Though there is hub-Spoke topology and eNodeB has to reach at the aggregation point for the messages exchange and for S1 handover also there is same path.
This is the case where according to you If this scenario occurs, what will be the behavior of the Network or else which handover will be advantageous.
As we know if we don't talk about this scenario, In normal scenario X2 based handover is better than S1 based handover because of less signalling, fast and quick. In a same way this scenario seems to me the same situation like the normal has.
X2 based handover will yield the advantage. Just consider the message flow.
Call Flow for X2 based handover:
Source-EnodeB------> Aggregation Point-------->Target EnodeB (Source and target MME is same)
This is how handover will be possible and messages which are to be exchange you know. Only one MME will be there. I mean to say source MME and Target MME is same so It saves the signaling.
Call flow for S1-Based Handover:
SeNB-->aggregation point--->SourceMME---->TargetMME--->Target eNodeB (Source and target MME is diff)
This is the extra signaling. So if both are in same condition and we have to compare between the preferable handover. I think the network should go for the X2. Exceptions are always there. It mainly depends upon the network condition and the implementation by the Operator side. It is the theoretical concept which I have been considered here. If something is slipped, please let me know.