The second part of this question is—so what? How remarkable are these wombled boundaries?
As you saw in the analysis, you find wombled boundaries by choosing the top x-percent of the boundary likelihood values (in this case, x-% = 20%). The next step is to determine whether these boundaries are more "boundary-like" than you would expect by chance.
By boundary like, we mean: longer, higher diameter, and lower branchiness than you would expect by chance. To assess this, we use subgraph statistics. The characteristics of the analyzed boundaries are compared to those boundaries calculated for Monte Carlo randomizations of the dataset.
You can do this once you have a set of wombled boundaries.
Choose Subboundary Analysis from the Boundary menu OR after right clicking on the local leukemia boundaries in the Boundaries tab of the project window.
Boundary: local leukemia boundaries
Monte Carlo Settings (accept the defaults):
99 randomizations
Complete randomization
Output Settings:
Accept the name chosen
Clear the box next to standardize results (useful for comparison to other boundaries, not necessary in this case)
Accept that you will view histograms after analysis.