LGEANet：用于呼吸运动预测的 LSTM-全局时间卷积-外部注意网络。

School of Biomedical Engineering, Southern Medical University, Guangdong, Guangzhou, China.

Guangdong Provincial Key Laboratory of Medical Image Processing, Southern Medical University, Guangdong, Guangzhou, China.

Med Phys. 2023 Apr;50(4):1975-1989. doi: 10.1002/mp.16237. Epub 2023 Feb 13.

PURPOSE

To develop a deep learning network that treats the three-dimensional respiratory motion signals as a whole and considers the inter-dimensional correlation between signals of different directions for accurate respiratory tumor motion prediction.

METHODS

We propose a deep learning framework, named as LSTM-Global Temporal Convolution-External Attention Network (LGEANet). In LGEANet, we first feed each of the univariate time series into the Long Short-Term Memory (LSTM) module respectively and utilize the strength of the global temporal convolutional layer to discover the temporal pattern of the univariate signals from hidden states of the LSTM. Then, External attention is adopted to capture the dynamic dependence of the multiple time series. Also, a traditional autoregressive linear model in parallel to the non-linear neural network part was integrated to mitigate the scale insensitivity of the networks. A total of 304 motion traces for 31 patients are acquired from a public dataset in the experiments and four representative cases were selected for model evaluation. The respiratory signals were sampled at intervals of about 37.5 ms (26 frames per second) for an average duration of 71 min.

RESULTS

The proposed LGEANet achieved better performance with higher empirical correlation coefficient value (CORRs) and lower mean absolute error value (MAEs) and relative squared error value (RSEs) than other investigated models. For the four representative datasets, when the response time is less than 231 ms, the model can achieve CORRs more than 0.96. And the averaged position error reduction by using the proposed model was about 67% in the superior-inferior (SI) direction, 41% in the anterior-posterior (AP) direction and 38% in the right-left (RL) direction compared to that without prediction. The proposed network achieved the greatest error reduction in the SI direction, which is the main direction of tumor motion.

CONCLUSIONS

The LGEANet achieves promising performance in minimizing the prediction error due to system latencies during real-time tumor motion tracking.

目的

开发一种深度学习网络，将三维呼吸运动信号作为一个整体进行处理，并考虑不同方向信号之间的多维相关性，以实现准确的呼吸肿瘤运动预测。

方法

我们提出了一种名为 LSTM-Global Temporal Convolution-External Attention Network（LGEANet）的深度学习框架。在 LGEANet 中，我们首先将每个单变量时间序列分别输入到长短期记忆（LSTM）模块中，并利用全局时间卷积层的强度从 LSTM 的隐藏状态中发现单变量信号的时间模式。然后，采用外部注意力来捕获多个时间序列的动态依赖性。此外，在非线性神经网络部分的平行部分还集成了一个传统的自回归线性模型，以减轻网络的尺度不敏感性。在实验中，我们从一个公共数据集获得了 31 名患者的 304 个运动轨迹，选择了四个有代表性的病例进行模型评估。呼吸信号以大约 37.5ms（每秒 26 帧）的间隔进行采样，平均持续时间为 71 分钟。

结果

与其他研究模型相比，所提出的 LGEANet 具有更高的经验相关系数值（CORRs）和更低的平均绝对误差值（MAEs）和相对平方误差值（RSEs），因此具有更好的性能。对于四个有代表性的数据集，当响应时间小于 231ms 时，该模型可以实现 CORRs 大于 0.96。与没有预测的情况相比，使用所提出的模型在上下（SI）方向上的平均位置误差减少了约 67%，在前后（AP）方向上减少了 41%，在左右（RL）方向上减少了 38%。在所提出的网络中，在 SI 方向上实现了最大的误差减少，这是肿瘤运动的主要方向。