Patient-derived tumor organoids can reproduce important features of an individual patient’s cancer, offering a more physiologically relevant model for drug screening than conventional two-dimensional cell cultures. Turning that biological complexity into a robust, scalable screening platform, however, presents significant technical and analytical challenges.
Meritxell B. Cutrona is a cell biologist, scientific advisor, and consultant specializing in organoid-based high-content screening. Her work brings together cancer organoids, phenotypic screening, high-content imaging, and laboratory automation to support drug discovery and pharmacotyping. Here, she discusses the predictive potential of organoid models, their application in pancreatic cancer, and what is needed to make complex 3D assays practical at scale.
Olli Dufva, a physician-scientist and postdoctoral fellow at the Wellcome Sanger Institute and Cambridge Stem Cell Institute, provides a complementary perspective on how single-cell, spatial, and multi-omics readouts can capture tumor heterogeneity and narrow the gap between in vitro models and clinical responses.
What makes a disease model or assay more predictive of human biology?
Three-dimensional (3D) self-organizing in vitro tumor models, such as patient-derived tumor organoids (PDTOs) can be initiated from epithelial tissue-resident adult stem cells present in tumor fragments and needle biopsies, as for example from gastrointestinal tract cancers such as colorectal (CRC) and pancreatic ductal adenocarcinoma (PDAC) (1). These 3D-cell cancer models provide a means to replicate the histopathological (e.g., tumor architectural organization, cellular markers expression and spatial distribution), molecular, and functional characteristics of human tumor tissues (2,3). PDTO collections allow to reconstruct the genotypic, epigenetic, and phenotypic landscapes of tumors, which facilitates the capture of the diversity of cancer pathways dysregulation and, in turn, unique disease states on a patient-by-patient basis. The representation of such disease heterogeneity through PDTOs – even if restricted to the epithelial tumoral component– has promoted the use of these cellular systems as dynamic predictive biomarkers of drug responses and for target validation (5-8). Hence, the correspondence between pharmaco-typed drug responses in PDTOs with patient clinical outcomes supports a PDTO-guided prediction of therapy responders, i.e., those patients who are most likely to benefit from the treatments under interrogation. However, the outcomes of ongoing clinical trials within the framework of precision oncology and predictive pharmacology for various types of cancers including PDAC (9,10), will be instrumental to establish the prediction power of functional profiling relying on this PDTO-based interrogation. This information will allow to validate the capabilities of the PDTO cultures to retain the pathways that sustain chemoresistance in addition to the representation of stem cell pathways (11), as promoted by the organoid culture protocols adopted currently.
Where are 3D models, organoids, complex cell systems, or organs-on-chip having the greatest impact in drug discovery?
As mentioned in the previous point, the capacity of pancreatic cancer PDTOs for recapitulating the molecular and functional states of tumors – including diversity of disease states – is critical for using 3D-cell models to investigate the response to chemotherapies and other treatments. So far, chemotherapy-sensitive profiling with pancreatic cancer PDTOs stands as the only means to achieve empirically-based patient stratification. Given the inter-patient heterogeneity in drug response and high recurrent rates of chemoresistance in PDAC, this tumor remains virtually incurable via pharmacological approaches. In this context, large-scale technologies such as high-throughput screening (HTS) can offer new perspectives for identifying effective novel therapies (8). The use of PDTO cohorts in drug-HTS will enable the integration of PDAC tumor heterogeneity profiles at the genomic, epigenetic and metabolic levels. This will play a pivotal role in drug discovery, considering that unified therapeutic options for PDAC will inevitably remain limited.
Recent studies have highlighted dual-and multi-agent in PDTO-based profiling as a means to achieve higher accuracies compared to monotherapy testing. Therefore, systematic multi-agent screening may be an ideal approach to assess the efficacy of drug combinations directly in PDTOs (12) to improve therapy prediction and to anticipate clinically informed regimens potentially eliciting tumor regression or preventing from the appearance of tumor resistance through drug synergy or other mechanisms that could confer benefit via patient-to-patient variability. In this instance, preclinical approaches that can help identify and prioritize new candidate combinations using PDTOs as a platform for drug synergy-HTS discovery urge to be developed (13).
A further incorporation of tissue microenvironment components – such as stroma, immune or endothelial cells – into epithelial PDTOs, using co-culture or organ-on-chip formats combined with low-throughput high-content imaging (HCI) assays, could pave the way for significant impact on drug discovery by enabling the ex vivo representation of pathological tissue (in this case the tumor) in its entirety and, as consequence, the appropriate contextualization for an assessment of on-target toxicities by cell type (5).
