Study of variation of pelvis positioning for patients suffering from rectal cancer using daily kilo-voltage cone beam CT images

Radiation therapy is a standard treatment for a large variety of cancers. Radiation treatment planning involves delineating tumor, areas-at-risk, as well as normal anatomical structures. Automated techniques for detection and segmentation of tumor is a challenging task because of the absence of any standard model to identify anatomical structures. As radiation therapy is delivered in multiple daily doses (called fractions) - treatment fields need to be positioned so that the daily uncertainties in the target position are covered adequately. This is done by adding a margin called the Planning Target Volume (PTV) margin around the clinical target volume. The PTV margin is determined from an understanding of the daily uncertainties. In this paper, we intend to aid doctors in determining the uncertainties of daily pelvis positioning for radiotherapy by estimating the degree of change in the pelvic bone position. We propose a technique for automatically detecting regions of pelvic bone anatomy from CBCT images. We track the change in the bone positions in pelvic regions on a daily basis. A small change suggests that the uncertainty involved in the position of tumorous cell is less, indicating that the planning target margin may be reduced substantially.