A revision to the protocol is different than a revision to the SAP. The SAP is part of the protocol, the criteria for how they adjudicate images is a different part of the protocol. If the sponsor wants to change how imaging is characterized, it takes more than a revised SAP. I am not saying that can't be done, but don't assume a revised SAP alone is sufficient to change adjudication of images because it is not.
As to your other question, either the trial meets the primary endpoint per the protocol (including any agreed amendments) or else it does not. If the trial meets the prespecified endpoint with the required statistical margin (95% CI which is the same as p <= 0.05), then the drug is approvable absent any significant safety concerns. However, if the endpoint can only be reached with a lot of statistical massaging then FDA would be likely to require an additional confirmatory trial before approval is granted.