Target volume definition for three-dimensional radiotherapy of cancer patients with a fuzzy rule based system

Abstract

Radiotherapy is an effective treatment for many cancer patients. The volume to be irradiated, i.e. the target volume, is determined by the radiotherapist who aims, in this process, at including the whole tumor at the same time as saving the surrounding healthy tissues as much as possible. The target volume definition is a difficult step, since the imaging techniques on which this process is based can visualize the gross tumor mass but not all microscopic tumor cell spread. This causes diagnostic uncertainties, leading to a fuzziness of the target volume. In order to assist the radiotherapist in the process of target volume definition, we developed a new approach. The radiotherapist defines a minimal target volume which surely contains tumor cells and a maximal target volume, outside of which no tumor cell spread is expected. The fuzziness region in between is processed by a fuzzy rule based system that calculates the membership values of individual subvolumes in the fuzziness region with respect to the target volume. The result is a fuzzy target volume which includes the diagnostic uncertainties in the target volume definition. Subsequently, after defuzzification a crisp target volume with optimized extension is proposed to the radiotherapist.