Shoppers of biotech headlines are spotting a fresh partnership: Waiv, the Paris-born AI pathology specialist, has teamed up with Daiichi Sankyo to hunt for digital biomarkers that could predict patient response to antibody‑drug conjugates , a move that matters because ADCs need clever screening when trial sizes are tiny.
Essential takeaways
- – Who’s involved: Waiv (formerly Owkin Dx) is partnering with Daiichi Sankyo to deliver AI-derived biomarkers for an ADC programme.
- – What it does: Waiv will apply computational pathology across H&E and IHC slides to identify treatment‑response signals and predict outcomes.
- – Why it’s notable: Their platform is built for low‑data settings, aiming to discover biomarkers even with fewer than 100 patients.
- – How it works: The approach blends bespoke AI models, whole‑slide image analysis, and interpretable outputs suitable for downstream clinical validation.
- – Feel of the tech: The work is data-driven but practical , reproducible, interpretable results rather than a black‑box claim.
Why this partnership matters now: ADCs need smarter biomarkers
Antibody‑drug conjugates are elegant but complicated , they carry toxic payloads and benefit only a subset of patients, so knowing who will respond is crucial. According to reporting in PharmaTech and industry press, drugmakers are increasingly pairing ADC programmes with sophisticated biomarker hunts. Waiv’s collaboration with Daiichi Sankyo plugs straight into that problem, promising a faster route to predictive tests using routine pathology slides. The idea of pulling meaningful signals from H&E and IHC images feels almost tactile: you can picture pathologists and algorithms reading the same slide, but for different insights.
Built for small trials: how Waiv tackles low‑data challenges
Most AI models crave vast datasets, yet early ADC trials can include fewer than 100 patients. Waiv says its foundation models were trained across hundreds of thousands of images from a global network, then fine‑tuned to perform in data‑constrained scenarios. That’s the practical benefit here , bespoke models optimised for small cohorts can extract predictive signals where standard approaches struggle. For teams planning early‑phase studies, this reduces the trade‑off between speed and confidence: you can explore biomarkers earlier without waiting for huge sample sizes.
From slides to clinical decision support: interpretability matters
One of the clearest industry complaints about AI in healthcare is opacity. Waiv stresses reproducible and interpretable outputs designed for clinical decision‑making, which is what separates a research signal from a deployable companion diagnostic. In practice that means algorithms that highlight spatial tumour microenvironment features on whole‑slide images, explain why a prediction was made, and produce metrics clinicians recognise. For sponsors and regulators, that narrative , transparent, testable, and validated , is what makes a partnership like this attractive.
How this fits the wider trend: pharma leans on AI and imaging
Big pharma has been courting imaging and AI partners across oncology pipelines, from Tempus collaborations with Daiichi Sankyo to other diagnostic alliances reported in industry outlets. The pattern is clear: companies are investing in digital pathology to de‑risk development and refine patient selection. Waiv’s global data network of hospitals, labs and academic centres also mirrors a broader move towards federated or collaborative data approaches that respect local governance while improving model robustness.
What sponsors and clinicians should look for next
If you’re following ADC development or running early trials, monitor three things: whether the biomarkers discovered by Waiv are validated prospectively, how the tests perform across different staining protocols and scanners, and whether outputs are accepted by regulators or adopted as companion diagnostics. Practically, teams should standardise slide preparation, consider multi‑site reproducibility checks, and ask partners for interpretability metrics so clinicians can trust the signal.
It’s a small change with big upside: better biomarkers could mean fewer failed trials, quicker approvals, and patients getting the right treatments sooner.
Source Reference Map
Story idea inspired by: [1]
Sources by paragraph:
– Paragraph 1: [2], [3]
– Paragraph 2: [5], [7]
– Paragraph 3: [2], [3]
– Paragraph 4: [4], [5]
– Paragraph 5: [6], [7]
Supercharge Your Content Strategy
Feel free to test this content on your social media sites to see whether it works for your community.
