I think if fraud can be shown, in an adequate quantity, the vote will be in enough question to call for a second runoff, regardless of who was voted for.
Certainly, enough evidence could be brought that there was fraud at a level sufficient to influence the result.
But I can't see an election being reversed on statistical testimony -- which must necessarily be premised on something like "the 95% level of confidence".
Statistical evidence, though, might be sufficient to authorize a re-vote as a remedy -- if Mississippi law provides for such.