In Execution Plan can one testcase be made contingent upon the successful completion of another?

Have a look at Handling Events, an event can be triggered if an error occurs, which you can then deal with programmatically. For example, restoring a database to it's original state for the next test.

In my opinion, there's no point in making the automation complicated. Plus it could be a nightmare to maintain too.