Revisiting the paradigm on the putative need for antigen-specific responses in cancer