Lost in Translation: Preclinical Modeling in Pancreatic Cancer - Where Do We Go from Here?

The success rate in oncology drug development is unacceptably low. In pancreatic ductal adenocarcinoma (PDAC), in particular, the overall likelihood of approval (LOA) for agents entering phase I development is a mere 2.3%, the lowest by far, not only in oncology. Failure of developing effective targeted agents in this disease is particularly bothersome. Methodological issues (trial design, patient selection, lack of solid predictive biomarkers, issues with go/no go decisions based on limited early clinical evidence) underpin the recently reported failure of EGFR/VEGFR inhibitors and MEK/AKT inhibitor combinations, among many other agents, to impact on PDAC patients’ survival. As a result, only three clinically meaningful agents or combinations (all conventional chemotherapy drugs) have been reported in the past 25 years. In addition to issues with clinical research methodology, better preclinical models, more closely predicting clinical situations, are urgently needed, especially if we are to successfully develop targeted agents that exploit the much deeper knowledge of PDAC biology acquired in the past few years.  In the era of precision oncology, preclinical models are required: (i) to be rapidly and efficiently established from almost all patients; (ii) to appropriately capture the genetic and non-genetic complexity of the human disease; and (iii) to serve as robust platforms to model complex cellular interactions occurring in vivo. Pancreatic cancer organoids show promise as such individualized models that will help improving outcomes of PDAC patients.