Daiichi Sankyo Taps Imagene AI for Biomarker Discovery

US-based Imagene AI has formed a collaboration with Daiichi Sankyo to advance multimodal biomarker discovery and response prediction across oncology programs. The partnership, announced on April 9, 2026, will apply Imagene’s multimodal platform, OI Suite, to support biomarker-driven decision making from early discovery through clinical development.¹

The deal comes at a time when multimodal AI is increasingly shaping how oncology drugs are developed. In the current landscape, biomarker discovery and patient stratification are emerging as critical levers for improving clinical success rates, especially in complex modalities like antibody-drug conjugates (ADCs).²

How can multimodal AI improve biomarker discovery in oncology trials?

The collaboration focuses on integrating histopathology data, including hematoxylin and eosin and immunohistochemistry (IHC) images, with molecular profiles and longitudinal clinical outcomes. Using its CanvOI foundation model and a large-scale real-world data lake, Imagene aims to generate biologically grounded insights that can inform biomarker hypotheses earlier in development.

This approach reflects an industry trend toward multimodal data integration, in which combining imaging, omics, and clinical datasets enables more robust identification of predictive biomarkers. In practice, such an approach can help refine patient selection criteria and reduce variability in clinical trial outcomes, which remain two persistent challenges in oncology drug development.³

“Collaborating with Daiichi Sankyo reflects a shared commitment to advancing biomarker discovery as a key driver of development success,” said Dean Bitan, co-founder and CEO of Imagene AI, in a company press release.¹ “By working together, we are integrating multimodal discovery and quantitative IHC scoring to move from biomarker hypothesis to patient stratification with greater confidence.”

Why is biomarker-driven patient stratification critical for ADC development?

The collaboration will focus in part on supporting select ADC programs from Daiichi Sankyo, for which precise patient selection is essential to maximize therapeutic benefit while minimizing toxicity, according to the companies. ADCs rely on target expression levels and tumor biology, making accurate biomarker identification a key determinant of clinical success.⁴

Under the agreement, Imagene will deploy AI-driven pipelines to identify biomarkers correlated with treatment response, map associated biological pathways, and evaluate histologic features. A central component is its composite continuous scoring system, which quantifies IHC-based target expression using a continuous scale rather than traditional categorical thresholds.

This quantitative approach could improve how patients are matched to therapies by capturing more nuanced variations in target expression, potentially leading to better response prediction and more efficient trial design.

What does this collaboration signal for AI adoption in clinical development?

The partnership highlights growing investment in AI-enabled infrastructure to address data fragmentation and complexity in oncology research.⁵ Imagene’s platform is supported by a real-world data lake of more than 3.5 million tissue samples, integrated with omics and clinical outcomes data, enabling model development across diverse patient populations.

For Daiichi Sankyo, the collaboration aligns with a broader strategy to differentiate oncology assets, particularly in the ADC space, through biomarker-driven development. As pipelines become more crowded and trial costs rise, the ability to identify responsive patient populations earlier may offer a competitive advantage.

The deal also reflects a shift from exploratory biomarker research to operational integration of AI tools within clinical development workflows. By embedding multimodal analytics into trial design and execution, companies aim to improve probability of success while accelerating timelines.²

As oncology pipelines grow more complex, collaborations such as this suggest that data integration and predictive modeling will play an increasingly central role in translating scientific insights into clinically meaningful outcomes.⁵

References

1. Imagene AI. Imagene AI announces collaboration with Daiichi Sankyo to advance multimodal biomarker discovery in oncology. Published April 9, 2026. Accessed April 10, 2026. https://imagene-ai.com/press-release/imagene-ai-announces-collaboration-with-daiichi-sankyo-to-advance-multimodal-biomarker-discovery-in-oncology/
2. Lipkova J, Chen RJ, Chen B, et al. Artificial intelligence for multimodal data integration in oncology. Cancer Cell. 2022;40(10):1095-1110. doi: 10.1016/j.ccell.2022.09.012
3. Vanguri RS, Luo J, Aukerman AT, et al. Multimodal integration of radiology, pathology and genomics for prediction of response to PD-(L)1 blockade in patients with non-small cell lung cancer. Nat. Cancer 2022;3:1151–1164. doi: 10.1038/s43018-022-00416-8
4. Ascione L, Guidi L, Prakash A, et al. Unlocking the potential: Biomarkers of response to antibody-drug conjugates. Am Soc Clin Oncol Educ Book 2024;44:e431766. doi:10.1200/EDBK_431766
5. García-Lezana T, Bobowicz M, Frid S, et al. New implementation of data standards for AI in oncology: Experience from the EuCanImage project. GigaScience 2025;14:giae101. doi: 10.1093/gigascience/giae101