Beyond Animal Testing: How AI and Organoids Are Reshaping Drug Discovery

As the predictive power of traditional animal studies faces increasing scrutiny, and as new FDA policies elevate organoids and organ-on-a-chip technologies to strategic prominence, a fundamental transformation in the underlying logic of new drug development is quietly underway. Recently, on the Pharmaphorum live broadcast, Wang Yunfang, Director of the Center for Clinical Translational Science at Beijing Tsinghua Changgeng Hospital, Tsinghua University; Zhuo Shu, CTO of SIGNET; Hu Ruilian, Chairman and CEO of ChemPartner; and Xu Yuzhou, Senior Director at Shanghai ChemPartner, engaged in an in-depth dialogue on how “AI + Organoids” are reshaping drug discovery and clinical translation pathways.

"Organoid technology did not become an overnight sensation; it is the inevitable outcome of the convergence between stem cell biology and the demands of drug development," stated Professor Wang Yunfang, getting straight to the point.

She pointed out that animal experimentation has served as the preclinical "gold standard" for 80 years, yet the gap between animal models and human responses is becoming increasingly pronounced—particularly regarding metabolic toxicity. Many animals lack human-specific metabolic enzymes, leading to the failure of numerous drugs in late-stage clinical trials or even after market launch due to hepatotoxicity.

The NAMS (Non-Animal Models) Act, released by the FDA in April 2025, elevated organoids and organ-on-a-chip technologies to a strategic level in regulatory science for the first time. Existing cases have demonstrated thatIND Approval Granted Solely on Organ-on-a-Chip Data, Without Animal Testing. This sends a strong signal: organoids have evolved from research tools into regulatory-approved “new benchmarks.”

However, Professor Wang emphasized that this does not mean a complete replacement of animal models—organoids address the issue of “human relevance,” while animal models address the issue of “systemic integration”; the two are complementary rather than opposing.

Professor Wang Yunfang’s team, leveraging the clinical resources of Tsinghua Changgung Hospital, has established a leading domestic biobank of hepatobiliary organoids. Its core philosophy is “beginning with the end in mind”—designing the application pathway for organoids in reverse, with the ultimate benefit to patients as the endpoint.

“We serve not only pharmaceutical companies but, more importantly, patients.” She explained that the company has established a “full-lifecycle” companion system, spanning from surgical tissue acquisition and organoid preparation to biobanking, and further linking organoid data with clinical diagnosis, treatment, and follow-up information. In the event of disease recurrence, reviving the patient’s autologous organoids provides an optimal screening platform for identifying the most suitable new drugs or therapies.

Furthermore, the team is “digitizing” physical organoids to construct virtual organoids, achieving the integration of carbon-based and silicon-based systems. “We aim for organoids to serve as a benchmark—visualizable, quantifiable, comparable, and referenceable.” This requires deep AI intervention and collaborative industry efforts to establish standards and break down data silos in organoid research.

Dr. Zhuo Shu, CTO of SIGNET, shared a distinct R&D pathway—starting with the construction of organoid models for rare tumors to reverse-engineer the discovery of novel targets, and then leveraging AI to accelerate drug design and iteration.

“Our first pipeline candidate is a targeted therapy for diffuse gastric cancer, which has advanced to the late stage of Phase I clinical trials. Although the FAK target has been previously studied, no one has ever established a strong association between it and diffuse gastric cancer,” said Dr. Zhuo Shu. The team constructed the world’s first organoid model of diffuse gastric cancer, recapitulating the gold-standard characteristics of stromal cells in vitro, thereby enabling in-depth mechanistic studies and target identification.

The foundational logic of SIGNET is the triadic closed-loop system of “AI-driven design – automated synthesis – organoid-based evaluation”:

• Early stage: Leveraging AI for drug scaffold design and crystal form prediction, combined with rapid screening using organoids to avoid the misleading results associated with traditional 2D cell cultures;

• Mid-term: Employ gene-edited organoids (high consistency) for iterative screening of candidate compounds, followed by final validation using patient-derived organoids (high heterogeneity);

• Late stage: Leveraging clinical big data and AI analytics to precisely identify susceptible populations and compress clinical trial timelines.

“We are at the inflection point of a fundamental paradigm shift in biomedicine,” said Dr. Zhuo Shu. He believes that the standardized implementation of organoids is the biggest bottleneck to industrialization, which happens to be the core competitive advantage his team has built over the past five years.

Hu Ruilian, Chairman of WisePharm, recalled that the company had already recognized the potential of organoids as early as ten years ago when investing in 3D bioreactors in the United States. Since 2021, the team has systematically evaluated organoid technology groups both domestically and internationally, ultimately choosing to engage in in-depth collaboration with Professor Wang Yunfang’s team at Beijing Tsinghua Changgung Hospital.

“Previously, clients were hesitant and indifferent toward AI and organoids, but inquiries have surged from last year to this year.” Hu Ruilian observed that clients are no longer satisfied with fragmented, single-item services; instead, they are proactively requesting the inclusion of organoid modules within “turnkey” R&D service packages to enhance PCC quality, IND approval success rates, and subsequent clinical translation efficiency.

Xu Yuzhou, Senior Director at Shanghai ChemPartner, further elaborated: The core value of organoids lies in“human relevance”In the field of non-oncology diseases (autoimmune, metabolic, and CNS), there is a significant genetic disparity between animal models and humans, often leading to substantial translational gaps in clinical applications. Organoids effectively bridge the gap between cell-based assays and animal studies, offering particular value in predicting hepatotoxicity, cardiotoxicity, and other adverse effects.

Xu Yuzhou revealed that Ruizhi is engaging in comprehensive collaboration with upstream (culture media, Matrigel), midstream (chips, 3D printing), and downstream (developers of various organoid models) segments of the industry chain to construct a standardized and implementable evaluation system for organoids. “What we aim to do is to begin with the end in mind, reverse-engineering model development from the genuine needs of the industrial sector, rather than developing organoids for their own sake.”

As the FDA sounds the policy clarion call, as the Center for Drug Evaluation (CDE) begins to lay out methodological validation frameworks for organoids, and as the closed-loop iteration between AI and organoids successfully establishes its first clinical pipeline, this transformation is no longer a multiple-choice question of “whether to do it,” but an imperative question of “how to do it faster, more standardized, and more effectively.”

From Professor Wang Yunfang’s “full-lifecycle companion diagnostics and treatment,” to SIGNET’s “AI + organoid closed-loop drug development,” to ChemPartner’s “comprehensive empowerment”—Chinese forces are defining their own rules on this new track.

Say goodbye to blind trend-following and embrace fundamental innovation. The integration of organoids and AI is not merely a tool upgrade, but a revolution in the “truth-seeking” paradigm of drug development. It enables new drugs to be validated at earlier stages, using models that more closely mimic human physiology, and at lower costs, ultimately making the promise of “bringing hope to life sooner” a reality rather than just a slogan.