I'll just add a supplement to @planetmaker's answer.
As long as a body is distinct and not connected to anything else, it will have a center of mass.
If the body is roughly spherical its center of mass will be near it's middle.
The body's rotational axis by definition passes through its center of mass, and is parallel to it's own angular momentum vector.
If there are two bodies orbiting each other, to a good approximation (but not exactly, see Which mass distributions guarantee two bodies have non-Keplerian orbits? Which non-spherical distributions still allow noncircular Keplerian orbits?) we can consider one's center of mass orbiting around the other's center of mass, and the center of both of their masses to be the pair's barycenter.
Their mutual rotation and orbital angular momentum will be defined by another axis passing through that barycenter, which may be inside one of them (like the Earth-Moon or Sun-Jupiter system) or in space between them like the Pluto-Charon system. It doesn't matter.
One is the rotation of a single body around its own center of mass, the other this the rotation of two centers of mass about their common center of mass. Apples and oranges.
If however a body is crazy-shaped, like a big letter "C" perhaps, that center of mass might be outside the body. That poses a conundrum for placing the "pole" of the body since the axis of rotation will not intersect the body's surface. IN that case the body will simply not have a true pole. But poles are constructs and not fundamental. It still has a center of mass and a rotational axis and angular momentum vector, and those are really what matter.
I don't know where the center of mass of comet 67P/Churyumov–Gerasimenko is exactly, nor where its poles are, so I think that that would be an excellent follow-up question!