Building what no organization could build alone

“What remains distinct about Bigpicture and its unique value in the pathology AI ecosystem is the size of the initiative and its non-proprietary, cross-organization nature, integrating industry, academia and regulatory,” says Page Bouchard. “That can not be replicated by any individual institution.”

That distinction matters at a time when activity in pathology AI is accelerating. Pharmaceutical and biotechnology companies, academic groups and technology developers are all developing tools, datasets and applications. Yet most are working within the boundaries of their own organizations, data or use cases. Bigpicture was designed to cross those boundaries.

What makes Bigpicture difficult to replicate

The idea behind Bigpicture appealed to Page from the outset because he had already seen both the potential of AI in pathology and the limits of what individual organizations could achieve on their own. “We recognized the power of a consortium in building a much larger, more robust and non-proprietary database,” he says. Scale is an important part of that proposition, but it is not only about the number of images. Bigpicture brings together data, infrastructure, expertise and perspectives from across industry, academia and other parts of the pathology ecosystem. An individual organization may be able to build its own dataset, and focused tools or AI capabilities, but recreating a shared infrastructure of this scale across organizations, sectors and countries is far more difficult. Bigpicture gives participants an asset they can build on collectively rather than having to recreate every element independently.

The value isn’t just in the data

The benefits of that collaboration already extend beyond the repository itself. Preparing pathology data for use at scale requires many different functions to work together. Pathologists, data scientists, IT specialists, legal teams, privacy experts and governance professionals all become part of the same process. Bigpicture has helped bring those conversations together. “It drove action within companies in a common way, something that would not have happened in the same comprehensive way without Bigpicture as the underpinning,” Page says. That experience matters as organizations move from experimenting with AI toward integrating it more broadly into research and pathology workflows. Technology is only one part of that transition. Data quality, governance, privacy, infrastructure and internal processes all have to develop alongside it. Many of those challenges are shared, and addressing them collectively gives participating organizations access not only to infrastructure, but also to experience and knowledge they would otherwise have to develop largely on their own.

The obstacles are part of the value

Building an infrastructure of this scale inevitably takes time. Early expectations around AI often focused on how quickly algorithms and digital tools could be developed. In practice, implementation involves far more than technology. Bigpicture has had to navigate privacy, legal and regulatory requirements, technical infrastructure and the realities of working across multiple organizations and countries. “If you don’t start, you’ll never get anywhere,” Page says. “Once you start, you always run into more obstacles than you anticipated, and it always takes longer than you anticipated.”

The value lies partly in solving those obstacles together. Questions around data, governance, access and implementation are not unique to Bigpicture. They are challenges many organizations face as pathology becomes more digital and AI-driven. The knowledge and infrastructure developed through Bigpicture can therefore support progress well beyond an individual project requirement.

Being unique isn’t enough

Bigpicture’s long-term strength lies in continuing to develop upon what has already been created. The infrastructure is not a static database. Its usefulness can grow as more data becomes available, new capabilities are added and more organizations find ways to use the resource. “The value proposition has to be very clear and very obvious to the stakeholders,” Page says. “If it doesn’t add value, then it’s not a good investment.”

That value will take different forms. In non-clinical pathology, Page sees opportunities for AI to improve consistency and quality while accelerating workflows and increasing productivity. Clinical and diagnostic applications bring additional privacy, legal, ethical and regulatory requirements, but the broader direction is clear: pathology is becoming increasingly digital and data-driven. A shared infrastructure can support that transition in ways that go beyond what any single participant can access or develop alone. As the resource evolves, the opportunity is to keep making those benefits visible to existing partners as well as organizations discovering Bigpicture for the first time.

Long-term development also requires looking beyond short planning cycles. “It’s got to have a vision of what’s the plan for three years and five years and ten years,” Page says. Communicating that vision is part of the challenge. As he puts it: “Now it becomes selling.” In this context, that means clearly demonstrating what Bigpicture makes possible and giving organizations compelling reasons to use it, contribute to it and support its continued development.

Success means becoming indispensable

Page has a straightforward measure of success: Bigpicture becomes something organizations actively rely on. “The sign that it has been successful is that stakeholders realize it is a critical component of their AI strategy in pathology, and that they could not be successful without it,” he says. That would be reflected in continued participation, but also in how the resource is used: in research, scientific publications, regulatory work and new AI applications. The more Bigpicture becomes embedded in those activities, the clearer its value becomes.

The opportunity may also reach well beyond the European community that created it. “There’s nothing like it in the United States. There’s nothing like it in Asia that I’m aware of,” Page says. “Therefore, it’s unique in the world. It’s not just unique in Europe.” That creates room for Bigpicture to develop into a broader international resource for pathology data and AI, with more organizations contributing to it, using it and building on the infrastructure already in place.