Query about homology modelling of missense mutants

We are studying alterations in structural parameters upon missense mutations in the enzyme. Let's consider we have generated a homology model of wild-type enzyme, say A, and a homology model of the missense mutant, say B, based on the same template. Now we want to find out the structural differences between A & B.

We want to know how logical it is to compare homology models of 2 variant proteins which differ from each other by a single residue? As the protein sequences (input for homology modelling) differ by only a single residue, could we expect significant and accurate differences between the two homology models? Is it at all possible to get information on such subtle differences in the structures by using current template based modeling platforms?

Thanks in advance.....