Bi Nan, Wang Jingbo, Zhang Tao, Chen Xinyuan, Xia Wenlong, Miao Junjie, Xu Kunpeng, Wu Linfang, Fan Quanrong, Wang Luhua, Li Yexiong, Zhou Zongmei, Dai Jianrong
Department of Radiation Oncology, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China.
National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital and Shenzhen Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Shenzhen, China.
Front Oncol. 2019 Nov 13;9:1192. doi: 10.3389/fonc.2019.01192. eCollection 2019.
To investigate whether a deep learning-assisted contour (DLAC) could provide greater accuracy, inter-observer consistency, and efficiency compared with a manual contour (MC) of the clinical target volume (CTV) for non-small cell lung cancer (NSCLC) receiving postoperative radiotherapy (PORT). A deep dilated residual network was used to achieve the effective automatic contour of the CTV. Eleven junior physicians contoured CTVs on 19 patients by using both MC and DLAC methods independently. Compared with the ground truth, the accuracy of the contour was evaluated by using the Dice coefficient and mean distance to agreement (MDTA). The coefficient of variation (CV) and standard distance deviation (SDD) were rendered to measure the inter-observer variability or consistency. The time consumed for each of the two contouring methods was also compared. A total of 418 CTV sets were generated. DLAC improved contour accuracy when compared with MC and was associated with a larger Dice coefficient (mean ± SD: 0.75 ± 0.06 vs. 0.72 ± 0.07, < 0.001) and smaller MDTA (mean ± SD: 2.97 ± 0.91 mm vs. 3.07 ± 0.98 mm, < 0.001). The DLAC was also associated with decreased inter-observer variability, with a smaller CV (mean ± SD: 0.129 ± 0.040 vs. 0.183 ± 0.043, < 0.001) and SDD (mean ± SD: 0.47 ± 0.22 mm vs. 0.72 ± 0.41 mm, < 0.001). In addition, a value of 35% of time saving was provided by the DLAC (median: 14.81 min vs. 9.59 min, < 0.001). Compared with MC, the DLAC is a promising strategy to obtain superior accuracy, consistency, and efficiency for the PORT-CTV in NSCLC.
为研究与非小细胞肺癌（NSCLC）接受术后放疗（PORT）的临床靶区（CTV）的手动轮廓（MC）相比，深度学习辅助轮廓（DLAC）是否能提供更高的准确性、观察者间一致性和效率。使用深度扩张残差网络实现CTV的有效自动轮廓。11名初级医师分别使用MC和DLAC方法在19例患者上勾勒CTV轮廓。与真实情况相比，通过使用Dice系数和平均一致距离（MDTA）评估轮廓的准确性。采用变异系数（CV）和标准距离偏差（SDD）来测量观察者间的变异性或一致性。还比较了两种轮廓勾勒方法各自所花费的时间。共生成了418组CTV。与MC相比，DLAC提高了轮廓准确性，且Dice系数更大（均值±标准差：0.75±0.06 vs. 0.72±0.07，<0.001），MDTA更小（均值±标准差：2.97±0.91 mm vs. 3.07±0.98 mm，<0.001）。DLAC还与观察者间变异性降低相关，CV更小（均值±标准差：0.129±0.040 vs. 0.183±0.043，<0.001），SDD更小（均值±标准差：0.47±0.22 mm vs. 0.72±0.41 mm，<0.001）。此外，DLAC节省了35%的时间（中位数：14.81分钟 vs. 9.59分钟，<0.001）。与MC相比，DLAC是一种有前景的策略，可为NSCLC的PORT-CTV获得更高的准确性、一致性和效率。
