Compugen (NASDAQ:CGEN) President and CEO Eran Ophir discussed the company’s partnered and wholly owned immuno-oncology programs in a prerecorded conference session hosted by Leerink Partners’ Daina Graybosch. The conversation focused on AstraZeneca’s rilvegostomig (a PD-1/TIGIT bispecific), Compugen’s wholly owned PVRIG antagonist COM701 in ovarian cancer, and an IL-18 binding protein (IL-18BP) program partnered with Gilead.
Rilvegostomig: AstraZeneca’s phase III program and trial design
Ophir said AstraZeneca is running 11 phase III trials with rilvegostomig and emphasized he was not speaking on behalf of AstraZeneca. He described a portion of the program as using rilvegostomig as an immuno-oncology “backbone” in combinations where the contribution of TIGIT blockade may not need to be explicitly demonstrated. Examples he cited included:
- TROPION-Lung10 in non-small cell lung cancer, combining rilvegostomig with a TROP-2 antibody-drug conjugate (ADC) and comparing to pembrolizumab.
- ARTEMIS-001, combining rilvegostomig plus chemotherapy compared to chemotherapy.
- Other studies combining with potent agents such as Enhertu, where efficacy may be driven substantially by the partner drug.
Why Compugen still sees potential in TIGIT—format, combinations, and patient selection
Graybosch pressed on why investors should still have confidence in a TIGIT approach given high-profile phase III setbacks across the sector. Ophir argued rilvegostomig differs from prior TIGIT efforts in three main ways: antibody format, combination strategy, and trial design.
On format, Ophir said AstraZeneca has published data supporting cooperative binding enabled by the bispecific design. He explained that the TIGIT-binding arm is based on Compugen’s high-affinity TIGIT antibody (COM902), paired with a lower-affinity PD-1 arm. According to Ophir, the design allows the antibody to “dock” first on TIGIT and then co-block PD-1 and TIGIT on the same effector cells.
He also pointed to the importance of an Fc-reduced format, saying Fc-reduced TIGIT approaches may avoid depletion of effector cells and improve safety and combinability. He referenced AstraZeneca’s reported ex vivo work in tumor fragment systems using patient material, where rilvegostomig showed greater activity than PD-1 alone and greater activity than a PD-1 plus TIGIT combination in that system. Ophir added that AstraZeneca also showed that making the bispecific Fc-active reduced activity in that setting.
On combinations, Ophir highlighted AstraZeneca’s focus on novel pairings, particularly with ADCs. He mentioned a recently added phase III study combining rilvegostomig with a claudin 18.2 ADC in a large trial he described as enrolling more than 2,000 patients, with stratification by PD-L1 and claudin 18.2 status and comparisons versus standards of care such as nivolumab plus chemotherapy.
On trial design, Ophir said AstraZeneca appeared to incorporate lessons from earlier TIGIT trials by focusing on more homogeneous populations (for example, separating squamous and non-squamous lung cancer cohorts), selecting based on PD-L1 in some settings, and powering studies with an appreciation for TIGIT’s potential magnitude of effect.
Lung cancer and TROP-2 biomarker strategy
Ophir provided a quick overview of AstraZeneca’s rilvegostomig lung cancer program, including phase III trials in PD-L1-positive disease combined with chemotherapy (separate squamous and non-squamous studies), a monotherapy comparison versus pembrolizumab in PD-L1-high patients, and the TROPION-Lung10 combination study in PD-L1-high disease.
He also discussed AstraZeneca’s QCS biomarker strategy for its TROP-2 ADC, describing it as an AI-based digital pathology approach that evaluates TROP-2 localization patterns (membrane versus cytoplasm) to infer internalization capacity. Ophir said retrospective analyses in TROPION-Lung01 suggested better efficacy in biomarker-positive patients compared with the all-comer population. He noted that, as discussed in the session, TROPION-Lung10 is the study specifically integrating the QCS biomarker with rilvegostomig.
On timing, Ophir said AstraZeneca has indicated that phase III readouts are expected after 2027, while non–phase III updates could emerge during 2026 and 2027. He cautioned against trying to predict which phase III trial would read out first.
COM701 in MAIA: a randomized maintenance approach in platinum-sensitive ovarian cancer
Turning to Compugen’s wholly owned program, Ophir outlined the company’s MAIA ovarian study of COM701, a PVRIG antagonist, in platinum-sensitive ovarian cancer in the second or third line. He described the trial as randomized and blinded with an initial design of 40 patients receiving COM701 monotherapy and 20 receiving placebo. He said the goal of the first arm is to demonstrate monotherapy activity before considering additional arms or expansion.
Ophir said the study excludes patients with liver metastases. He connected this to retrospective observations from Compugen’s earlier phase I work in heavily pretreated, platinum-resistant ovarian cancer patients, where excluding patients with liver metastases yielded what he described as “almost 40%” clinical benefit. He said MAIA moves earlier in the disease course to patients with less compromised immune systems and lower tumor burden, positioning COM701 as a potential maintenance therapy.
He stated the primary efficacy focus is progression-free survival (PFS), and suggested that if placebo PFS is around four to five months, a three-month improvement would be “very clinically meaningful,” while smaller effects would need to be interpreted in the context of hazard distribution and confidence intervals. Ophir said the trial is intended to provide a clear signal of monotherapy activity rather than an immediate registrational p-value.
Gilead-partnered IL-18BP program: early clinical progress
Ophir also discussed Compugen’s IL-18 binding protein program licensed to Gilead. He said the first patient was dosed in early 2025 in a dose-escalation study evaluating monotherapy and combination with Gilead’s PD-1 antibody zimberelimab, structured with dose escalation/backfill cohorts and subsequent expansion. While he said enrollment is progressing, he indicated Compugen is limited in what it can disclose and did not provide timing guidance for data.
On mechanism, Ophir said Compugen’s computational platform identified that IL-18 can accumulate in tumors in an inactive form due to high IL-18BP levels that block its activity. He described the program’s antibody approach as “inhibit the inhibitor,” aiming to enable IL-18 activity primarily in the tumor microenvironment rather than the periphery, potentially avoiding the therapeutic window issues associated with administering engineered cytokines.
Finally, Ophir briefly described Compugen’s computational discovery engine, Unigen, as built on years of tumor microenvironment data across multiple modalities (including proteomics, spatial methods, and transcriptomics), with AI used to query and operationalize the dataset. He said the company remains focused on oncology targets and new mechanisms and formats, but did not give guidance on when new programs might be announced.
About Compugen (NASDAQ:CGEN)
Compugen Ltd. (NASDAQ: CGEN) is a clinical-stage therapeutic discovery company that leverages proprietary computational discovery platforms to identify novel immuno-oncology targets and biomarkers. The company combines large-scale biological datasets with machine learning algorithms to generate and validate new therapeutic and diagnostic candidates. Founded in 1993 and headquartered in Tel Aviv, Israel, Compugen also maintains a presence in the United States to support its clinical development and commercial collaborations.
Compugen’s predictive discovery engine scans complex biological systems in silico to reveal previously unrecognized pathways and immune checkpoints involved in cancer progression.
