If you don't use optimization and really want to convert it, you basically have no choice but to do it all by hand, since PDF fundamentally doesn't have the information necessary to auto-convert and keep the format. (I believe there is an ongoing project -- as in years-old and still massively flawed -- to improve how to read PDF formatting.)
A good place to start learning about the problems -- and some solutions -- is here:
https://www.mobileread.com/forums/sho...d.php?t=118605
But after reading it you will probably want to stick with k2pdfopt.