Q: Ben, you're the EIPM's Director of Ex Vivo Models. Can you explain what you do?
A: Sure. The EIPM Ex Vivo Models Group receives tissues from patients with cancer, usually advanced cases. Our team builds organoid models of these tumors to test different therapeutics, aiming to find new treatments and improve patient care. The overarching goal is to translate research findings into the clinic to benefit patients.
Q: What drew you to this field, and what excites you about the research?
A: There’s a big gap between what should work on paper and what actually works in the clinic. Historically, we have lacked good cancer models. With advances in precision medicine, we can identify driver mutations but still struggle to treat tumors effectively. Our models help answer: What’s the best drug for a specific patient? It’s about bridging the gap between understanding tumor biology and effective treatment.
Q: Where are those numbers today compared to when you started?
A: Over the past 12 years, we've collected over 4,000 patient samples. We've run hundreds of drug screens and participated in numerous clinical trials. Our platform is increasingly focused on identifying therapies for patients who are in need of better therapeutic options.
Key Statistics:
- Over 4,000 patient samples collected and modeled over the past 12 years.
- More than 400 organoid models available for scientists to use for their research.
- It takes 17 people to process each sample.
- Our team of 38 researchers coordinate and execute projects.
- Interested in collaborating? Just follow this link: https://www.cognitoforms.com/ipm3/eipmcollaborationrequest
Q: How unique is this work compared to other institutions? Are many leading centers doing similar work?
A: We're quite unique. Our approach is distinct in that it covers a broad and diverse range of tumor types, giving us a strength that others lack. This allows us to identify tumor specific vulnerabilities and ask questions while we compare across multiple tumors, which is more powerful for translational research.
Q: What have you learned about organoids’ clinical applications over the last five years?
A: They’re now used as clinical tests in some cases, like cystic fibrosis—testing drug responses in patient-derived organoids. For cancer, the fidelity varies across tumor types. For rectal cancer, organoids have shown a 100% correlation with patient responses in one study. But in other cancers, like lung or pancreatic, the models have limitations, such as not fully capturing tumor dynamics or recapitulating resistance mechanisms. So, it is critical when using these models to be aware of both their strengths and limitations.
Q: Are there misconceptions about organoids in the field?
A: Yes. Some think organoids are like “mouse models in a dish” or that they fully replicate the tumor microenvironment. In reality, they’re complex but limited models. For example, they often lack vasculature or immune components. Overestimating their capabilities can lead to unrealistic expectations, while underestimating their potential may miss opportunities to capture key tumor behaviors.
Q: Tell us about the resource of organoids at EIPM and how accessible they are.
A: We’ve attempted to develop models from over 4,000 cases and have around 400 established models that grow well enough for broad distribution and use in research. Many have genomic data, like whole exome sequencing, matched to the primary patient tumors. These resources are available for research, helping to find models that match specific conditions or questions.
Q: What are some tumor types that don’t grow well or are difficult to model?
A: Prostate cancer is very challenging, less than 10% of attempts succeed. Also, certain mutations like STK11 are rare in organoids, possibly due to culture conditions or metabolic requirements that we do not accurately recapitulate in culture. Some combinations of mutations are hard to recapitulate because of biological constraints.
Q: Can you talk about collaborations within Cornell, Tri-I, or external partners?
A: We have broad collaborations—ranging from academic research with partners from Cornell in Ithaca on technology and molecular studies, to co-clinical trials where we get sequential tumor samples and are able to compare both the clinical and molecular responses of tumors to a test agent. We also collaborate with industry—pharma startups to big companies—testing new drugs and evaluating drug sensitivities to identify biomarkers. Internally, we have many potential collaborations across disciplines.
Q: Can companies purchase organoid lines from you?
A: Yes, our lines can used by anyone, but for industry in particular it’s complicated by patents and licensing. We do license some models, and often, they are accessible to external partners within current frameworks.
Q: Our Director Olivier Elemento, Ph.D. is enthusiastic about organoids advancing precision medicine. Where do you see this field in three to five years?
A: In the next 2-3 years, I believe organoids will help identify which patients respond best to specific drugs, especially for rare or hard-to-treat tumors. So I expect that they will be integrated into clinical decision-making, providing rapid, personalized treatment guidance. In five years, I envision organoids being used for diagnostic tests—such as directly informing treatment choices based on a patient’s tumor response in vitro.
Q: How will Artificial Intelligence (AI) impact your work?
A: Organoids let you generate massive data sets to integrate into big data questions. We don’t have that scale of data yet, but we are generating it. We do have an automated system at the EIPM that is Application Programming Interface (or API) compatible, so our robot on the 15th floor can be run autonomously. You could have an AI agent that generates a hypothesis and runs a test—with a human in the loop or not--but someday I might just be able to sit back and watch the robots run!
Q: Can you tell us more about the API?
A: An API is a system that can be run by another program, such as an AI. Currently we are working on a few projects that seek to develop AI that run entire experimental cycles, from hypothesis to validation. Such systems could use our automated system to run the experiments.
Q: Who typically reaches out to you for collaboration, are they grad students, clinicians, basic researchers, industry partners?
A: It’s all of the above. If you have a patient-centric question, KRASi-resistant tumors for example, or if a researcher just wants to collaborate on prostate cancer or is studying cancer samples from African American women, the EIPM team can work with you to identify the resources that already exist to help forward your research.
# # #
Related publications
- “Pan-Cancer PDOs Preserve Tumor Heterogeneity and Uncover Therapeutic Vulnerabilities,” bioRxiv, April 2025.
- “Functional genomics pipeline identifies CRL4 inhibition for the treatment of ovarian cancer,” Clinical and Translational Medicine, January 2025.
- “Unlocking the therapeutic potential of rigosertib as a selective therapy for ovarian cancer,” Cell Reports Medicine, July 2025.
- “Personalized In Vitro and In Vivo Cancer Models to Guide Precision Medicine” Cancer Discovery, July 2017.