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.
Lunit recently announced that it has received preliminary approval for an IPO on the South Korean KOSDAQ index. Following the approval, the Seoul-based AI developer now plans to submit a listing within the first half of the year.
The outfit already has partnerships with several large international imaging vendors, including GE Healthcare, Philips and Fujifilm to incorporate its AI capability into their imaging systems, but Lunit says the money from the funding round will enable it to further develop its AI product range, and expand its global commercial reach. This will entail promoting its products, which include an AI tool for analysing mammograms, an AI solution for analysis of tissue slides for cancer biomarkers, and a comprehensive AI solution for chest X-rays that detects 10 of the most common findings, in more markets across the world.
The Signify View
That Lunit has decided to place its fortunes in the hands of public investors should come as no surprise. As detailed in a recent Premium Insight discussing medical imaging AI vendors with more than $100m of venture funding, Lunit, like many of its successful peers shares some common traits. The vendor has, for instance, taken a very robust approach to product development. Instead of relying solely on one of the available training datasets, which can introduce some ambiguity and aren’t always perfectly labeled, Lunit has chosen to use training data validated against biopsy results. This helps ground the AI tool’s algorithm and minimise the likelihood of errors. A tool is only as good as the data it is fed so supplementing images with other clinical data is a prudent approach.
In a similar vein, the vendor has also been thorough with regard to clinical validation. One of the hurdles stymieing the broader adoption of AI is a lack of robust evidence. To be profitable AI vendors need to prove their solutions are valuable to providers, to warrant hospital budgets and convince policymakers that their tools deserve reimbursement. To do this these vendors must undertake strenuous, detailed clinical validation studies. These are expensive and time consuming to conduct, but they are necessary. This need is compounded as AI vendors look to grow globally, with developers obliged to prove their solutions are as effective on non-local populations. Lunit has been able to meet this requirement, and can point to a wealth of published studies open to scrutiny, as well as regulatory approvals for its Insight CXR and Insight MMG chest X-ray and mammography solutions in the USA, Europe, Japan and South Korea and for SCOPE PD-L1 in Europe.
In addition to this technical capability, the vendor has also been commercially savvy. As well as selling its products directly, Lunit has made its products available on several AI marketplaces, allowing providers which use, for example, Sectra’s imaging IT solutions to incorporate its tools through that company’s Amplifier Marketplace.
More significantly, the vendor has also sought to utilise partnerships to target new markets. It has received $26m from liquid biopsy specialist Guardant Health. As well as boosting Lunit’s bank balance, the collaboration will help the AI vendor target the US oncology market with its SCOPE tissue analysis platform. Lunit has also looked to international imaging vendors to grow, with the company inking deals with GE Healthcare, Philips and Fujifilm helping encourage Lunit’s adoption among those vendors’ install bases.
Such endeavours are for naught, if the products themselves are inadequate. Lunit, however, has avoided this trap. While its range of products is small, comprising of three solutions, one for chest X-ray, a second for mammography and a third for pathological tissue analysis, they are focused on valuable areas. Insight CXR, for example is a solution that is on the rising tide of increasingly comprehensive AI tools which, like those from Annalise AI and Oxipit AI seek to provide greater clinical value to doctors than narrow AI solutions which are often more limited. Similarly, Lunit’s SCOPE solution is set to benefit from the growth of digital pathology and the establishment of closer ties between diagnosis and treatment.
A Time for Temerity?
Despite these strengths, however, listing publicly also represents a risk for the vendor. Reports suggest the listing offers Lunit a valuation of around $500m. This is a far cry from the valuations of the likes of HeartFlow and Viz.ai, which saw valuations as high as $2.4bn and $1.2bn respectively in their recent fundraising endeavours, however it is still a significant sum for a vendor that, according to Signify Research’s AI in Medical Imaging report, achieved just over $1 million in revenue and a loss of $16.5 million in 2020. While the IPO will furnish the vendor with ample capital, it also adds considerable pressure. It risks facing many private investors necessitating consistently strong quarterly performances, rather than sympathetic private investors au fait with the medical imaging AI market, which are likely to be understanding of more modest returns in the name of sustainable long-term progress.
There are other hurdles too. The enormous valuations that some medical imaging AI vendors have been able to achieve through the high availability of funding for AI firms over recent years, have, in some instances proved a hindrance to these AI firms when they have looked to list. The likes of HeartFlow and Keya Medical both sought to go public, before being forced to postpone their plans, in part due to the inability of public investors to match these lofty valuations. Furthermore, in many instances, vendors that did go on and list publicly have seen their share prices fall as they were unable to meet the high expectations of their public investors. This has been true of all Lunit’s closest South Korean peers.
Vuno listed at an initial price of KRW32,150 in February 2021, and is now trading at around KRW10,500. DeepNoid shares initially traded at a price of KRW25,200 in August 2021 but now sit languishing at an all time low of KRW11,400. JLK, meanwhile went public in December 2019 with an initial price of KRW8,330, rallied to KRW14,150 in September 2020 but now trades at KRW6,220. The market as a whole has suffered, with the KOSDAQ falling 13% year-to-date, but JLK, DeepNoid and Vuno have all underperformed even relative to this benchmark, falling 20%, 38% and 44% respectively.
A Hard Market
Making headway amidst these underperforming peers will prove difficult. This challenge will also be compounded by market complexities facing its products. While the tools themselves are sound, they are competing in difficult markets. Mammography for instance has well-established incumbent vendors such as Hologic and iCad, as well as a plethora of breast imaging AI start-ups each trying to eke out a share of the market. Similarly, Lunit’s Insight CXR product will not only face stiff competition, but even when used it will be fighting for a small percentage of the limited reimbursement available for chest X-rays. Both could be successful products, but could be slow to return sizable revenues. Lunit SCOPE is likely to be similar. There are considerable opportunities pertaining to digital pathology AI, clinically as well as in other areas such as drug discovery. However, the digital pathology market is so nascent, it is unlikely to significantly contribute to Lunit’s bottom line for several years.
These weaknesses do point to one area in which Lunit should prioritise following its IPO. As well as commercialisation efforts, building sales and support networks in new markets, Lunit must also spend heavily on the development of new tools. The vendor must channel its expertise into targeting new areas that offer high potential returns, whether for better-reimbursed modalities, high-value use cases, or care coordination tools that expand beyond image analysis itself, Lunit needs to supplement its current strength with tools that will be lucrative into the long term.
Don’t Look Down
Ultimately, Lunit’s growth up to this point has positioned it in a difficult spot. While the vendor must move forward, doing so requires overcoming considerable adversity. This is not a problem that is unique to Lunit, but each medical imaging AI vendor that makes it to this pivotal moment, must navigate it in its own way. Balancing the needs of its investors while continuing to invest in product development and market creation, will be tough. The need to create significant revenues and profits to justify its lofty valuation, while not neglecting the robust way it has approached algorithm training and clinical validation, which helped establish its credentials, means walking a tightrope, where one slip can send a share price tumbling.
Lunit will see a future in which the company is successful. Consistently generating revenues, with sustainable growth, a secure user base and an innovative roadmap. The vendor can lead its investors to this future, but its challenge will be in ensuring they don’t lose faith and fall along the way.
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