How do we balance biological complexity with the need for assays that are scalable, reproducible, and practical for screening? | What needs to happen for more predictive models and assays to reduce late-stage failures and improve confidence in drug discovery decisions?
By definition, microscopy-based HTS applied to classical cell-based assays involves interrogating the effects of large compound libraries (e.g., thousands of drugs) at the phenotypic level (14-16). When applying this approach to examine the response of PDTOs to drugs and chemotherapeutic agents, morphological or visual viability readouts are taken into account (8, 17-20). Such readouts offer more biological insight over measurement of bulk ATP levels via luminescence assays. HCI enables real-time monitoring, spatial and multiparametric resolution or dissection of cellular pathways and mechanism-of-action inference. Therefore, visual parameters – such as cellular markers, viability stains or organoid size and shapes (which can also be detected in brightfield images) – support scalability of the assay while keeping the biological complexity provided by PDTOs.
The principal role of the HTS microscopist – and cell/organoid biologist – is that to find and define an informative visual readout by few parameters, tailor the HCI regime and establish a cost-effective high-content analysis (HCA) pipeline during the assay development phase. All this will ensure the deployment of a streamlined workflow intended to capture hits with confidence, in turn, to limit the generation of superfluous image data to ensure the throughput of the assay with PDTOs. Therefore, implementing advanced approaches such as image-based HTS with PDTOs can be seen as a counter-measure over traditional drug screening assays using immortalized two-dimensional cell cultures because PDTO-based HTS could increase the reliability and confidence in early steps of the drug discovery pipeline. The only means to drive successful drug candidates into late-stage clinical research is through the construction of robust pre-clinical drug screening. Thus, PDTO-based HTS, while integrating more physiologically relevant and predictable tumor models, holds promise to prevent attrition failures in pharma industry.
What role does automation play in making complex models more useful for drug discovery?
From the point of view of PDTO-based drug-HTS, lab automation plays a pivotal role as it constitutes the core technology for drug screening at early stages of the drug discovery workflow according the standards of the pharma industry (21). In the context of organoid-based HCI, the interrogation of the effects of compound libraries relies on the use of optical multi-well plates (8,17-20,22,23). Therefore, all those operations that regard compound handling, seeding of organoid cells into plate wells, drug dispensing, or in situ well processing (e.g., addition of fluorescence dyes, organoid fixation and whole-mount immunofluorescence staining) are supported by lab automation (13,22). The automation of organoid culture with liquid handlers is emerging as a new proposition to enable to scale up organoid production and automate maintenance of complex models, also supporting dynamic tracking of PDTO cultures. Although this newest automation “module” is truly attractive, more experience is needed to evaluate its transformative potential when integrated into drug discovery workflows in industry, core facilities or biotech companies. Clearly, this automated organoid culture module should be kept separate from the liquid handling equipment mentioned above, as the two workflows must operate completely independently.
Meritxell B. Cutrona presented “In-organoid HCS: Perspectives on Cancer Therapy Development and Synergistic Drug Discovery by Harnessing Cohorts of PDOs” at the SLAS Europe 2026 Conference and Exhibition. Her presentation formed part of the “Automating Drug Discovery in 3D Models” session.
Beyond Viability: Reading the Organoid
By Olli Dufva, physician-scientist and postdoctoral fellow at the Wellcome Sanger Institute and Cambridge Stem Cell Institute, University of Cambridge
Ultimately, the predictive value of a disease model or assay will be determined by its ability to accurately identify targets or biomarkers that go on to have clinical impact. As such results are largely unavailable as yet, we have to infer that data derived from models such as organoids – which more faithfully recapitulate the complex cell states and tissue niches found in vivo – will more accurately reflect the biology of real human tissues.
These models are beginning to have a particularly strong impact on understanding the role of tumor heterogeneity in drug responses. This includes differences between patients and genetic and phenotypic variation within a single tumor, along with co-culture systems that capture how stromal and immune cells influence tumor biology and therapeutic response.
However, these more physiologically relevant models also introduce analytical challenges. Complex multicellular models such as organoids require single-cell and spatial readouts to fully extract the unique information they provide on cell-cell interactions and tissue niches.
Newer analytical readouts are helping researchers access that information. Multi-omics, particularly at single-cell resolution, allows us to assess how drugs rewire cell states and reverse disease-associated programs, rather than relying solely on measures such as cytotoxicity or target engagement.
To improve confidence in drug discovery decisions, we need to achieve increased fidelity of in vitro models through systems such as organoids that recapitulate the genetic and epigenetic complexity of patient tissues. Combined with high-content single-cell and spatial profiling, these models should allow more seamless integration with patient-derived data, reducing the gap between in vitro assays and clinical responses.
