Aoyama Takahiro, Shimizu Hidetoshi, Koide Yutaro, Kamezawa Hidemi, Fukunaga Jun-Ichi, Kitagawa Tomoki, Tachibana Hiroyuki, Suzuki Kojiro, Kodaira Takeshi
Department of Radiation Oncology, Aichi Cancer Center, Nagoya, Japan.
Division of Radiological Sciences, Graduate School of Health Sciences, Teikyo University, Fukuoka, Japan.
J Med Phys. 2024 Jan-Mar;49(1):33-40. doi: 10.4103/jmp.jmp_122_23. Epub 2024 Mar 30.
This study aimed to develop a deep learning model for the prediction of V (the volume of the lung parenchyma that received ≥20 Gy) during intensity-modulated radiation therapy using chest X-ray images.
The study utilized 91 chest X-ray images of patients with lung cancer acquired routinely during the admission workup. The prescription dose for the planning target volume was 60 Gy in 30 fractions. A convolutional neural network-based regression model was developed to predict V. To evaluate model performance, the coefficient of determination ), root mean square error (RMSE), and mean absolute error (MAE) were calculated with conducting a four-fold cross-validation method. The patient characteristics of the eligible data were treatment period (2018-2022) and V (19.3%; 4.9%-30.7%).
The predictive results of the developed model for V were 0.16, 5.4%, and 4.5% for the , RMSE, and MAE, respectively. The median error was -1.8% (range, -13.0% to 9.2%). The Pearson correlation coefficient between the calculated and predicted V values was 0.40. As a binary classifier with V <20%, the model showed a sensitivity of 75.0%, specificity of 82.6%, diagnostic accuracy of 80.6%, and area under the receiver operator characteristic curve of 0.79.
The proposed deep learning chest X-ray model can predict V and play an important role in the early determination of patient treatment strategies.
本研究旨在开发一种深度学习模型，用于利用胸部X光图像预测调强放射治疗期间肺实质接受≥20 Gy剂量的体积（V）。
本研究使用了91例肺癌患者在入院检查期间常规获取的胸部X光图像。计划靶体积的处方剂量为60 Gy，分30次给予。开发了一种基于卷积神经网络的回归模型来预测V。为评估模型性能，采用四重交叉验证方法计算决定系数（ ）、均方根误差（RMSE）和平均绝对误差（MAE）。符合条件数据的患者特征为治疗时间（2018 - 2022年）和V（19.3%；4.9% - 30.7%）。
所开发模型对V的预测结果，决定系数为0.16，RMSE为5.4%，MAE为4.5%。中位误差为 - 1.8%（范围： - 13.0%至9.2%）。计算得到的V值与预测的V值之间的Pearson相关系数为0.40。作为V <20%的二元分类器，该模型的灵敏度为75.0%，特异性为82.6%，诊断准确率为80.6%，受试者操作特征曲线下面积为0.79。
所提出的深度学习胸部X光模型可预测V，并在患者治疗策略的早期确定中发挥重要作用。
