It looks like the PDF > ePub conversion left in the headers and footers and all those lines are part of the header.
The only way to do the PDF > ePub conversion properly is to take the resulting ePub and A/B compare the PDF to the ePub and manually fix all the errors in the ePub due to the conversion process. You cannot convert that PDF > ePub without errors. It just isn't possible. So that is why an A/B conversion is needed.