Currently any nearest neighbor analysis that cannot be done with clustering can be done in R, after you export the Detections list (which come with X,Y coordinates).
saveDetectionMeasurements() works in the same way that the annotations version does, see elsewhere on the forums.
As far as quantifying and classifying, see here. Standard warnings about cytoplasmic stain apply.
I have been playing around with some ideas for NN analysis, but haven’t written any scripts yet. It is fairly straightforward in R (entire modules devoted to it), so there hasn’t been a huge need.
If you can classify an area (like tumor), you can use the distance transform function in 0.2.0m2 to find the distance of a cell type from the tumor annotation.