Segmentation of rectum from CT images using K-means clustering for the EBRT of prostate cancer

Abstract

Cancer is one of the most haunted conditions in the society even though there have been treatment methods developed to cure. External Beam Radiation Therapy (EBRT) is a type of radiation therapy. Radiation therapy mainly includes three phases-treatment planning, treatment plan evaluation and treatment delivery. During the treatment plan phase, the organs at risk in the respective cancer context need to be segmented from acquired medical images. In the yester years, the segmentation of target and the organs at risk was done by the doctor himself which is very tedious. But now, with the developments in the fields of technology, computerized algorithms are used for the purpose of delineating the organs at risk. The images used for treatment planning are usually CT images. This is because, the density information obtained in CT is used directly for dose calculation. In this work, rectum is segmented from CT images and segmented rectum images are useful in the treatment plan. The work deals with the segmentation of rectum during EBRT plan for treatment of prostate cancer and image processing algorithm developed based on k-means clustering. The segmentation was carried out using MATLAB. The method proves to be efficient and nearly accurate in segmenting rectum from the CT images of male pelvic region.