Prada:
The interesting part is that after the tail breaks off, HS failure and APU cone breakaway become much less likely. A HS failure on the other hand can result in APU cone break away but after the HS is gone the tail breakaway is much less likely. Even a scenario where a 'pressure event' causes both the APU cone ejection and a HS damage, one expects that this causes either the tail to break after which HS failure is less likely or one expects the HS to break off fully, after which the tail is much less likely to break off.
A possible scenario is that there were multiple explosions because of the fuel and that the first one blew off the APU cone. Problem is that the damage pattern offers only weak support for it. I think this is the puzzle at the moment.