Dr. David Klimstra’s Presentation to PNPL’s Digital Pathology & AI Focus Group

On June 27, 2025, Dr. David S. Klimstra, Professor of Pathology at Yale University School of Medicine and Co-Founder of Paige AI, gave a compelling presentation to PNPL’s Digital Pathology and AI Focus Group. Drawing on his leadership in both academic pathology and AI innovation, Dr. Klimstra shared unique, cutting-edge insights into how artificial intelligence is transforming diagnostic workflows, shaping business strategy, and creating new opportunities for pathology practices.

Artificial intelligence (AI) is rapidly transforming the healthcare landscape, and pathology is emerging as one of its most promising frontiers. While the technology is advancing at an unprecedented pace, its integration into day-to-day pathology operations depends as much on business and workflow considerations as on scientific breakthroughs. For pathology practice administrators, understanding these developments is critical for planning investments, staffing, and operational strategies in the years ahead.

The Expanding Role of AI in Pathology

AI in pathology is no longer a theoretical exercise—it is delivering tangible results in cancer detection, grading, biomarker identification, and workflow optimization. Today’s AI systems can assist in diagnosing cancers, classifying tumor types, staging disease, and predicting treatment response. Importantly, these tools don’t replace pathologists; rather, they augment their work by reducing error rates, standardizing interpretation, and accelerating turnaround times.

For example, AI-assisted prostate cancer diagnosis has been shown to reduce detection errors by up to 70% and improve both sensitivity and specificity for pathologists across varying experience levels. In breast cancer, AI models can identify invasive carcinoma, atypical hyperplasia, and in situ lesions with high accuracy, while also expediting complex tasks such as lymph node metastasis detection and mitotic counting—areas traditionally prone to subjectivity and variability.

Why Digital Pathology Is the Gateway

The foundation for AI in pathology is the shift to fully digital workflows. Scanning glass slides into high-resolution whole-slide images enables secure archiving, rapid retrieval, and remote consultation, while also unlocking the potential for AI-driven computational analysis. For administrators, digital adoption offers clear operational advantages: improved slide custody for patient safety, streamlined access for consultations, and cost-saving potential through more efficient processes.

However, full digital adoption faces persistent challenges—chief among them cost, lack of reimbursement, and inconsistent interoperability across systems. Without these hurdles addressed, AI deployment at scale will remain limited.

Demonstrated Operational Impact

From an administrative standpoint, the operational benefits of AI are particularly compelling. Time-and-motion studies show that pathologists spend over a third of their day on slide review and another third on reporting, with the rest split between workflow-related and miscellaneous tasks. AI tools can help reduce non-value-added work, enabling pathologists to focus on complex diagnostic decision-making.

In practical trials, AI has cut lymph node review time in breast cancer cases by more than half while improving detection sensitivity. In prostate biopsies, AI integration significantly reduced missed cancer cases and improved diagnostic consistency between generalists and subspecialists. These efficiency gains translate directly into improved case throughput, reduced backlogs, and potentially lower per-case costs.

Emerging Business Opportunities

Beyond traditional diagnostic support, AI is opening new revenue and service opportunities for pathology practices. Novel biomarker detection—once reliant on expensive genomic sequencing—can now be performed from standard H&E slides, potentially reducing costs for clients and expanding testing menus. AI can also identify rare tumor subtypes, stratify patients for targeted therapies, and support clinical trial enrollment, creating partnerships with pharmaceutical companies and research institutions.

Foundation models, a new class of AI trained on massive, diverse datasets, promise to make developing these specialized applications faster and more cost-effective. For administrators, this means that future AI capabilities may be delivered as scalable software updates, rather than requiring extensive new infrastructure.

Implementation Considerations

Integrating AI into a pathology practice requires thoughtful planning. Administrators must evaluate vendor offerings not only for accuracy but also for interoperability with existing LIS, regulatory compliance, and data privacy safeguards. Equally important is ensuring that AI tools are validated for the practice’s specific case mix and patient demographics, as real-world pathology is more variable than the curated datasets used in development.

There is also a cultural component: adoption depends on pathologists trusting the technology. Building confidence involves training, demonstrating value in real-world workflows, and allowing for “human-verifiable” AI output rather than opaque “black box” recommendations.

The Strategic Imperative

The question is no longer whether AI will be part of pathology, but how quickly practices can position themselves to benefit from it. As Dr. Klimstra noted, pathologists who use AI will likely replace those who don’t—not because machines will take over, but because practices leveraging AI will operate with greater accuracy, efficiency, and diagnostic range.

For administrators, the strategic priorities are clear:

• Invest in digital pathology infrastructure to enable AI integration.
• Evaluate AI tools for their operational impact, interoperability, and regulatory readiness.
• Plan for reimbursement by tracking developments in coding and payer policies.
• Engage staff early to build confidence and smooth adoption.
• Explore new service lines made possible by AI capabilities, from advanced biomarker testing to telepathology consultations.

Conclusion

AI in pathology represents both a challenge and an unprecedented opportunity for practice leaders. By taking a proactive approach—building the digital foundation, selecting the right tools, and aligning staff and workflow—administrators can ensure their practices are not only ready for AI but positioned to thrive in an increasingly competitive, data-driven healthcare environment. The future of pathology will still center on the expertise of skilled physicians, but their most successful partners will be those who understand how to harness AI to enhance both diagnostic excellence and operational performance.

Panel of National Pathology Leaders’ Member Practices have access to a wealth of information and tools to help navigate management challenges, including exclusive access to reports, data, and surveys. For information on how to join PNPL, see our website or email ak****@*********rs.org.