基于自动勾画和快速蒙特卡罗计算的放射治疗全身器官剂量数据库平台的可行性研究

Feasibility of establishing a whole-body organ dose database for radiotherapy using automatic segmentation and fast Monte Carlo calculation

目的:探讨建立一种放射治疗全身器官剂量数据库平台的可行性。方法:使用基于深度学习的自动勾画软件DeepViewer^(®)在1例食管癌患者的全身CT上勾画全身器官,然后利用基于GPU加速的蒙特卡罗软件ARCHER计算相应的器官剂量分布,最后利用Lyman-Kutcher-Burman(LKB)模型评估放疗患者正常组织并发症概率(NTCP)。结果:针对该病例,成功建立基于DeepViewer^(®)、ARCHER和LKB模型的全身器官剂量数据库,发现距离靶区越近的器官剂量越大,其中心脏与靶区间距离最小,剂量为14.11 Gy,但因其模型参数特殊,通过LKB模型计算的NTCP为0.00%;左、右肺的剂量分别为3.19和1.16 Gy,但是NTCP值却很大,分别为2.13%和1.60%。对于距离靶区较远的头颈部器官(视交叉、视神经和眼)和腹部器官(直肠、膀胱和股骨头)剂量分别约为9和2 mGy,并且NTCP均近似为0.00%。结论:研究结果证明通过自动勾画软件DeepViewer^(®)、蒙特卡罗软件ARCHER和LKB模型建立全身器官剂量数据库的可行性。

Objective To discuss the feasibility of establishing a whole-body organ dose database platform for radiotherapy.Methods The automatic segmentation software DeepViewer^(®)based on deep learning was used to delineate organs in wholebody CT of a patient with esophageal cancer,and the Monte Carlo software ARCHER accelerated by GPU was used to calculate the dose profile.Finally,the Lyman-Kutcher-Burman(LKB)model was used to calculate the normal tissue complication probability(NTCP).Results For this case,the whole-body organ dose database was established successfully using DeepViewer^(®),ARCHER and LKB model.It was found that the closer to the target area,the greater the organ dose.The heart was closest to the target area,and it was irradiated with 14.11 Gy,but its NTCP calculated by LKB model was 0.00%due to its special model parameters.The left and right lungs were irradiated with 3.19 and 1.16 Gy,respectively,but their NTCP were larger(2.13%and 1.60%,respectively).Head and neck organs(optic chiasm,optic nerve and eyes)and abdominal organs(rectum,bladder and femoral head)which are far from target area were irradiated with about 9 and 2 mGy,respectively,and their NTCP were approximately 0.00%.Conclusion The establishment of whole-body organ dose database using automatic segmentation software DeepViewer^(®),Monte Carlo softwareARCHER and LKB model is practicable.