No. For cases like that, you'd normalize and perhaps case-fold the text and pattern first, then break on a specific string. (Normalization is handled via text-icu.)