Compugen (CGEN) Leerink’s Global Healthcare Conference 2025 summary

Strategic partnerships and pipeline overview

• Focus on immuno-oncology with a computational platform (Unigen) for novel target discovery, leading to assets like COM701 (PVRIG), COM902 (TIGIT), and COM503 (IL-18BP), with key licensing deals to AstraZeneca and Gilead.

• AstraZeneca's rilvegostomig (TIGIT/PD-1 bispecific) is advancing aggressively in multiple phase 3 trials, especially in non-small cell lung and gastric cancers, with promising early efficacy and safety data.

• Biomarker-driven patient enrichment strategies, including AI-developed TROP2 QCS, are being used in trials, though prospective validation remains ongoing.

• COM902 benefits from validation through rilvegostomig's progress and is positioned for combinations or as a potential out-licensing asset.

• The discovery platform has repeatedly identified first-in-class targets, differentiating it from other AI-driven approaches.

Clinical development and trial design insights

• COM701 targets PVRIG, showing activity in platinum-resistant ovarian cancer, with plans to move into platinum-sensitive maintenance settings due to unique biology and clinical signals.

• Early biomarker data suggest PVRL2 expression may enrich for responders, but prospective selection is not yet feasible; ongoing translational efforts continue.

• The platinum-sensitive ovarian trial uses an adaptive design with PFS as the primary endpoint, aiming for early decision-making and flexibility for expansion or combination strategies.

• Maintenance therapy is seen as a rational setting for checkpoint inhibitors, with the identified patient population representing a significant unmet need.

• Competitive landscape includes other IO agents in ovarian cancer, but strategic focus is on earlier lines and maintenance rather than direct competition.

Scientific and mechanistic perspectives

• Rilvegostomig's combination with ADCs is expected to show synergistic effects, with initial data anticipated this year.

• Fc engineering differentiates TIGIT antibodies; non-active Fc versions (like COM902 and rilvegostomig) may offer better safety and efficacy profiles compared to active Fc competitors.

• COM503 (anti-IL-18BP) leverages computational discovery to enable localized cytokine activation in the tumor microenvironment, with phase 1 trials underway and broad potential indications.

• Ongoing efforts include deep translational work, such as paired biopsies and computational biomarker discovery, to inform future patient selection and mechanistic understanding.

• Differentiation of COM902 from other TIGIT antibodies is based on binding affinity and Fc characteristics, though clinical impact remains to be seen.