Fueling the future of pathology
The Bigpicture initiative is heading towards its mid-term and is progressing steadily. The involved partners know that pioneering is hard work, and it requires determination, passion, and maybe a little courage. But one thing is certain: Bigpicture is here for the long run. “An initiative on this kind of scale, a vision like this, like having an absolutely massive database of images that will continue to grow and continues to fuel research; so much more will come from this”, says Dr. Mike Montalto, Chief Scientific Officer, PathAI.
Whilst attending the Second Bigpicture Annual Meeting as a keynote speaker earlier this year, Mike was surprised by the scale of the consortium and the combined efforts to overcome barriers. Mike: “Bigpicture is doing more than just pushing for the digitizing of WSI and archives of slides. From what I have seen during the Annual Meeting is that Bigpicture is working to solve the challenges associated with such a large effort, including tackling both ethical and legal considerations and operational challenges. Bigpicture is doing more than ‘just’ digitizing science but will hopefully create a paradigm by which others can follow. I think the people in Bigpicture are doing incredible work.”
The promise of AI-powered pathology
“When AI first started to come up, I didn’t think it was that different from the perspective of imaging automation”, says Mike. At this point, Mike worked in digital pathology for almost his entire career. It was when he started working in the AI space that he realized how quickly and efficient models and image analysis could be developed with AI, and he understood this was going to be different. “Not just in our field, but in every single industry it touches. It creates opportunities to do things faster, better, and cheaper.”
About 15 years ago, as Mike remembers, pathologists would have been tremendously skeptical about choosing digital technology over ‘real’ slides. Big barriers that interfered with the technological development were that the images would load too slowly, the quality of the images wouldn’t be good enough, and the regulatory path wasn’t established. “But all these problems have been overcome for digital pathology, and now the foundation is set, most pathologists agree today that this is the new way of working. AI is the newest technological development accompanied with similar barriers and skepticism.”
There is no way back
One of the achievements in his career that Mike is particularly proud of is the creation of the regulatory pathway for digital pathology in the United States, during which time he was president of the DPA. “I was very happy to be part of the industry efforts to remove the barrier of regulatory uncertainty. And we are at this crossroads again with the regulatory that needs to be established for AI-based technologies.”
Mike is certain that AI will set a new standard, and not just in the field of pathology. “It will touch every single aspect of our lives. From medicine to consumer products. It’s like investing in electricity. After electricity was invented and we knew what it could do, who would think in their right mind that we’d go back to candles and steam engines? That is where we are headed with AI.”
And this is where Bigpicture comes in by building a large-scale repository packed with pathology data that will accelerate the AI developments in the industry. Mike: “Access to high quality data is an important cornerstone for the development of AI-based machine learning. The Bigpicture initiative has an essential role in this, as accessible data is a real bottleneck in the industry today both for academic researchers as well as for industry to create models.”
Future impact of embracing AI in our industry
There are challenges on many different levels for AI-powered pathology. For instance, the accessibility of data to create models, the regulatory challenges and payment mechanisms, and the establishment of an infrastructure at an industry level to embrace AI in technology as common practice. Mike: “Pathology is still built on an infrastructure that doesn’t yet support AI. And this is, I think, the biggest challenge as it takes the most time to establish. We could have great models and great applications, but if pathology still evolves around glass today, you will still have to re-design the workflow infrastructure. And not only the systems need changing, but also people need to be re-trained. That takes even more time.”
What if we take it another step further; if we are changing existing workflows for AI to be implemented, couldn’t we also look at other aspects and processes? Mike: “What if, in a distant future, we don’t need any stained glass slides at all? If we take a section of an image, and it’s translucent, we could use AI to recolor it. This artificial coloring of the slide would show the same as if it were to be stained. People say that the history of disease has been written in pink and purple stains which will never change, but once you live in a world where computation exists ubiquitously, you can really rethink the whole paradigm.”
About Mike Montalto
Mike Montalto is Chief Scientific Officer at PathAI, where he is responsible for the strategy and execution of the entire AI-based product life cycle . He attended the Second Bigpicture Annual Meeting in Vienna as an invited keynote speaker early 2023. He has experience developing digital pathology diagnostic products and driving cutting-edge translational research in a large biopharmaceutical setting. Mike joined PathAI in November 2019 to use AI-powered pathology to support the development of new therapies and diagnostics for patients.