Determination of target cancer types and prediction of therapeutic efficacy are essential for the successful development of anticancer drugs. Cell lines have been used in preclinical studies to select cancer types. However, it has been pointed out that the homogenization of tumor cells during the establishment process leads to changes in genomic information and tumor characteristics, resulting in poor prediction of clinical response.
As a bioresource, patient derived-xenografted (PDX) models is recognized as an essential tool. Although it is complicated and costly to manage, it maintains tumor cells' heterogeneity, microenvironment, and tissue structure and is highly predictive of clinical response (60-80%).
The National Cancer Center has created a lot of PDX models derived from Japanese cancer patients for predicting treatment effects. It has established the J-PDX library, Japan's largest bioresource of PDX. This library has collected much information on biological characteristics such as growth behavior, pathology, genomic analysis, gene expression analysis, epigenomic analysis during PDX establishment, and donor patient's detailed clinical information.
We have been conducting preclinical studies of novel anticancer drugs using PDX models to evaluate drug responsiveness across cancer types, to make "Go or No-Go" decisions to move into Phase I clinical trials, and to select cancer types to be developed.
In this symposium, we will present the results of a Co-Clinical study comparing the clinical response of PDX models and cancer patients and the utility of tumor tissue derived from cancer patients in drug discovery research.