[R] 3D generalization of correlogram?

[Dieter Menne]

We have 3-Dimensional MRI density recordings of tumor tissue and would like
to have a measure of "patchiness", reflecting cluster size in the tissue.

For 2-D slices, correlogram from MASS works well. Does someone know of a
packages that provides a 3-D generalization of this measure? Or any
alternatives? I assume this is quite a common question in climate research.

Dieter