Department of Radiological Technology, Faculty of Medical Technology, Mahidol University, Bangkok, 10700, Thailand.
Johns Hopkins School of Medicine, Johns Hopkins University, Baltimore, Maryland, 21218, USA.
Med Phys. 2021 Dec;48(12):7877-7890. doi: 10.1002/mp.15303. Epub 2021 Oct 31.
Y selective internal radiation therapy (SIRT) has become a safe and effective treatment option for liver cancer. However, segmentation of target and organ-at-risks is labor-intensive and time-consuming in Y SIRT planning. In this study, we developed a convolutional neural network (CNN)-based method for automated lungs, liver, and tumor segmentation on Tc-MAA SPECT/CT images for Y SIRT planning.
Tc-MAA SPECT/CT images and corresponding clinical segmentations were retrospectively collected from 56 patients who underwent Y SIRT. The collected data were used to train three CNN-based segmentation algorithms for lungs, liver, and tumor segmentation. Segmentation performance was evaluated using the Dice similarity coefficient (DSC), surface DSC, and average symmetric surface distance (ASSD). Dosimetric parameters (volume, counts, and lung shunt fraction) were measured from the segmentation results and were compared with clinical reference segmentations.
The evaluation results show that the method can accurately segment lungs, liver, and tumor with median [interquartile range] DSCs of 0.98 [0.97-0.98], 0.91 [0.83-0.93], and 0.85 [0.71-0.88]; surface DSCs of 0.99 [0.97-0.99], 0.86 [0.77-0.93], and 0.85 [0.62-0.93], and ASSDs of 0.91 [0.69-1.5], 4.8 [2.6-8.4], and 4.7 [3.5-9.2] mm, respectively. Dosimetric parameters from the three segmentation networks show relationship with those from the reference segmentations. The overall segmentation took about 1 min per patient on an NVIDIA RTX-2080Ti GPU.
This work presents CNN-based algorithms to segment lungs, liver, and tumor from Tc-MAA SPECT/CT images. The results demonstrated the potential of the proposed CNN-based segmentation method for assisting Y SIRT planning while drastically reducing operator time.
Y 选择性内放射治疗（SIRT）已成为肝癌安全有效的治疗选择。然而，Y SIRT 计划中的目标和危险器官的分割是劳动密集型且耗时的。在这项研究中，我们开发了一种基于卷积神经网络（CNN）的方法，用于在 Y SIRT 计划中对 Tc-MAA SPECT/CT 图像进行自动肺部、肝脏和肿瘤分割。
回顾性收集了 56 例接受 Y SIRT 的患者的 Tc-MAA SPECT/CT 图像和相应的临床分割。所收集的数据用于训练三个基于 CNN 的分割算法，用于肺部、肝脏和肿瘤分割。使用 Dice 相似系数（DSC）、表面 DSC 和平均对称表面距离（ASSD）评估分割性能。从分割结果中测量了剂量学参数（体积、计数和肺分流分数），并与临床参考分割进行了比较。
评估结果表明，该方法可以准确地分割肺部、肝脏和肿瘤，中位数[四分位距] DSCs 分别为 0.98[0.97-0.98]、0.91[0.83-0.93]和 0.85[0.71-0.88]；表面 DSCs 分别为 0.99[0.97-0.99]、0.86[0.77-0.93]和 0.85[0.62-0.93]，ASSD 分别为 0.91[0.69-1.5]、4.8[2.6-8.4]和 4.7[3.5-9.2]mm。三个分割网络的剂量学参数与参考分割的参数具有相关性。在 NVIDIA RTX-2080Ti GPU 上，每位患者的整体分割时间约为 1 分钟。
本研究提出了基于 CNN 的算法，用于从 Tc-MAA SPECT/CT 图像中分割肺部、肝脏和肿瘤。结果表明，所提出的基于 CNN 的分割方法具有辅助 Y SIRT 计划的潜力，同时大大减少了操作人员的时间。
