Abstract: The segmentation of the pathological lungs adhesion tumor division is the radiotherapy target area limits key. This article uses the CT sequence picture to segment the tumor boundary information in the complete lung area, then applies these information to further segment the adhesion tumor in its neighboring CT scans.The segmentation result and radiology doctor manually outline the boundary has made the comparison, used overlapping rate and the boundary position appraisal method. The result showed this method is feasible.

Key words: attached tumor, pathological lung, rotundity degree

摘要： 病理性肺部粘连肿瘤的分割是放射治疗靶区界定的关键。本文利用CT序列图像分割出完整肺区中的肿瘤边界信息，然后应用这些信息对其相邻CT图片中粘连肿瘤作进一步分割。使用重叠率和边界位置评价方法,对分割结果与放射科医生手工勾画的边界作了比较。结果说明该方法是可行的。

关键词: 粘连肿瘤, 病理性肺, 圆形度