155: AI Pathology & Genomics_ A New Benchmark for Predicting Gene Mutations

Digital Pathology Podcast

Content provided by Aleksandra Zuraw, DVM, PhD, Aleksandra Zuraw, and DVM. All podcast content including episodes, graphics, and podcast descriptions are uploaded and provided directly by Aleksandra Zuraw, DVM, PhD, Aleksandra Zuraw, and DVM or their podcast platform partner. If you believe someone is using your copyrighted work without your permission, you can follow the process outlined here https://podcastplayer.com/legal.

12d ago 23:26

MP3•Episode home

Send us a text

AI Pathology & Genomics: A New Benchmark for Predicting Gene Mutations

If you still think visual quantification is “good enough” in pathology, think again.
In this 27th episode of DigiPath Digest, I break down four transformative abstracts that show how AI is shifting our diagnostic landscape—from breast cancer segmentation to fibrosis assessment, and all the way to spatial immunology and the evolving immunoscore.

If you’re still relying on manual scoring, static staging systems, or single-marker immunohistochemistry, this episode will challenge you to look deeper—literally and algorithmically.

🔬 Episode Highlights & Timestamps

[02:00] Abstract 1 – AI + IHC for epithelial cell segmentation in breast cancer
[07:30] Abstract 2 – Deep learning quantifies TILs in esophageal cancer
[14:30] Abstract 3 – Biopsy size impacts SHGTPF-based liver fibrosis staging
[22:30] Abstract 4 – Immunoscore in colorectal cancer: promise & limits

🧬 Key Insights & Takeaways

1. IHC-Guided Segmentation for Breast Cancer
Using immunohistochemistry as a ground truth for AI segmentation reveals how effective our models can be—but also where they fall short. The challenge? Accurately subclassifying benign, in situ, and invasive epithelial cells. Spoiler: We’re not quite there yet.

2. Tumor-Infiltrating Lymphocytes in Esophageal SCC
A Chinese team trained deep learning algorithms to analyze TILs spatially. Result? High TIL counts in both intra- and peritumoral zones correlated with better survival—highlighting the emerging power of spatial immunology.

3. Liver Fibrosis Staging with SHGTPF Microscopy
Second harmonic generation two-photon microscopy gives us label-free imaging of unstained tissue. The takeaway: bigger biopsies (20–26mm) yield better fibrosis quantification. Biopsy position? Surprisingly irrelevant. A game-changer for MASLD diagnostics.

4. Immunoscore for Colorectal Cancer
This image analysis-based tool outperforms traditional TNM staging, helping stratify patients for immunotherapy. But adoption is hampered by cost and digital slide access. Integrating AI could take it to the next level—something we should all watch closely.

🎓 Resources from This Episode

Breast cancer segmentation using IHC-guided AI (Trondheim, Norway)
Esophageal SCC & spatial TILs (Cancer Medicine, China)
SHGTPF microscopy in liver fibrosis (UK/US multi-center study)
Immunoscore in colorectal cancer (Jerome Galon group origins)

💡 Bonus: I show off some histology-inspired earrings and talk about the story behind them—multinucleated giant cells, cartilage, and more. Check them out if you’re into pathology fashion!

We’re not just validating AI anymore—we're redefining diagnostics. From high-res, label-free imaging to robust spatial biology insights, the path forward in pathology is clearer and more precise than ever. Whether you’re a practicing pathologist, researcher, or innovator, this episode offers tools and perspectives you can apply today.

Support the show

Become a Digital Pathology Trailblazer get the "Digital Pathology 101" FREE E-book and join us!

160 episodes

#Science #Tech #Natural Sciences #Aleksandra Zuraw, DVM, PhD #Aleksandra Zuraw #DVM #Pathology #Digital #Image #Analysis #Artificial #Intelligence #Pharma #Pharmaceutical #Healthcare #Veterinary #Clinical #Fitness #Medicine