Send us a text

What happens when artificial intelligence moves beyond images and begins interpreting clinical notes, kidney biopsies, multimodal cancer data, and even healthcare costs?

In this episode, I open the year by exploring four recent studies that show how AI is expanding across the full spectrum of medical data. From Large Language Models (LLM) reading unstructured clinical text to computational pathology supporting rare kidney disease diagnosis, multimodal cancer prediction, and cost-effectiveness modeling in oncology, this session connects innovation with real-world clinical impact.

Across all discussions, one theme is clear: progress depends not just on performance, but on integration, validation, interpretability, and trust.

HIGHLIGHTS:

00:00–05:30 | Welcome & 2026 Outlook
New year reflections, global community check-in, and upcoming Digital Pathology Place initiatives.

05:30–16:00 | LLMs for Clinical Phenotyping
How GPT-4 and NLP automate phenotyping from free-text EHR notes in Crohn’s disease, reducing manual chart review while matching expert performance.

16:00–23:30 | AI Screening for Fabry Nephropathy
A computational pathology pipeline identifies foamy podocytes on renal biopsies and introduces a quantitative Zebra score to support nephropathologists.

23:30–29:30 | Is AI Cost-Effective in Oncology?
A Markov model evaluates AI-based response prediction in locally advanced rectal cancer, highlighting when AI delivers value—and when it does not.

29:30–38:30 | LLM-Guided Arbitration in Multimodal AI
A multi-expert deep learning framework uses large language models to resolve disagreement between AI models, improving transparency and robustness.

38:30–44:30 | Real-World AI & Cautionary Notes
Ambient clinical scribing in practice, AI hallucinated citations, and why guardrails remain essential.

KEY TAKEAWAYS

• LLMs can extract meaningful clinical phenotypes from narrative notes at scale
• AI can support rare disease diagnosis without replacing expert judgment
• Economic value matters as much as technical performance
• Explainability and arbitration are becoming critical in multimodal AI systems
• Human oversight remains central to responsible adoption

Resources & References

• Digital Pathology Place: https://www.digitalpathologyplace.com
• Digital Pathology 101 (free PDF, updates included)
• Automating clinical phenotyping using natural language processing
• Zebra bodies recognition by artificial intelligence (ZEBRA): a computational tool for Fabry nephropathy
• Cost-effectiveness analysis of artificial intelligence (AI) for response prediction of neoadjuvant radio(chemo)therapy in locally advanced rectal cancer (LARC) in the Netherlands
• A multi-expert deep learning framework with LLM-guided arbitration for multimodal histopathology prediction

Support the show

Get the "Digital Pathology 101" FREE E-book and join us!

Send us a text

What really changed in digital pathology this year—and what still needs work?

As we close out 2025 and step into 2026, I wanted to pause, reflect, and share what I’ve seen shift from theory to real-world practice across labs, conferences, and clinical workflows.

I look back at the most meaningful developments in digital pathology and AI in 2025—from wider adoption of primary diagnosis on digital slides to more grounded, evidence-driven use of AI tools. We’ve moved past hype and pilots and started asking harder questions about validation, workflow integration, regulation, and trust.

I also share what I believe matters most as we move into 2026: building real-world evidence, upskilling pathologists, and focusing on tools that genuinely support patient care rather than distract from it.

This episode is for anyone navigating change in pathology and wondering where to invest their time, energy, and curiosity next.

Episode Highlights:

• [00:00–02:10] Why 2025 marked a turning point for digital pathology adoption
• [02:10–05:40] From pilot projects to clinical workflows: what actually changed
• [05:40–08:30] How AI usage shifted toward triage, quantification, and decision support
• [08:30–11:45] Why validation and real-world evidence became central topics
• [11:45–14:20] The growing role of pathologists in AI governance and quality assurance
• [14:20–17:10] Lessons from conferences, labs, and conversations worldwide
• [17:10–20:00] What I expect to see more of in 2026—and what I hope we leave behind
  
  

Key Takeaways:

• Digital pathology is no longer experimental—it’s becoming routine in more labs.
• AI tools are shifting from novelty to practical clinical support.
• Validation, regulation, and workflow fit matter more than algorithm performance alone.
• Training and continuous learning are now essential career components for pathologists.
• 2026 will reward teams that test, measure, and iterate thoughtfully.
  
  
  

Resources Mentioned

• Digital Pathology Place – education, podcasts, and community
• 2025 CONFERENCE insights and real-world lab experiences
  
  

Support the show

Get the "Digital Pathology 101" FREE E-book and join us!

Send us a text

What if the biggest breakthrough in pathology AI isn’t a new algorithm—but finally sharing the data we already have?

In this episode, I’m joined by Jeroen van der Laak and Julie Boisclair from the IMI BigPicture consortium, a European public-private initiative building one of the world’s largest digital pathology image repositories. The goal isn’t to create a single AI model—but to enable thousands by making high-quality, legally compliant data accessible at scale.

We unpack what it really takes to build a 3-million-slide repository across 44 partners, why GDPR and data-sharing agreements delayed progress by 18 months, and how sustainability, trust, and collaboration are just as critical as technology. This conversation is about the unglamorous—but essential—work of building infrastructure that will shape pathology AI for decades.

⏱️ Highlights with Timestamps

• [00:00–01:40] Why BigPicture focuses on data—not algorithms
• [01:40–03:16] Scope of the project: 44 partners, 15–18 countries, 3M images
• [03:16–06:20] The 18-month delay caused by legal frameworks and GDPR
• [06:20–11:52] Extracting data from heterogeneous lab infrastructures
• [11:52–13:38] Current status: 115,000 slides uploaded and growing
• [13:38–18:39] Why LLMs and foundation models make curated data more valuable than ever
• [18:39–23:49] Industry collaboration and shared negotiating power
• [23:49–28:06] Data access models and governance after project independence
• [28:06–31:59] Sustainability plans and nonprofit foundation model
• [37:02–43:18] Tools developed: DICOMizer, artifact detection AI, image registration
  
  
  

📚 Resources from This Episode

• IMI BigPicture Consortium
• GDPR & Data Sharing Agreements (DSA)
• DICOMizer & SEND metadata tools
• Artifact detection AI for slide QC
• European AI Factories initiative

Support the show

Get the "Digital Pathology 101" FREE E-book and join us!