In the WGCNA why we want to calculate de dissimilarity TOM? The other question that I have is, once that I have my network visualization with the igraph package how I can export it to cytoscape? Which steps I should follow?
WGCNA uses the dissimilarity matrix from the topological overlap matrix to reduce the effects of noise and spurious associations (from the manual), before doing the hierarchical clustering.
As to your second question, probably the easiest way to get from igraph in R to Cytoscape is to use the RCy3 package, which includes specific functions that will talk to the running cytoscape and push the igraph network to it. You can read more about RCy3 at https://github.com/cytoscape/cytoscape-automation/wiki