Abstract
As the prescribed dose difference between tumor cell and normal tissue increases due to the advances of treatment technology, the treatment planning becomes more sensitive to treatment uncertainties. Among various treatment uncertainties, accurate evaluation of organ motion is a key factor to quality improvement in treatment planning. In this study, we propose an automatic boundary extraction algorithm to extract medium-contrast enhanced esophagus in fluoroscopy. The algorithm is based on image thresholding, image edge detection, morphological closing to identify the esophageal boundary, and motion quantification of mass center. The results of boundary extraction can provide an accurate motion estimate of the internal margins in superior-inferior direction around 0.6~ 1.74cm, and in anterior-posterior around 0.2~ 0.6cm. The algorithm can also extract edge position in the image sequence automatically and identify esophageal boundary efficiently. Our experimental results have indicated the amount of esophageal motion extracted by the proposed algorithm serves as a good reference for the internal margin needed in radiotherapy. It is clinically feasible to increase the prescribed tumor dose while to reduce the normal tissue absorption.