Can PDC models improve oncology drug attrition rates?

Can PDC models improve oncology drug attrition rates?

by Imagen Therapeutics|March 1, 2022 at 2:00 PM

A large and well characterised biobank of patient-derived cell models can significantly accelerate drug development timelines and fast track decisions of what new agent or combination strategy to move forward in the clinic. But how can you ensure the right drugs are reaching the right patient population?

Are we using the right preclinical models?

Cancer is a complex disease and new oncology drug attrition rates are significantly higher than for other diseases. New drugs reaching the clinic have usually shown promising pharmacological profiles at the preclinical stage but while tested on patients they fail to deliver the desired patient benefits.

There are multiple reasons for this failure and the poor translatability of currently available preclinical models is regarded as an important factor.

Immortalised cell lines have been used extensively in drug development to assess target engagement and test efficacy and potency of new compounds. Although they have contributed to much of our understanding of the biology of cancer, they have clear limitations when it comes to predicting patient response in the clinic. Cell lines poorly represent the heterogeneity of the tumours in vivo and therefore are not able to recapitulate mechanisms of drug resistance.

Cell line-derived xenografts have been developed to try to more closely mimic tumours in vivo and still extensively used in drug development. However, they can grow at a much faster rate compared to primary tumours and, as a result, are much more likely to respond to antiproliferative agents, potentially leading to false indications on the efficacy of a drug.

Human xenografts, such as PDX better recapitulate human tumours pathology, but they can be costly and time consuming and as a consequence are not a viable alternative for applications such as high-throughput drug screening, or co-clinical investigations.

Getting the target right is crucial

Not all targets are exclusively present on tumour cells and not all tumour cells express the target consistently. Some targets are transient, and their level of expression may have different effects on drug efficacy. Too much or too little of the target, could allow tumour cells to evade the cell killing activity.

Therefore, it is easy to see how designing a new drug target based on results from immortalised cell lines or xenograft that do not necessarily preserve the heterogeneity or target expression profile of the original tumour, may lead to later failures in the clinic.

How can PDC help?

Patient-derived cell models or (PDC) are freshly derived from patient tumours and kept at low passage to ensure they remain faithful to the tissue of origin.

Here is how they can help

  • They can be established with high success rate from multiple solid tumours much quicker than any patient-derived in vivo models
  • They can be more easily expanded and enrolled in a screening to test multiple drugs and combinations in parallel
  • They have a strong predictive power of patient response and can offer in vitro pharmacology data to support IND application.
  • A large collection of PDC models has already been established and benchmarked for response to more than 60 SOC agents, including chemotherapeutics and targeted agents, and can be interrogated to validate the efficacy of a new agents against or in combination with standard treatments.
  • PDC models NGS data is available to test your agent on a clinically relevant cohort of patient-derived models expressing your target of interest.
  • PDC are amenable to high content phenotypic screening which provide better insights into your drug mechanism of action.

My drug has reached IND application, how can I ensure it will succeed

Having identified and validated a druggable target, a PDC-based platform can also enhance the quality and effectiveness of the subsequent clinical studies as a companion diagnostic tool.

For example, a PDC screen could help identify biomarkers indicative of increased susceptibility of tumours to a candidate drug, which can then be measured in patients enrolled for a Phase 1 trial to identify patients with tumours which are more likely to respond. Increasingly, patients are being expected to provide tumour biopsies on recruitment, for stratification purposes and these can be used to generate PDC and examine the phenotypic response of each tumour to the trialled drug. This allows the clinical cohort to be enriched with patients more likely to benefit from the trial itself, and in turns reduces attrition rates. Almost as important, it allows the deselection of patients who are unlikely to benefit.

This concept can be extended to improve the quality of data that can be generated from later-stage trials as well. By taking biopsies from all patients enrolled in Phase 2 & 3 trials, we can build an in vitro population of PDC “avatars” which can be used to better understand patient responses, both phenotypically and based upon genetics, even long-after the trial itself has finished. The value of such a set of models in the development of the next generation of drugs could be tremendous.


Advanced and patient-relevant preclinical models are sorely needed to reduce oncology drug attrition rates. PDC are unique patient-derived models that preserve relevance to original tumours and can help predict response of patients in the clinic, identify new targets, and better understand your compound mechanism of action, thus helping to de-risk your new drug development program. At the clinical stage PDC derived from patient in the clinic can be adopted in co-clinical in vitro trials to identify the most promising treatment regimen among existing SOC or predict which patient population will benefit the most from a new therapeutic strategy.