Jorge Reis-Filho, Chief of AI for Science Innovation, Enterprise AI Unit, AstraZeneca, shared a post on LinkedIn:
“I am delighted to share a new paper published in Cancer Discovery alongside my colleagues Richard Goodwin, Simon Barry, James Weatherall and Stefan Platz: ‘Enabling AI to drive innovation and precision across Oncology R&D.’
We have presenting these advancements as part of the AACR 2026 agenda.
The paper discusses how at AstraZeneca, we are catalyzing efforts to accelerate and improve decision-making across the entire R&D lifecycle, not merely enhanced by AI, but reimagined through the AI lens. In Q2 2026, this is actively moving from aspiration toward operational practice and, in some areas, already beyond it.
AI is already impacting our work across R&D in multiple ways, in particular through the adoption of domain-specific multimodal foundation models trained on diverse data types including pathology images, omics, EHRs and clinical text. These models, reinforced and fine-tuned with our unique and proprietary data, can generate insights that support early discovery decisions through to downstream clinical development at scale, with greater precision and more rigorous validation than previously possible.
Our AI-enabled computational pathology platform, Quantitative Continuous Scoring (QCS) is a fully supervised AI solution that allows us to quantify target expression at the subcellular compartments of cancer cells across seven classes of interpretable features. It delivers quantitation with a level of accuracy, precision and reproducibility, as well as with a dynamic range that exceeds the levels attained by traditional pathology approaches. We are now compounding technologies: the precision of QCS powered by the capabilities of foundation models, such as Modella.ai UNI2, for the development of multi-cancer/ multi-target QCS solutions.
At the R&D enterprise level, the paradigm shift will materialize through the development of systems, platforms and solutions ‘through‘ AI rather than simply by ‘using AI‘. This means reimagining how questions are asked, how evidence is generated, how to avoid the ‘more-plex fallacy’ of confusing higher dimensionality for greater insight, and how decisions are made across the ecosystem. Germane to the success of these endeavors will be the combination of deep domain knowledge, world-class technical expertise, the right infrastructure (e.g. compute, models and agentic frameworks) and, most importantly, high-quality and sufficiently diverse data for the development and benchmarking of robust AI solutions.
We are beginning to witness substantial progress from individual scientist enablement, to process-enabled AI, to enterprise-wide transformation in which AI-driven research enables the derivation of novel biological insights, concepts and paradigms. This will undoubtedly bring us closer to delivering AstraZeneca’s bold ambition in oncology: to eliminate cancer as a cause of death.”
Title: Enabling AI to Drive Innovation and Precision across Oncology R&D
Authors: Richard Goodwin, Simon Barry, James Weatherall, Stefan Platz, Jorge Reis-Filho
Other articles featuring Jorge Reis-Filho on OncoDaily.