Proscia: From Potential to Pipeline: How Pathology AI Is Rewiring Drug Discovery and Development
The latest from the Proscia blog:
Everyone’s talking about AI in therapeutic R&D, but it’s not always easy to cut through the hype to understand where it’s making a real impact today. One of the clearest signals of progress? Pathology.
In a recent Drug Discovery World webinar, our Chief Strategy Officer Nathan Buchbinder sat down with industry leaders Vipul Baxi (Amgen), Dr. Deirdre De Steno (Owkin), and Dr. Ken Bloom (Nucleai) to explore how AI-driven pathology is actively reshaping drug discovery and development—from biomarker discovery to clinical trial design.
Nathan opened the discussion with a clear message: pathology AI is becoming foundational to how leading pharma and biotech organizations operate.
Key Applications Transforming R&D Today
1. Solving the Biomarker Bottleneck
For decades, biomarkers have faced a critical challenge: reproducibility. As Dr. Ken Bloom noted, only about 1% of biomarkers published in peer-reviewed journals ever make it into the clinic, often due to variability in human interpretation. AI is directly addressing this by enhancing what pathologists can do and, crucially, by seeing what they can’t.
In the complex field of multiplex imaging, AI can distinguish and quantify co-localized targets that are incredibly difficult for the human eye to see, especially with new, complex therapies.
Beyond discovering novel biomarkers, AI is expanding access to known ones. For example, only 43% of eligible breast cancer patients currently receive BRCA testing. Owkin’s work on its BRCAura tool demonstrates how AI can predict mutation status from a routine H&E slide, creating a scalable, low-cost pre-screening tool that alleviates bottlenecks and drives access to precision therapies.
2. Driving Efficiency in Clinical Trials
Perhaps the most significant ROI of AI today is in making clinical trials faster and smarter. Instead of running expensive molecular assays on every sample, teams can now use AI to pre-screen tissue, stratify patient populations, and make better use of existing assets.
This isn’t just theoretical. It’s being used to enroll patients in trials today, proving its real-world value.
3. Shifting from Judgment Call to Quantitative Biomarker
For IHC scoring—a cornerstone of oncology trials—AI is replacing subjective interpretation with standardized, reproducible quantification. This allows teams to capture subtle expression patterns and analyze early trial data to define the most optimal scoring paradigms for future studies.
This move towards quantitative, continuous scoring is foundational for the next generation of AI-based companion diagnostics (CDx), promising to match patients to targeted therapies with unprecedented precision.
The Augmented Pathologist: A New Standard of Care
A recurring theme was that AI is here to augment, not replace, the pathologist. The consensus is clear: the combination of physician and AI delivers the best results, creating a powerful partnership that leverages technology to solve complex challenges that require human expertise for final application.
Looking Ahead: Every Slide as a Learning Asset
The ultimate vision shared by the panel is one where the immense, untapped value of archived tissue is fully realized. The data to fight disease is already present in millions of H&E slides stored in labs worldwide. By creating a feedback loop where every slide from every trial becomes a learning asset, the industry can continuously mine historical data to uncover novel insights, refine biomarker strategies, and accelerate the entire R&D cycle.
As the first AI-based companion diagnostics make their way through regulatory approval, the floodgates are poised to open, making this vision an everyday reality for patients.
For an even closer look at how this shift is playing out, and the evidence behind it, download our new scientific report “Rewiring Pathology for Precision Drug Discovery and Development.”
SOURCE: Proscia
