A look ahead at what’s to come from Paige’s healthcare experts
Digital pathology and AI have come a long way over the past year; in 2023, we saw news of major healthcare centers around the world adopting digital platforms, the introduction of advanced AI algorithms for various tissue types and biomarkers, increased regulatory approvals and AI safety protocols, as well as numerous academic papers proving the value of these technologies. Paige has been at the forefront of this progress, working with partners like Leica Biosystems and Microsoft to launch groundbreaking solutions and drive adoption across labs of all shapes and sizes, earning a new FDA breakthrough designation, and setting the stage for the most advanced AI algorithms yet thanks to our Foundation Model, to name a few.
Building on this tremendous progress toward our goal of revolutionizing the industry, Paige asked our team of experts – whose experience in the field spans decades – to look ahead at what’s to come in digital pathology, AI, and beyond for 2024. See why it’s already shaping up to be one of the most exciting years for pathology yet:
AI in Healthcare
- Multi-modal AI models that combine imaging, genomics, pathology, and language data, embedded in platforms, will emerge as key drivers for AI adoption.
- Big tech companies, such as Microsoft, will lead the way in providing healthcare platforms. Similarly, Cloud providers such as Microsoft, AWS, and Google will recognize pathology labs and digital pathology companies as top targets for cloud storage and services and continue to build their capabilities in pathology.
- The pharmaceutical industry will realize the power of AI-driven patient stratification for precision medicines, and more than one clinical trial will utilize AI as an endpoint.
- Payors will be overwhelmed with AI solutions and will need to set up a regulatory body similar to MolDX to help understand how to pay for AI clinical decision support tools.
- There will be a strong focus for regulators and legislators on the quality of data and ownership of the data used to build AI systems
- There will be a push for regulatory agencies (FDA in particular) to take an even more active role in assessing the quality of AI tools being introduced into the market, in particular to address how to validate AI algorithms that generate an output that is not human-verifiable.
Technology and Infrastructure Innovation
- Cloud will become the storage and compute backbone of choice for digital pathology. The scale, security, convenience and reduced cost of cloud services will be recognized as the preference for digital pathology labs, as it is in other areas of health provision.
- There will be widespread development and adoption of Foundation Models in healthcare and pathology. Not all of these will be strong – as Foundation Models, by their very definition, need to be built on the largest and best annotated data sets. Only very few companies will be able to offer Foundation Model services that will accelerate the research capabilities of academic and commercial R & D programs.
- The GPU race will continue, and new players will emerge.
- Generative AI and Large Language Models (LLMs) will be used to accelerate and improve pathology reporting. There will be an upsurge in the number of academic research publications on large language models, generative AI, and automated reporting in pathology in 2024. This will lead to a number of new start-ups in this field as well as LIS/LIMS companies adopting the technology to accelerate reporting in pathology.
- Pathology reporting, and diagnostic reporting more broadly, will increasingly focus on integration of findings from multiple studies, both to bring together diverse data related to a single specimen or patient (“data aggregation”), but also to enable physicians to integrate all of the diagnostic findings to achieve of more comprehensive interpretation, which can be facilitated by AI.
- Multimodal AI in oncology will supersede pathology-only AI. In the past, AI in pathology has focused on replicating tasks that pathologists do today or identifying new predictive markers from pathology images. This year, we will see an increase in the number of “multimodal” applications that pull together datasets from different fields that include pathology imaging but are not exclusively pathology related. This will be largely restricted to academic studies to begin with, but industry will start driving some of these changes to better define cancer and predict patient signatures and predictors of outcome or response to therapy.
Adoption of Digital Pathology in Clinical Settings
- 2024 will see increased adoption of digital pathology for routine diagnostic use, with new scanner clearances coming into the market and increased recognition of the positive impact AI can have on patients and overwhelmed healthcare systems
- Last-mile delivery of AI solutions, the inclusion of AI-enabled products in the workflow, and overall user experience will be key to recognizing the value that AI can bring to healthcare
- Further integration of AI tools that address other aspects of the pathology workflow, like automated reporting, case prioritization and triage, will increase rather than limit AI use to disease detection within the pathology image.
- Prospective clinical research will quantify the benefits of the digital work environment and the use of AI for pathology workflow, both in terms of efficiency and accuracy but also related to patient care benefits.
- While we might not see it come to fruition in 2024, we’re inching closer to reimbursement.
Paige is always striving to drive the industry forward, and throughout 2024, we remain dedicated to our vision of progressing what is possible in pathology through a full end-to end digital pathology platform that can integrate powerful and innovative AI applications. Our unique partnerships, commitment to collaborating with practicing physicians, and unrivaled technological capabilities have positioned us to drive the field forward for the benefit of cancer care at large.
