Liu Yaoying, Chen Zhaocai, Wang Jinyuan, Wang Xiaoshen, Qu Baolin, Ma Lin, Zhao Wei, Zhang Gaolong, Xu Shouping
Department of Radiation Oncology, the First Medical Center of the People's Liberation Army General Hospital, Beijing, China.
School of Physics, Beihang University, Beijing, China.
Front Oncol. 2021 Nov 11;11:752007. doi: 10.3389/fonc.2021.752007. eCollection 2021.
This study focused on predicting 3D dose distribution at high precision and generated the prediction methods for nasopharyngeal carcinoma patients (NPC) treated with Tomotherapy based on the patient-specific gap between organs at risk (OARs) and planning target volumes (PTVs).
A convolutional neural network (CNN) is trained using the CT and contour masks as the input and dose distributions as output. The CNN is based on the "3D Dense-U-Net", which combines the U-Net and the Dense-Net. To evaluate the model, we retrospectively used 124 NPC patients treated with Tomotherapy, in which 96 and 28 patients were randomly split and used for model training and test, respectively. We performed comparison studies using different training matrix shapes and dimensions for the CNN models, i.e., 128 ×128 ×48 (for Model I), 128 ×128 ×16 (for Model II), and 2D Dense U-Net (for Model III). The performance of these models was quantitatively evaluated using clinically relevant metrics and statistical analysis.
We found a more considerable height of the training patch size yields a better model outcome. The study calculated the corresponding errors by comparing the predicted dose with the ground truth. The mean deviations from the mean and maximum doses of PTVs and OARs were 2.42 and 2.93%. Error for the maximum dose of right optic nerves in Model I was 4.87 ± 6.88%, compared with 7.9 ± 6.8% in Model II (=0.08) and 13.85 ± 10.97% in Model III (<0.01); the Model I performed the best. The gamma passing rates of PTV for 3%/3 mm criteria was 83.6 ± 5.2% in Model I, compared with 75.9 ± 5.5% in Model II (<0.001) and 77.2 ± 7.3% in Model III (<0.01); the Model I also gave the best outcome. The prediction error of D for PTV was 0.64 ± 0.68% in Model I, compared with 2.04 ± 1.38% in Model II (<0.01) and 1.05 ± 0.96% in Model III (=0.01); the Model I was also the best one.
It is significant to train the dose prediction model by exploiting deep-learning techniques with various clinical logic concepts. Increasing the height (Y direction) of training patch size can improve the dose prediction accuracy of tiny OARs and the whole body. Our dose prediction network model provides a clinically acceptable result and a training strategy for a dose prediction model. It should be helpful to build automatic Tomotherapy planning.
本研究聚焦于高精度预测三维剂量分布，并基于鼻咽癌（NPC）患者危及器官（OARs）与计划靶区（PTVs）之间的个体差异，生成了螺旋断层放疗治疗的鼻咽癌患者的预测方法。
使用CT图像和轮廓掩码作为输入，剂量分布作为输出，训练卷积神经网络（CNN）。该CNN基于“3D密集U-Net”，它结合了U-Net和密集连接网络（Dense-Net）。为评估该模型，我们回顾性地使用了124例接受螺旋断层放疗的鼻咽癌患者，其中96例和28例患者分别被随机拆分用于模型训练和测试。我们对CNN模型使用不同的训练矩阵形状和维度进行了比较研究，即128×128×48（模型I）、128×128×16（模型II）和二维密集U-Net（模型III）。使用临床相关指标和统计分析对这些模型的性能进行了定量评估。
我们发现训练补丁大小的高度越大，模型结果越好。该研究通过将预测剂量与真实剂量进行比较来计算相应误差。PTV和OARs的平均剂量和最大剂量的平均偏差分别为2.42%和2.93%。模型I中右侧视神经最大剂量的误差为4.87±6.88%，而模型II为7.9±6.8%（P = 0.08），模型III为13.85±10.97%（P<0.01）；模型I表现最佳。模型I中PTV在3%/3 mm标准下的伽马通过率为83.6±5.2%，模型II为75.9±5.5%（P<0.001），模型III为77.2±7.3%（P<0.01）；模型I同样给出了最佳结果。模型I中PTV的剂量预测误差D为0.64±0.68%，模型II为2.04±1.38%（P<0.01），模型III为1.05±0.96%（P = 0.01）；模型I也是最佳的。
利用具有各种临床逻辑概念的深度学习技术训练剂量预测模型具有重要意义。增加训练补丁大小的高度（Y方向）可以提高微小OARs和全身的剂量预测准确性。我们的剂量预测网络模型提供了临床可接受的结果以及剂量预测模型的训练策略。这对于建立自动螺旋断层放疗计划应该是有帮助的。
