You don't need to repeat the whole line when doing the replacement---you can use a backreference, or simply use this construct:
sed -i '/regex1/ s/regex2//' filename
reads: For any line matching regex1, delete the text matching regex2 (and modify the file in place).
Does "=" need to be escaped?
Finally, you very seldom need to match an entire line---look for a regex that is unambiguous in the context of the contents of the file.