Human and Animal Tumour Efficacy Evaluation: Endpoints and Methods from a Translational Perspective

Despite substantial advances in oncology drug development, promising preclinical efficacy signals frequently fail to translate into durable clinical benefit. This gap reflects not only the limitations of experimental models, but also a fundamental mismatch in how tumour efficacy is defined, measured, and interpreted across preclinical and clinical settings. In this review, we examine tumour efficacy assessment from a translational perspective, focusing on the endpoint frameworks and measurement modalities used in both domains. We highlight key differences between clinical efficacy assessment, which is anchored in imaging, biomarker-informed evaluation, and patient-centred outcomes, and preclinical efficacy assessment, which more often relies on tumour burden, survival, pharmacodynamic readouts, and mechanistic profiling. Although both settings share a common inferential basis, differences in biological context, sampling feasibility, and evidentiary standards limit direct cross-species comparability. These challenges are further compounded by interspecies variation in genetics, immune architecture, and biomarker relevance. We therefore propose a fit-for-purpose framework based on endpoint crosswalks, biologically aligned model selection, orthogonal validation, spatially resolved analysis, and transparent reporting. Taken together, this perspective argues that tumour efficacy should be interpreted as an integrated, clinically meaningful evidence framework rather than a singular model-derived signal, thereby strengthening translational decision-making in oncology.