The pursuit of active specific immunotherapy of cancer re-emerged vigorously in the 90s. More than 50 vaccines are currently under clinical testing, and more than 400 clinical trials have been conducted. This wave of enthusiasm is rooted in fundamental immunology, as new paradigms, such as the dominant tolerance through T-regulatory cells and the instructive role of the innate immune system on the adaptive immune system, opened the possibility that an efficient cancer vaccination could be achieved even without the need of cancer neoantigens, provided that antigen presentation could be increased, and that regulatory circuits could be controlled. However, recent failures in some large trials have brought disappointment and have highlighted the differences between experiments in young, healthy mice with small transplanted tumours, and clinical testing in aged, ill patients with advanced spontaneous tumours, driving the attention to issues such as tumour editing, tumourinduced immunosuppression, and immunosenescence. The molecular basis of these phenomena is only partially known. Additionally, the inherent complexity of the immune system as a network of multiple interactions and redundant control loops among a huge diversity of components sets another barrier to the translation of in vitro reductionist knowledge into rationally designed clinical trials. All this calls for a new therapeutic paradigm in cancer vaccines, moving beyond the analogy with the classic drug-target approach, and targeting the immune system regulation as a whole, and its interaction with the tumour, in all its complexity. Early mathematical modelling of cancer immunotherapy has suggested how to go about it. This reevaluation of the cancer vaccine landscape, suggests that future successful cancer immunotherapy will be combined immunotherapy, will be exquisitely schedule-dependent and will need new experimental models allowing for the exploration of the mechanisms of resistance and tumour escape, such as tumour editing and tumour induced immunosuppression, in the context of the physiology of the immune system of the elderly. These shifts will put cancer vaccines closer to pharmacology than to conventional preventive vaccinology, or at least at the midway. A change in the design and the ultimate goals of the clinical trials will also be needed, identifying long term stabilization of the disease and quality of life as main endpoints, again closer to the clinical management of most chronic noncommunicable diseases.