Department of Health Sciences, Graduate School of Medical Sciences, Kyushu University, 3-1-1, Maidashi, Higashi-ku, Fukuoka 812-8582, Japan.
Department of Health Sciences, Faculty of Medical Sciences, Kyushu University, 3-1-1, Maidashi, Higashi-ku, Fukuoka 812-8582, Japan.
J Radiat Res. 2021 Mar 10;62(2):346-355. doi: 10.1093/jrr/rraa132.
The aim of this study was to develop an automated segmentation approach for small gross tumor volumes (GTVs) in 3D planning computed tomography (CT) images using dense V-networks (DVNs) that offer more advantages in segmenting smaller structures than conventional V-networks. Regions of interest (ROI) with dimensions of 50 × 50 × 6-72 pixels in the planning CT images were cropped based on the GTV centroids when applying stereotactic body radiotherapy (SBRT) to patients. Segmentation accuracy of GTV contours for 192 lung cancer patients [with the following tumor types: 118 solid, 53 part-solid types and 21 pure ground-glass opacity (pure GGO)], who underwent SBRT, were evaluated based on a 10-fold cross-validation test using Dice's similarity coefficient (DSC) and Hausdorff distance (HD). For each case, 11 segmented GTVs consisting of three single outputs, four logical AND outputs, and four logical OR outputs from combinations of two or three outputs from DVNs were obtained by three runs with different initial weights. The AND output (combination of three outputs) achieved the highest values of average 3D-DSC (0.832 ± 0.074) and HD (4.57 ± 2.44 mm). The average 3D DSCs from the AND output for solid, part-solid and pure GGO types were 0.838 ± 0.074, 0.822 ± 0.078 and 0.819 ± 0.059, respectively. This study suggests that the proposed approach could be useful in segmenting GTVs for planning lung cancer SBRT.
本研究旨在开发一种基于密集 V 网络（DVN）的自动分割方法，用于在 3D 计划计算机断层扫描（CT）图像中分割小的大体肿瘤体积（GTV），与传统 V 网络相比，密集 V 网络在分割较小结构方面具有更多优势。当对患者进行立体定向体放射治疗（SBRT）时，根据 GTV 质心从计划 CT 图像中裁剪出感兴趣区域（ROI），其尺寸为 50×50×6-72 像素。对 192 例接受 SBRT 的肺癌患者[肿瘤类型如下：118 例实体瘤、53 例部分实体瘤和 21 例纯磨玻璃密度（纯 GGO）]的 GTV 轮廓分割准确性进行了评估，使用 Dice 相似系数（DSC）和 Hausdorff 距离（HD）进行了 10 折交叉验证测试。对于每个病例，通过三次不同初始权重的运行，从 DVN 的两个或三个输出组合中获得 11 个分割的 GTV，每个 GTV 由三个单输出、四个逻辑与输出和四个逻辑或输出组成。AND 输出（三个输出的组合）实现了平均 3D-DSC（0.832±0.074）和 HD（4.57±2.44mm）的最高值。实体瘤、部分实体瘤和纯 GGO 类型的 AND 输出的平均 3D-DSC 分别为 0.838±0.074、0.822±0.078 和 0.819±0.059。本研究表明，该方法可用于分割计划肺癌 SBRT 的 GTV。
