Antibody–drug conjugates (ADCs) are a type of oncology drugs that have been gaining tremendous momentum, but predicting patient benefit remains a challenge. Current biomarker strategies rely only on target expression, which does not capture the full biological context driving response.

ADCs are modular therapeutics. Their activity depends on multiple biological processes, including intracellular trafficking, payload sensitivity, DNA damage response, and tumor context. As a result, patients with similar target expression can show very different outcomes.

The Genialis Supermodel is a biology-first foundation model that maps RNA-seq data into interpretable biomodules representing key biological functions. For ADCs, this enables modeling of mechanism-relevant biology and identification of transcriptomic patterns associated with response.

We developed an RNA-based survival model Genialis adcID to predict real-world response to trastuzumab deruxtecan (T-DXd). It demonstrated predictive performance in a real-world cohort and identified biological programs aligned with ADC activity, including DNA damage response, stress pathways, and hormone signaling.

At AACR in April 2026, we presented an RNA-based survival model built with the Genialis™ Supermodel. The model predicts real-world response to trastuzumab deruxtecan (Enhertu) and captures biological mechanisms underlying ADC sensitivity.