Immunotherapy and targeted therapies in animal models of cancer: a systematic approach from static reviews to living evidence synthesis

This dissertation, entitled “Immunotherapy and Targeted Therapies in Animal Models of Cancer: A Systematic Approach from Static Reviews to Living Evidence Synthesis,” addresses the translational relevance and reproducibility of preclinical cancer research by systematically evaluating therapeutic efficacy and animal models, and by introducing strategies to enhance their validity. Part I investigated the efficacy of immunotherapy in preclinical models of metastatic breast cancer through a systematic review and meta-analysis. While many studies reported therapeutic benefits, the results varied widely depending on the model and experimental design. This pronounced heterogeneity underscored the need for more rigorous methodologies and transparent reporting, while also highlighting both the potential and the limitations of existing breast cancer models in recapitulating clinically relevant disease biology. Part II examined preclinical murine models of pancreatic cancer with a focus on small-molecule inhibitors. The analysis revealed that although small molecules frequently induced tumour regression in vivo, reproducibility was inconsistent and clinical translation remained limited. These results exposed important methodological weaknesses but also provided guidance for more informed model selection and experimental design in pancreatic cancer research. Part III introduced the Cancer-SOLES pilot, a novel platform that integrates systematic review methodology with automation and visualization. For researchers, the platform reduces the manual burden of evidence synthesis and supports the identification of suitable models. For funders and policymakers, it offers a transparent map of research activity that can inform strategic investment and oversight. For the broader scientific community, it contributes to the 3Rs by minimizing unnecessary or redundant animal use and directing efforts toward models with the greatest translational value. Together, the three parts present a unified strategy to overcome barriers in preclinical cancer research and to make it more reproducible, ethical, and clinically impactful.