Where is the proof someone got access to the backend and not those specific accounts? Seems more likely an API client got hacked, possibly one that high profile people might use like a tweet scheduler, but not Twitter, given their threat profile and resources. That would explain why 2FA accounts were affected.