple AI algorithms to solve existing problems and to enhance radiologist efficiency and performance. This way, AI becomes an ingredient in a solution, rather than positioned as a product in its own right.
  4. Complete workflow solutions are expected to be more cost-effective for healthcare providers than the ad-hoc purchase of individual algorithms.

TeraRecon plans to promote its AI Suites through the EnvoyAI marketplace and will also seek distribution partnerships with other marketplace or imaging IT vendors.

 Intellectual Property

Alongside the acquisition announcement, TeraRecon also announced two first of their kind patents for AI in medical imaging. The patents pertain to how healthcare providers will consume, integrate and adopt AI algorithms into their medical imaging workflows.

The first patent pertains to the radiologist’s ability to modify, accept or reject findings from AI algorithms and the ability of the AI platform to synthesize these. The second patent pertains to orchestration in the workflow, and how AI platforms will use new and old imaging data to enable AI results to be delivered in a single user experience.

The SymphonyAI acquisition will give TeraRecon the financial firepower to protect and enforce these patents. The company has signalled it plans to pursue infringements of its patents, but also wants to position itself as a technology enabler by licensing its technologies to other vendors.

 Impact on Core AV Business

TeraRecon is a significant vendor in the AV software market, holding a top 10 share of the market globally and top 5 position in the North American market, based on revenue share. However, it has for some time been one of the few remaining independent AV firms, competing against  major healthcare technology vendors like GE Healthcare, Siemens Healthineers, Philips and  Canon (Vital Images), all of which have extensive advanced imaging modality hardware businesses and considerably larger operational and R&D war chests. While TeraRecon has used its independence as an advantage in past years, recent market changes have softened this message. Broadly, large healthcare providers in North America and internationally are looking to consolidate imaging informatics, under a broader enterprise imaging strategy. This has placed more focus on a “one throat to choke” play, with major deals leaning towards vendors that can offer not just AV platforms and tools, but also offer the full range of clinical viewing, diagnostic toolset (PACS), operational and workflow modules, image exchange and vendor neutral archiving. Of the top five market leaders, Terarecon is the only one without a broader informatics portfolio.

TeraRecon has (and correctly in our opinion) expanded its focus to the emerging AI market. By positioning itself as a best-of-breed AI vendor with established AV capabilities, it has stolen a lead on its larger competitors in terms of AI platform development. This has also paid dividends for the firm in terms of retaining its existing AV customer base; the optics of leadership in the AI segment has in some cases prompted contract renewals and upgrades to higher tiers of its AV products. In doing so, this has for the near-term stemmed the tide in terms of addressing how Terarecon can compete with the growing push towards imaging IT consolidation and single vendor purchasing.

The SymphonyAI deal will have the most impact in the mid- to long-term. Given the greater R&D firepower and advanced modality businesses of its market peers, its arguable that TeraRecon may have struggled to compete as an independent AV firm in the AI era for AV. However, with a considerable backer in SymphonyAI, it is now better positioned to stay ahead of the curve and develop a compelling best-of-breed, AI-enabled imaging analysis platform that customers will prioritize over single vendor consolidation. Of note from the announcement was the focus on “image-related decision support”, with neurology and oncology highlighted as new areas of focus outside of its more familiar radiology, cardiology and vascular surgery segments. This broader capability will be necessary to compete with its peers, but also factors in the growing use of imaging and decision support in multidisciplinary care pathways, especially in oncology where demand for solutions to support “tumour board” care team meetings has ramped-up considerably.

However, given its focus on AV and AI for image analysis, TeraRecon will also need to address how it can play more heavily in the diagnostic and clinical workflow sector to fit with established and emerging care pathways; here its peers perhaps have the edge, with GE Healthcare partnering with Roche for oncology solutions (Navify) and Philips already working on clinical decision support with its Intellisite Oncology offering.

The new patents described above will go some way to protecting the use of AI within the diagnostic reporting process, but this is only part of the broader decision support requirement, meaning TeraRecon will need to decide whether to more extensively partner or self-develop its own offerings. One of the agenda items for its new owners will be working with TeraRecon on its mid to long term strategy. Three questions should be front and centre of this discussion:

  1. To drive focus on best of breed AI for image analysis and AV, or to broaden its focus on establishing a larger platform around complete diagnostic and clinical care pathways (potentially leveraging SymphonyAI’s other healthcare assets)?
  2. How to address the rising tide of enterprise imaging implementations and customer preference for consolidated imaging informatics platforms from a single vendor?
  3. How to make a financial success of AI (and which business model), given the vast majority of TeraRecon revenues today stem from traditional AV platform sales?