TIA Centre: News

European Congress on Digital Pathology, Barcelona (26th - 28th June 2025)

The European Congress on Digital Pathology (ECDP) 2025, held in the vibrant and picturesque city of Barcelona, brought together a dynamic mix of researchers, clinicians, and industry professionals pushing the boundaries of digital and computational pathology. From deep learning and 3D reconstruction to pan-cancer biomarkers and cell-level interpretability, this year's conference showcased not only the technical depth of our field but also the collaborative spirit that drives it forward.

Four members of the TIA Centre community shared their experiences and highlights from the conference:

Mostafa Jahanifar presented on Quantifying Mitotic Topology as a Pan-cancer Prognostic Biomarker Using a Deep-Learning and Network-Based Approach (Friday 27th June, 09:50)

“I have been to the ECDP conference twice before, but my third time this year — en la ciudad bonita, Barcelona de Cataluña (fancy way of saying "in Barcelona") — went beyond my expectations in so many ways.

To start with, it was bigger; bigger in the number of participants, the number of industrial exhibitions, and the number of papers and posters presented. But even more impressive than the quantity was the quality of the work, across all of those areas. I saw presentations and posters focused on large-scale projects using state-of-the-art methods, or even proposing brand-new approaches to long-standing problems.

I also had the chance to present my latest work on "Quantifying mitotic topology and mitotic errors as pan-cancer prognostic biomarkers." In this project, which is part of my PhD, I used deep learning–based models and social network analysis techniques to quantify mitotic activity in tumors across 31 different cancer types.

I was really pleased to see how well my work was received. I got great feedback from enthusiastic pathologists and researchers. And the cherry on top was that my presentation won the "Best Presentation Award"; a big surprise for me, especially considering all the amazing work presented at ECDP.

I enjoyed the experience so much, and was so impressed by the incredible job ESDIP is doing, that I couldn’t help but want to be a part of it. So right after the conference, I asked the organisers if I could get involved in any way. They kindly said yes!

Hopefully, I will be able to contribute to the mission of ESDIP and ECDP in the coming years. I will definitely be back for ECDP in Graz next year; and I hope to see you there!”

Jiaqi Lv presented on Deep Learning for detection of Mononuclear Leukocytes from PAS Stained Whole-Slide-Images of Kidney Biopsy (Thursday 26th June, 15:15)

“It was a fantastic experience attending ECDP in Barcelona this year. I gave a short oral presentation on my work in deep learning for inflammatory cell detection. Throughout the conference, I had the opportunity to engage in insightful conversations with researchers from both academia and industry. It was a great opportunity to share ideas and connect with the wider computational pathology community!”

Neda Zamanitajeddin presented on Pan-Cancer Microsatellite Instability Prediction via Foundation Models and Cell Social Network Analysis Reveals Shared Tumor Microenvironment Dynamics (Saturday 28th June, 10:10)

“I had the opportunity to present our project at ECDP 2025. We developed a pan-cancer approach for predicting microsatellite instability (MSI) using H&E whole-slide images from colorectal, gastric, and endometrial cancers. By combining a foundation model (UNI) with a transformer-based aggregator, we achieved high MSI prediction accuracy. To improve interpretability, we used cell social network analysis (CSNA), which revealed distinct tumor microenvironment (TME) interactions between MSI and MSS cases and uncovered shared molecular patterns across different cancer types. This work highlights the potential of cross-cancer learning and interpretable AI in advancing precision oncology.

Presenting this work at ECDP 2025 was a fantastic experience, an inspiring environment filled with innovative research, insightful discussions, and a great opportunity to connect with the digital pathology community.”

Mark Eastwood presented High Resolution 3D Reconstruction and Visualization of Tubulo-villous Adenoma Tissue (Thursday 26th June, 17:50)

“I had the opportunity to present my work on 'High resolution 3D reconstruction of TVA tissue from sequential WSIs', in the 'Novel Interdisciplinary Approaches' session at this years ECDP, which I thought was a strangely named session - all work in Digital pathology is interdisciplinary, given that it is right at the intersection of the pathology and computer science fields!

I talked about a pipeline we developed (based on opensource tools including DeeperHistoReg, CloudVolume, and NeuroGlancer) for flexible 3D visualisation of tissue samples. Our approach requires only a sequence of adjacent sections taken using standard H&E slide protocols, which we applied to a dataset of 50 sequential sections from a tubulo-villous adenoma. Slides were registered and a volumetric representation of the tissue saved in a 3D pyramidal format, allowing the complex 3D structure of a TVA to be visualised so we can gain a better understanding of the growth patterns of such lesions.

As my talk was on the first day, I could then relax and enjoy the rest of ECDP conference, which was excellent as always with lots of interesting talks, and delicious spanish food and the sights of Barcelona to enjoy in the evenings.”

Looking Ahead

Whether it was the scientific breakthroughs, the spirit of collaboration, or the lively setting of Barcelona, ECDP 2025 left a lasting impression on all who attended. As the field continues to evolve, gatherings like ECDP play a vital role in fostering innovation, community, and shared purpose and the TIA Centre will continue to be involved in future events sharing the latest thinking and taking the opportunity to share our continuing work.

We’re already looking forward to reconnecting in Graz in 2026!

We would like to take this opportunity to thank Big Picture for supporting Mostafa, EPSRC for supporting Jiaqi, the Warwick Chancellor's Scholarship for supporting Neda, and Cancer Research UK for supporting Mark.

Contributions from the TIA Centre:

Mostafa Jahanifar [Oral Presentation, and Best Talk]: Quantifying Mitotic Topology as a Pan-cancer Prognostic Biomarker Using a Deep-Learning and Network-Based Approach.

Jiaqi Lv [Oral Presentation]: Deep Learning for detection of Mononuclear Leukocytes from PAS Stained Whole-Slide-Images of Kidney Biopsy.

Neda Zamanitajeddin [Oral Presentation]: Pan-Cancer Microsatellite Instability Prediction via Foundation Models and Cell Social Network Analysis Reveals Shared Tumor Microenvironment Dynamics.

Mark Eastwood [Oral Presentation]: High Resolution 3D Reconstruction and Visualization of Tubulo-villous Adenoma Tissue.

By Mostafa Jahanifar, Jiaqi Lv, Neda Zamanitajeddin, Mark Eastwood