I think it needs to be approached like any other end project retrospective. The difference here is the project has been canceled.
1. It's important to review the last scrum process and understand what went wrong and what went right and how can we continue improving for future projects. (Lessons learned)
2. It's important to understand why the project has been canceled and how it can be avoided in future projects. (Important for the executives and team)
There are many reasons for a project to be canceled: Don't achieve the needed time to market, lack of budget wrong technology, expectative management, change in the company strategy, lack of skills needed for the project, and many others...
And if the reason is identified it needs to be feedback for the company for future improvement.
**The most important thing is to keep the focus on the problem** , never on the people. In that way, you avoid exposing the team and step in the executive's foot keeping the morale of the team.