INTRODUCTION
Oncology drug attrition rates are significantly higher than for other therapeutic areas. This is
partly attributed to the poor translatability of current preclinical models and to their suboptimal
application in drug development programs.
Patient-derived xenografts (PDX) have been widely adopted over the last two decades for
highly predictive in vivo therapeutic testing in translational research, as they preserve key
patient tumor features and the response to certain treatments. PDX, however, can be costly
and time consuming to develop and are not amenable to large scale screens encompassing
multiple test agents or combinations across several models in parallel.
New patient-relevant models that can be utilized early in the drug development process (e.g.
more clinically relevant in vitro platforms) are urgently needed to better identify target patient
populations and improve anticancer agent success rate. Cell lines currently used for in vitro
investigations are poorly reflective of original disease and maintained in a 2D environment, but
are routinely used in the early stages of drug development.
Tumor organoids are a novel 3D in vitro system that, similarly to PDX, preserve the morphological,
genomic, and pathophysiological identity of their corresponding in vivo tumor by showing similar
pharmacological profiles and treatment response. Organoids can be generated from tumors
directly from a patient´s biopsy or resection, and sufficiently expanded for large scale screens
without losing the original tumor identity, demonstrating high predictive power and great
potential to revolutionize the drug discovery workflow
THE NEED FOR EARLY STAGE PATIENT-RELEVANT MODELS
Since the US National Cancer Institute (NCI) announced the phasing out its NCI-60(1) panel of
1)
cancer cell lines for drug screening purposes in favor of a PDX repository, numerous libraries of
PDX models have been established representing the diversity and heterogeneity of the patient
population. PDX are widely adopted in vivo models in preclinical drug discovery and are more
patient relevant than conventional xenografts, as original tumor morphology and genetic
make-up are preserved. Patient tumors are never manipulated in vitro when establishing PDX
models, which supports the maintenance of original patient tumor heterogeneity.
PDX models have impacted drug development in many different ways. Their most prominent use
has been as patient surrogate models, which are enrolled in Phase II-like mouse clinical trials to
identify responder populations and predictive biomarkers. The power of PDX in mouse clinical
trials is highlighted in high profile scientific publications(2) demonstrating the feasibility of using
2)
these models as patient "avatars" to predict response.
Mouse clinical trials are run as prospective studies, with study design varying to focus on either
a specific cancer mutation across various cancer indications or a tissue specific cancer type.
Typically, these studies take a population approach and require a fairly large cohort of mice to
produce statistically relevant predictions of agent efficacy. N of 1 studies are also run where 1
animal is enrolled in the treatment and 1 in the control arm but these require large numbers of
different models of the same cancer type to produce meaningful data.
Co-clinical trial approaches have been trialed where PDX are generated and run at the same
time, with the same treatment as the patient in the clinic. However, the applicability of this
approach to a wide patient population has been challenging due to the time taken to establish
PDX, associated costs, and ethical concerns around animal use.
is now part of
huborganoids.nl
2
