During the last two decades, the interest for computer-interpretable guidelines has kept growing to become a major issue in medical informatics. Clinical guidelines usually contain decisional knowledge that can be represented by decision trees. This paper presents KcatoS, a semantic decision tree editor, which provides a collaborative tool to simplify knowledge acquisition. Using a simple graphical language, KcatoS allows exporting decision trees to formalised knowledge, by proposing an original export algorithm to OWL. Easy to integrate in web applications, KcatoS is part of a larger work about collaborative editing of clinical guidelines. The overall objective is to provide tools to assist editing and storage of decisional knowledge in the field of oncology.