The dimensions of a waveguide are determined by the wavelength of the
signal. For predictable operation, the smaller dimension of a
rectangular guide is beyond cutoff (i.e. less than half a wavelength),
and the larger dimension about twice that. Only one mode can exist in a
waveguide so built. The modes in circular guides become complex after
long runs, so they are mostly used in short sections near rotary joints.
Waveguides for X-rays would be exceedingly small. I wonder if their
existence is consistent with atomic spacing in ordinary materials. I can
imagine you thinking "This guy's full of it: what about light pipes?"
Light pipes, light guides, what have you, are analogous to what radar
and broadcast people call waveguides, but they are not the same. In
general, they support many modes simultaneously (and so have even worse
dispersion) or employ other means of become unimodal.
Waveguides and light guides work by a the wave zig-zagging from boundary
as it proceeds down the length of the guide. X-rays tend to penetrate
material, rather than be reflected. Reflectivity increases, however, as
the angle of incidence approaches 90 degrees.
Lenses to focus X-rays are made in many ways. One way uses concentric
shaped baffles that rely on low-angle reflection, another is basically
similar, but uses polycapillary channels. Both of those can be seen as
X-ray guides of a sort, so it can be done. I believe that the transport
efficiency is low, however, which would mean the the leakage is high.