The given solutions in the thread did not work for me, I did find
another one which brings me very close to the result I want:

http://images.amazon.com/images/P/<xsl:value-of
select="substring(isbn/isbnprim, 6, 10)"/>.01.THUMBZZZ

This eliminates the first 5 characters, which was what I needed. But
now it turns out that some of the ISBN-numbers are written like this:
ISBN 91-12-304-403-x

Thus, I need to filter out the '-' and 'x' characters, anyone knows
how to do that?

thanx,
Gerard