Department of Radiation Oncology, Hubei Cancer Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430079, China; Department of Radiation Oncology, The First Affiliated Hospital of USTC, Division of Life Sciences and Medicine, University of Science and Technology of China, Hefei 230001, China.
School of Computer Science and Technology, University of Science and Technology of China, Hefei, Anhui 230026, China.
Phys Med. 2024 Jul;123:103393. doi: 10.1016/j.ejmp.2024.103393. Epub 2024 Jun 8.
One of the current roadblocks to the widespread use of Total Marrow Irradiation (TMI) and Total Marrow and Lymphoid Irradiation (TMLI) is the challenging difficulties in tumor target contouring workflow. This study aims to develop a hybrid neural network model that promotes accurate, automatic, and rapid segmentation of multi-class clinical target volumes.
Patients who underwent TMI and TMLI from January 2018 to May 2022 were included. Two independent oncologists manually contoured eight target volumes for patients on CT images. A novel Dual-Encoder Alignment Network (DEA-Net) was developed and trained using 46 patients from one internal institution and independently evaluated on a total of 39 internal and external patients. Performance was evaluated on accuracy metrics and delineation time.
The DEA-Net achieved a mean dice similarity coefficient of 90.1 % ± 1.8 % for internal testing dataset (23 patients) and 91.1 % ± 2.5 % for external testing dataset (16 patients). The 95 % Hausdorff distance and average symmetric surface distance were 2.04 ± 0.62 mm and 0.57 ± 0.11 mm for internal testing dataset, and 2.17 ± 0.68 mm, and 0.57 ± 0.20 mm for external testing dataset, respectively, outperforming most of existing state-of-the-art methods. In addition, the automatic segmentation workflow reduced delineation time by 98 % compared to the conventional manual contouring process (mean 173 ± 29 s vs. 12168 ± 1690 s; P < 0.001). Ablation study validate the effectiveness of hybrid structures.
The proposed deep learning framework achieved comparable or superior target volume delineation accuracy, significantly accelerating the radiotherapy planning process.
全骨髓照射（TMI）和全骨髓及淋巴照射（TMLI）广泛应用的当前障碍之一是肿瘤靶区勾画工作流程的挑战性困难。本研究旨在开发一种混合神经网络模型，以促进多类临床靶区的准确、自动和快速分割。
纳入 2018 年 1 月至 2022 年 5 月期间行 TMI 和 TMLI 的患者。两名独立的肿瘤学家在 CT 图像上手动勾画 8 个靶区。使用来自一个内部机构的 46 名患者开发并训练了一种新的双编码器对齐网络（DEA-Net），并在总共 39 名内部和外部患者上进行了独立评估。使用准确性指标和勾画时间评估性能。
DEA-Net 在内部测试数据集（23 名患者）中获得了 90.1%±1.8%的平均骰子相似系数，在外部测试数据集（16 名患者）中获得了 91.1%±2.5%的平均骰子相似系数。内部测试数据集的 95%Hausdorff 距离和平均对称表面距离分别为 2.04±0.62mm 和 0.57±0.11mm，外部测试数据集分别为 2.17±0.68mm 和 0.57±0.20mm，优于大多数现有的最先进方法。此外，与传统的手动勾画过程相比，自动勾画工作流程将勾画时间减少了 98%（平均 173±29s 与 12168±1690s；P<0.001）。消融研究验证了混合结构的有效性。
所提出的深度学习框架实现了可比或更高的靶区勾画精度，显著加速了放射治疗计划过程。
