Teaching AI to see: How the next era of digital pathology is being driven by curated biospecimen data
In this guest editorial, Daryl Waggott, Head of Data Products, BioIVT, highlights the critical role of ethically sourced, clinically linked biospecimen datasets in advancing computational tumor biology.
12 Nov 2025In oncology, the microscope has long served as a window into disease. Today, it’s becoming a gateway to artificial intelligence. As digital pathology helps transform tumor biology research, AI models are only as accurate as the data that trains them. Better data, not more algorithms, will likely drive the field's next leap forward.
The data behind the diagnosis
Whole-slide imaging (WSI) now captures tumor morphology at unprecedented resolutions compared to a decade ago. However, these images alone can’t teach AI what to look for. To make pathology truly computational, imaging data must be connected to the story behind every sample, including the clinical context, molecular profile, and patient history that shape each case.
This is where the concept of biospecimen intelligence begins. By bringing together detailed clinical annotations, biomarker data, and high-resolution pathology slides, researchers can create AI systems capable not only of pattern recognition but also of biological understanding.
From raw pixels to intelligent insights
BioIVT’s Visionaire™ initiative was developed to address this gap. Drawing on over three decades of ethically sourced biospecimens, Visionaire transforms real-world tumor samples into deeply curated AI-ready datasets that pair each digital slide with its originating tissue block, molecular data, and clinical context, all from the same consented donor.
In digital pathology, this means that an image is no longer just a static snapshot but a living biological data object. Researchers can trace cellular patterns back to genomic variants or treatment responses, directly linking what’s seen under the microscope to what’s happening in the tissue itself. This continuity between digital and physical specimens can accelerate biomarker discovery, therapeutic validation, and reproducibility across studies.
A collaborative future for tumor biology
As digital pathology and computational oncology evolve, communication between data providers, AI developers, and clinicians is becoming critical. The best AI models in tumor biology will be created not from separate datasets, but from connected systems that reflect real clinical and biological diversity.
By building datasets that are traceable not only to the clinical record but also to the original consented human tissue itself and aligned with ethical and regulatory standards, BioIVT aims to support a future where every slide contributes to scientific insight and every dataset brings AI one step closer to clinical relevance.
The next frontier
AI is beginning to redefine how we see cancer, not as static tissue but as dynamic biology. With the right data, machines can learn to see what even the trained eye might miss.
And that’s the promise of biospecimen intelligence: transforming years of carefully collected tissue, molecular, and imaging data into the foundation for tomorrow’s discoveries.
Explore BioIVT’s commitment to advancing biospecimen intelligence in oncology and beyond.