AI May Be the Missing Link in ADC Cancer Precision

AI-guided tissue analysis is sharpening patient selection, cutting trial risk, and speeding more precise ADC cancer drug development

5 Nov 2025

A small change in how tumours are read may have large effects on how cancer drugs are made. Across America’s oncology industry, artificial intelligence is moving from a back-office tool to a core part of antibody drug conjugate (ADC) development. The promise is not flashier science, but fewer mistakes, better choices about which drugs to build and which patients to test them on.

ADCs are often described as smart bombs, antibodies that ferry toxic payloads directly to cancer cells. They are powerful, expensive and risky. Many fail late in development, when costs are highest. That is why Pfizer’s $43bn purchase of Seagen in 2023 mattered. It was not just a bet on Seagen’s drugs, but on its data and know-how. The deal signalled that future progress in oncology would depend as much on analysing information as on inventing molecules.

AI’s most immediate impact is in tissue analysis and biomarker discovery. Traditional pathology tends to ask a simple question, does a tumour express a given target? AI-assisted pathology asks harder ones. Where is that target located? How evenly is it spread? What surrounds it? By scanning high-resolution digital slides, algorithms can detect spatial patterns inside tumours, relationships between cancer cells, immune cells and supporting tissue, that human eyes often miss. These patterns can influence whether an ADC reaches its target, or fails before it gets there.

Firms such as Nucleai are building tools to map these microscopic neighbourhoods. By modelling how cells interact within the tumour microenvironment, they aim to help drugmakers decide earlier which patients are most likely to benefit. That matters because the biggest losses in oncology come not from weak science, but from testing the right drug in the wrong group. Improving selection upstream can save both money and time.

The shift reflects a broader change in cancer research. Genetics still matter, but they are no longer enough. Tumour architecture and context are gaining weight. AI-driven pathology offers a way to measure them systematically, and to design trials around more precise definitions of disease. The technology is still being validated, and questions remain about data quality, explainability and regulation. Regulators will want to know not just what an algorithm predicts, but why.

Even so, the direction is clear. As complex drug platforms meet more sophisticated data tools, ADC development is becoming narrower and more deliberate. For patients, that should mean treatments chosen with greater care. For drugmakers, it suggests that intelligence, artificial and otherwise, is becoming as valuable as chemistry.