Guidance on Applications for Studies with Animal Models 

Animal models have proven useful in answering many questions in cancer biology. Following NIH guidelines, NCI continues to fund research using animal models. However, it is essential that researchers whose proposals call for the use of animals explicitly justify why animal use is necessary, and why alternatives would be inadequate. NIH RePORTER makes each award's Abstract publicly searchable, but it often does not include the animal-use justification section.

Justifying animal model use in NCI grant applications

I request that applications starting with the October 5, 2026, deadline that include the use of animal models include a brief, explicit justification of the use of animal models and why alternatives are inferior. This justification must be included in the application’s Abstract section. A sentence or two explaining why the use of an animal is necessary and why alternatives are inferior is likely sufficient.

Planning human model validation for animal studies

Even in experimental contexts where animal models are the best way to answer an important question, it is necessary that findings be validated in a human model, as there remain gaps in the translation of animal findings to human disease. For example, while genetically engineered mouse models (GEMMs) of cancer have been informative in many cases, there are also cases where GEMMS did not accurately reflect the human condition. This might be due to differences in tumor latency, genetic background, epigenetic patterning, or unknown factors. Whatever the cause, if we are to translate concepts derived from animal models, we need to have confidence that the principle learned in a mouse is valid in a human.  

Therefore, when grant proposals contain animal experiments, the proposal should also contain a plan for validation of the findings of this animal experiment in a human model. Human models might include organoids, fresh tumor biopsy samples, early passage cultures, patient-derived xenografts, or other models initially derived from human patient tumor tissue. It may be that the actual performance of the validation falls beyond the fundable scope of the proposal. It is nonetheless important that a path to validate these findings in a human model be evident. It is up to the applicant to provide a description of this path. While a more thorough description might be provided in the Vertebrate Animal Section, I request that this brief validation plan, again a sentence or two should suffice, be included in the Abstract.  

Choosing the best research models for human impact

Investigators should not consider this requirement as “business as usual with a couple extra sentences of boilerplate.” Human models, be they organoids, fresh tumor biopsy samples, early passage cultures, or other approaches are easier to access and use than ever before and should be strongly considered. Methods now exist to directly interrogate the genome, proteome, metabolome, transcriptome and pharmacotype of almost any human cancer sample. Information gathered from human cancer samples is directly informative for the human cancer one is trying to study. Moreover, proper sampling allows recapitulation of the heterogeneity of the human disease in a way no mouse model can.  

Scientific studies should remain inspired by asking great questions with high potential for human impact, using the best and most relevant scientific methods to accomplish this goal. For some projects and scientific goals, mouse models may be among the tools necessary to answer such questions. However, it would be regrettable if a scientist identified too closely with a single tool in this rapidly evolving research landscape with growing opportunities for implementation of novel, rapidly translatable approaches. A scientist should no more identify solely as a “mouse modeler” than as an “immunoblotter” or an “organoidologist.” Identifying the best tool to answer an important question is a critical part of the job of a scientist. Modern advances have rendered human models an increasingly useful tool that should be considered by cancer scientists.  

NCI looks forward to continuing to support meritorious research utilizing the full array of tools, leveraging the ones that best contribute to our understanding ways to prevent, detect, and treat human cancer.