Olli Dufva presented “Population-Scale Single-Cell Genomics for Precision Blood Diagnostics” at the SLAS Europe 2026 Conference and Exhibition as part of the session “Capturing Heterogeneity: Biomarker Signatures from Complex Data.”
Newsletters
Receive the latest analytical science news, personalities, education, and career development – weekly to your inbox.
References
- Clevers, H. Modeling development and disease with organoids. Cell165, 1586–1597 (2016).
- van de Wetering, M. et al. Prospective derivation of a living organoid biobank of colorectal cancer patients. Cell161, 933–45 (2015).
- Boj, S. F. et al. Organoid models of human and mouse ductal pancreatic cancer. Cell160, 324–338 (2015).
- Wensink, G. E. et al. Patient-derived organoids as a predictive biomarker for treatment response in cancer patients. NPJ Precis. Oncol.5, 30 (2021).
- Verstegen, M. M. A. et al. Clinical applications of human organoids. Nat. Med.31, 409–421 (2025).
- Driehuis, E. et al. Pancreatic cancer organoids recapitulate disease and allow personalized drug screening. Proc. Natl. Acad. Sci. U. S. A.116, 26580–26590 (2019).
- Tiriac, H. et al. Organoid profiling identifies common responders to chemotherapy in pancreatic cancer. Cancer Discov.8, 1112–1129 (2018).
- Cutrona, M. B., Wu, J., Yang, K., Peng, J. & Chen, T. Pancreatic cancer organoid-screening captures personalized sensitivity and chemoresistance suppression upon cytochrome P450 3A5-targeted inhibition. iScience27, 110289 (2024).
- Beutel, A. K. et al. Organoid‐based precision medicine in pancreatic cancer. United Eur. Gastroenterol. J.13, 21–33 (2025).
- Thorel, L. et al. Patient-derived tumor organoids: a new avenue for preclinical research and precision medicine in oncology. Exp. Mol. Med.56, 1531–1551 (2024).
- Seghers, S. et al. Comparative characterization of OncoProTM and Wnt-Based media reveals distinct phenotypic and pharmacologic states in patient-derived tumor organoids. at https://doi.org/10.64898/2025.12.13.693944 (2025).
- Wansch, K. et al. Multi-drug pharmacotyping improves therapy prediction in pancreatic cancer organoids. Cancer Cell Int.25, 321 (2025).
- Cutrona, M. B. A high-content screening workflow for pancreatic cancer therapy development using patient-derived tumor organoid cohorts: An adaptation to enable in organoid synergistic drug discovery. in Methods in Molecular Biology Online ahead of print (Springer, New York, NY, 2026). doi:10.1007/7651_2026_698.
- Giuliano, K. A. et al. High-content screening: A new approach to easing key bottlenecks in the drug discovery process. SLAS Discov.2, 249–259 (1997).
- Boutros, M., Heigwer, F. & Laufer, C. Microscopy-based high-content screening. Cell163, 1314–1325 (2015).
- Way, G. P., Sailem, H., Shave, S., Kasprowicz, R. & Carragher, N. O. Evolution and impact of high content imaging. SLAS Discov.28, 292–305 (2023).
- Mertens, S. et al. Drug-repurposing screen on patient-derived organoids identifies therapy-induced vulnerability in KRAS-mutant colon cancer. Cell Rep.42, 112324 (2023).
- Herpers, B. et al. Functional patient-derived organoid screenings identify MCLA-158 as a therapeutic EGFR × LGR5 bispecific antibody with efficacy in epithelial tumors. Nat. Cancer3, 418–436 (2022).
- Betge, J. et al. The drug-induced phenotypic landscape of colorectal cancer organoids. Nat. Commun.13, 3135 (2022).
- Kawai-Kawachi, A. et al. Replication stress is an actionable genetic vulnerability in desmoplastic small round cell tumors. Cancer Res.85, 154–170 (2025).
- Hughes, J. P., Rees, S., Kalindjian, S. B. & Philpott, K. L. Principles of early drug discovery. Br. J. Pharmacol.162, 1239–1249 (2011).
- Lukonin, I., Zinner, M. & Liberali, P. Organoids in image-based phenotypic chemical screens. Exp. Mol. Med.53, 1495–1502 (2021).
- Angeli, C. et al. High-throughput drug screening in advanced pre-clinical 3D melanoma models identifies potential first-line therapies for NRAS-mutated melanoma. J. Exp. Clin. Cancer Res.44, 278 (2025).
