Earlier this year, my colleague, Cameron Braverman, looked at the digital pathology and MTA abstract collaborations at AACR to identify key research contributors within the space. Inspired by this, I decided to look at digital pathology partnerships more broadly, to identify the companies and collaborations that are enabling more widespread access to digital pathology tools.
‍

Spurred by the pandemic, wider access to large datasets, and the promise of emerging analytics tools, digital pathology is increasingly utilized by select stakeholders to improve consistency, efficiency, and collaboration within the pathology lab. In the coming years, improving diagnostic, prognostic, and predictive algorithms are expected to find their way into routine care, cementing the use of DP in primary diagnostics. However, the vast majority of hospital labs do not currently have DP infrastructure in place. To reach this goal, labs will need to adopt the many components (i.e., slide scanners, AI tools, workflow software, and data storage capabilities) that make up a fully functioning digital pathology environment.
‍

The digital pathology landscape is quite fragmented today, with no single player offering all of the elements needed for pathologists to adopt and utilize these tools. In an effort to combat this, digital pathology companies have rapidly been expanding their partnerships, with ~80 publicly announced partnerships since the beginning of 2021 alone. Approximately half of these deals include slide / workflow management software companies, who are partnering to develop networks of scanner, AI, data storage, and LIS companies in an effort to streamline adoption. Within the other 50%, there are a number of partnerships between algorithm companies and research partners, which are facilitating the development of these AI tools. A combination of technology, research, and market access partnerships will be of the utmost importance for a developing digital pathology ecosystem that seeks to revolutionize how patients are diagnosed and treated.

Notes: 

^1. Company size corresponds with number of partnerships, link thickness corresponds with recency of the partnership (thicker lines are more recent; partnerships for which a year could not be established are set at minimum thickness). 

^2.Partnership activity was pulled from company websites and press releases and is based on publicly available information. Research collaborations that have resulted in pubs / conference abstracts but that have not been publicized as partnerships may not be included. Partnerships announced or referenced in the last five years were prioritized; older partnerships may be missing. Not all partnerships may be active and ongoing. Diagram is not exhaustive of all digital pathology partnerships. 

^3.Many companies have offerings spanning multiple product categories; the primary or most comprehensive product offering was used for node shading.

^4.Incubators, accelerators, research groups, and investors were excluded.

^5.There are additional pathology companies that have been excluded because they do not have digital pathology offerings. Partnerships with companies that are unrelated to digital pathology have been excluded 

^6.Research consortia were generally excluded unless there were multiple industry participants.

^7. Some other partnerships that have been excluded due to our criteria, but are of note include Reveal Biosciences partnering with Quantumcyte, Akoya partnering with Nikon, Crestoptics, and Andor, Owkin partnering with Arkhn, Roche Diagnostics partnering with Sonrai Analytics and QUB, Scorpio partnering with Beckman Coulter, Corista partnering with Elsevier, and the One Dorset Pathology Network

^8.Olympus has rebranded their life sciences division as Evident

‍

1. Slide Scanners - Like most of the digital pathology landscape, the scanning market is fragmented, with no single player dominating the space. The expertise needed to develop scanning hardware is fairly distinct from that needed to develop standalone digital pathology software applications, thus many of the leading scanning companies (e.g., Hamamatsu, Fujifilm, Zeiss, Olympus) are optics / lens companies, rather than healthcare companies. And though there are a number of companies with scanners that are building out more comprehensive offerings (e.g., Leica, Roche, Philips), the fragmented landscape (and difficulty in getting labs to switch scanners once adopted) incentivizes partnering over acquisition / building for market capture. The most frequent collaborations in this category are with workflow software companies, to ensure access to the software needed to upload, analyze and store scanned images with ease, though scanner companies may also partner with AI tools companies directly.

‍

2. AI tools / Algorithms - AI tools companies address a wide range of applications, including workflow efficiencies (e.g., PD-L1 counting, ROI identification), primary diagnosis (e.g., Prostate Dx), and novel biomarker identification for prognosis and therapy selection. Given the diversity of applications, there is room for a long tail of players to carve out their own niche. Though there are select companies that have developed slide viewers on which their algorithms can be deployed (e.g., Paige, Deep Bio, PathAI*), most rely on partnerships for widespread access. Thus, the primary category of collaboration is with workflow software companies, which can serve as an “app store” for algorithms. These partnerships will be key for ensuring the algorithms can be deployed across clinical and pharma labs, regardless of which scanner is used. Additionally, there are a number of research collaborations with biopharma and Academic Medical Centers (AMCs), which are key for algorithm development (but have not been exhaustively covered here). 

