The FDA's April 2025 roadmap for reducing animal testing in preclinical safety studies and its March 2026 draft guidance on validating new approach methodologies (NAMs) have made human-relevant systems such as patient-derived organoids a regulatory expectation in oncology drug development. Most drug developers cannot yet execute organoid-based testing at the pace regulators now require. Organoid models vary from laboratory to laboratory, and protocols remain largely unstandardized.

Seahee Kim, PhD
Director of Organoids Technology
Samsung Biologics
Few drug development programs can connect a drug's organoid response to actual patient outcomes for the same treatment. Seahee Kim leads the organoids technology team at Samsung Biologics and helped launch its oncology screening service, Samsung Organoids. Through her work at Samsung Biologics, Kim offers a perspective on how the company’s Samsung Organoids platform is helping address the gap between regulatory ambitions and the practical challenges of organoid screening.
Drug Discovery News spoke with Kim about where organoid screening data break down, what evaluation criteria the field overlooks, and what organoids need to become reliable decision-making tools.
What is the gap between the FDA’s push away from animal models and what developers can execute today?
A significant gap exists between regulatory expectations for organoid testing and what most drug developers can realistically implement today.
Organoid and organ-on-a-chip technologies are advancing rapidly, but challenges remain in model availability, reproducibility, and clinical validation. Many drug developers also lack the specialized expertise and infrastructure needed to implement these technologies internally.
Bridging this gap requires standardized testing platforms and reliable, clinically relevant data. Specialized contract research organizations (CROs) can provide both for developers who lack in-house organoid expertise, allowing them to adopt human-relevant testing approaches more efficiently.
Beyond 2D culture, what makes organoid data credible rather than just biologically interesting?
Traditional 2D cell cultures cannot reproduce the three-dimensional structure and complex biological characteristics of human tumors. In addition, long-established cell lines can also lose some of the genetic diversity and characteristics of the original patient tumors over time.
In contrast, patient-derived organoids can better preserve the three-dimensional architecture, genetic features, cell-to-cell interactions, and tumor heterogeneity of the original tumors. Those structural and molecular differences can influence drug penetration and treatment responses, making organoids potentially more predictive of how patients respond to therapies.
Where does integrating high-content screening and multi-omics data actually break down?
The biggest challenge in integrating high-content screening and multi-omics data is not simply combining different data formats but ensuring that both datasets accurately reflect the same biological state.
High-content screening provides image-based information on drug responses—changes in organoid morphology, growth inhibition, and cell death—while multi-omics analysis reveals molecular characteristics, including genetic alterations and gene expression patterns.
However, differences in culture conditions, sampling time points, and drug treatment conditions can make it difficult to establish meaningful correlations between these datasets. In addition, imaging and omics data require different quality control standards and analytical approaches, making standardized workflows essential.
Ultimately, successful integration is not just about combining datasets. It is about connecting phenotypic and molecular information through standardized experimental conditions and robust quality controls to generate reliable and biologically meaningful insights.
When a drug program runs organoid screening through a vendor without GxP-aligned quality systems, where the data typically break down?
Without a robust quality management system, the biggest risk in organoid screening is inconsistent, unreproducible data.
Organoid responses can be influenced by culture conditions, passage number, cell quality, and drug treatment procedures. Without standardized protocols and appropriate quality controls, the same drug can produce inconsistent results across experiments.
That variability goes beyond experimental noise. It can steer critical drug development decisions in the wrong direction, potentially leading to programs dropping promising candidates or advancing less effective ones. Additional validation and repeated experiments to correct for unreliable data also add significant delays and development costs.
What are R&D directors missing when they evaluate organoid CROs?
Tumor type coverage and screening throughput are important metrics when evaluating organoid CROs. However, data quality and clinical relevance matter more than the model volume or screening speed.
- Model specificity: It is essential to demonstrate that organoids retain the key genetic characteristics and biomarker profiles of the original patient tumors. Beyond tumor type diversity, having models with specific mutations and therapeutic targets is also critical.
- Data quality and reproducibility: Standardized culture conditions, validated assays, and robust quality management systems are necessary to ensure consistent and reliable results.
- Clinical data integration: The ability to integrate organoid screening results with clinical data is a key differentiator. Linking drug responses to patients' treatment histories and clinical outcomes can provide more meaningful insights for drug development decisions.
Where does the line sit between organoid data that can support an IND application and data that regulators still treat as supplementary?
The distinction between organoid data that can support an IND application and data that regulators treat as supplementary depends on data quality, validation, and intended regulatory use.
The FDA's March 2026 draft guidance outlined validation principles for NAMs. Additional guidance issued in May 2026 introduced approaches to reducing or replacing certain animal toxicity studies for oncology drugs.
Organoid data can provide valuable evidence on drug mechanisms, antitumor efficacy, biomarkers, and human-relevant toxicity. Replacing specific animal studies, however, requires fit-for-purpose validation and sufficient scientific justification, often developed through early discussions with regulatory authorities. Ultimately, the goal is not simply to include organoid data in an IND submission, but to generate scientifically robust, reliable, and validated evidence that regulators can confidently consider.
After a year of client engagements, what must organoid services still solve?
Over the past year, client engagements have revealed strong interest in organoid technologies alongside persistent barriers to broader adoption. First, the field needs greater diversity in organoid models. Clients are looking not only for different tumor types, but also for patient-derived models with specific genetic mutations, target expression levels, and drug resistance profiles.
Second, assay capabilities need to expand. As drug modalities such as antibody-drug conjugates, bispecific antibodies, and immunotherapies become more complex, demand is growing for organoid assays that go beyond simple cell viability measurements to evaluate mechanisms of action and interactions with immune cells.
Finally, clinical validation remains the field’s essential unfinished task. The industry must demonstrate how well organoid drug responses correlate with actual patient outcomes and continue building robust, clinically linked datasets. Until that evidence exists at scale, organoid services will remain a promising approach rather than a source of reliable decision-making tools across drug development.








