Just leverage Microsoft Flow's filtered trigger mechanism based on a new item creation or update on a combination of columns or any column.

For example, use Nintex Workflow to update a Status column with the value "LaunchFlow". Then based on your filter in Microsoft Flow, when it "sees" the Status column value change to "LaunchFlow" then it will be the trigger to commence processing.

There is no need for Nintex to be involved directly in the execution path transfer.

---