‍

3. Slide / Workflow Management Software- Particularly important in the clinical setting, workflow management software covers a number of applications, including slide viewing, slide storage, and algorithm deployment. These platforms serve as the central hub of digital pathology within a lab, ensuring slides can be viewed remotely, analyzed, and archived for future clinical care or training. This software needs to be interoperable with slide scanners upstream and AI tools, data storage, and LIS downstream, resulting in the highest average number of collaborations per company of any category; Inspirata has the most partners of any company with 25. Though scanner and LIS interoperability is possible without explicit partnerships, it will likely be important for regulatory approval going forward. While many of these companies offer basic image analysis capabilities and are eyeing ML / AI algorithm development (e.g., Proscia, 3DHistech), these softwares primarily serve as a hub through which 3rd party tools can be accessed.

‍

4. Data - Data partnerships are not always explicitly called out, but data storage is a key piece of the digital pathology ecosystem. Not only are vast amounts of data generated (~10x more than radiology images), but also, once stored, these slides need to be readily retrievable for future use. Though many hospitals will have some amount of data storage onsight, widespread slide digitization is not possible without cloud storage enabled by companies such as AWS, Dell, and Microsoft. For DP data generated in retrospective studies, pharma companies may work with these cloud storage providers directly, however AMCs are likely to access them via workflow management software partnerships. In addition to storage, data providers can also provide data tokenization, integration, management, and processing.

‍

5. LIS / LIMS - Laboratory Information Systems (LIS) have become increasingly interoperable with lab hardware, reducing the need for explicit partnerships. However, there are a handful of partnerships between workflow software providers and LIS companies to ensure seamless data transfer.  This can be especially beneficial for hospital labs where there is limited IT expertise for linking disparate software within a lab.

‍

6. Multiplex tissue analysis reagents - Though not always digital pathology companies by definition, most multiplex analysis (above 3-4 plex) requires digital pathology tools for interpretation. As a result, multiplex reagent companies are nestled within the digital pathology ecosystem. Companies such as Ultivue and Akoya have partnered with AI tools companies to supplement their own internal analytical capabilities. Additionally, Akoya has developed their own image management software which is compatible with a variety of tools. (Read more about AI tools / Algorithm companies, MTA companies, and their pharma partners in our Spatial Biology + Digital Pathology Competitive Intelligence Report - H1 2022)

‍

In addition to these companies, there are a number of other categories of participants that contribute to data generation and technology access, rather than technology commercialization. Of these, AMC / hospital partners are particularly valuable due to the large swaths of real-world clinical data to which they have access. These partnerships, as evidenced by MSK and Paige, are immensely important for AI algorithm development, which requires large clinical datasets for training and validation. Pharma partners are also be key for algorithm development, particularly when tied to therapy selection. Pharma will be vital for driving long term adoption of DP through CDx deals, which push DP from a nice-to-have tool to a must-have. (These collaborations can be read more about here). Finally, while the ultimate goal of DP is decentralized access, in the near term, reference lab partners are vital for clinical uptake. Large players such as Quest, Labcorp, and ARUP have taken steps to digitize their pathology workflows, which could serve as a short-term bridge to algorithm access while DP infrastructure remains limited to large U.S. AMCs and certain European health systems.
‍

Some consolidation is expected in the longer term, with larger players acquiring smaller fish, both within categories (e.g., large algorithm developers acquiring smaller Dx algorithm developers) and between categories (e.g., DP workflow software developers acquiring clinical algorithm developers for their platform). Furthermore, digital pathology companies are expected to build out increasingly comprehensive software solutions for their customers (e.g., algorithm companies developing slide viewing software). However, as long as the DP landscape remains segmented (and specialized) and algorithm companies target specific indications / applications, collaboration will remain an essential aspect of the digital pathology ecosystem.

‍

Note: *Though historically a pharma services company, PathAI recently announced FDA and CE-IVD clearance of a slide viewing platform; additionally, they have partnered with LabCorp (and acquired Poplar Health, a U.S. reference lab) for access to clinical channels

‍

Source: Company websites; Company press releases; DeciBio expertise; Chlipala, E., et. al., “Archival and Retrieval in Digital Pathology Systems”, Digital Pathology Association. https://digitalpathologyassociation.org/_data/cms_files/files/Archival_and_Retrieval_in_Digital_Pathology_Systems.pdf

‍