Computer-aided Delineation of Lung Tumor Regions in Treatment Planning CT Images and PET/CT Images Using Localized Level Set Approach

摘要

The purpose of this study was to develop a computer-aided delineation method of lung tumor regions in treatment planning computed tomography (CT) images and positron emission tomography (PET)/CT images using a localized level set approach. Six data sets of treatment planning CT and PET/CT images of lung cancer patients were selected for this study. PET images were registered to the planning CT images by using an affine transformation matrix, which was the same transformation matrix derived from registration between the planning CT images and CT images of PET/CT images. The initial regions of the lung tumors were identified by thresholding the PET images at a certain standardized uptake value (SUV). We proposed a localized level set method (LLSM), which determines an optimum contour of the GTV region in the planning CT images based on the initial regions. For performance evaluation, we employed the Dice similarity coefficient (DSC), which denotes the degree of region similarity between the gold standard determined by radiation oncologists and the GTV region extracted by the proposed method. Also, we made a comparison between the results, which were obtained by using the proposed method and a conventional method, i.e., the fast level set method. Our proposed method reached an average DSC of 0.77 for the six cases, which was greater than that of 0.44 obtained by the conventional method. The proposed method may be useful for assisting treatment planners in delineation of the tumor region.