try this:
<br>\s*(title1|title2|title3|title\s+with\s+spaces )\s*<br> - changing title to whatever your chapters are.
You can also add <hr>\s* or \s*<hr> to the beginning or end (depending on whether it's header or footer), to more accurately tie it just to the page headers. If you tie it to the <hr> tag you might be able to get away with something like this:
<br>\s*(\w+\s*)+\s*<br>\s*<hr>
Use the test function with some of those examples to see if you can get what you need.
http://www.regular-expressions.info/ is the best place to read up on how to use regex.
edit - here's a sample regex I used for a file which also had chapter title headers:
Code:
((Castello\s|The\s(Phleg|nun|night|prince\sof\smus|garden|secret\spalac)|Epilogu|Prefac|Four\scarnival|Amalf|La\sSiren|Marriage\sto|Montevergin|Spaccanapol|A\sstiletto|Gesualdo\sC)[^<]+<br>\s*)?(\d|[xvi])+<br>\s*(The\sD\s*e\s*v\s*i\s*l\s*[^<]+<br>)?\s*((Bh|27)[^<]+<br>\s*){4,4}\s*<hr>\s*<A name=\d+></a>
I believe in this case it was a footer, <A name=\d+></a> also shows up on every page break, so it's another way to tie the regex to the header/footer by including that in the pattern.