Department of Radiation Oncology, The University of Texas Southwestern Medical Center, Dallas, TX 75390, United States of America.
School of Software Engineering, South China University of Technology, Guangzhou, Guangdong 510006, People's Republic of China.
Phys Med Biol. 2020 Nov 20;65(23). doi: 10.1088/1361-6560/abc363.
Deep learning (DL)-based auto-segmentation has the potential for accurate organ delineation in radiotherapy applications but requires large amounts of clean labeled data to train a robust model. However, annotating medical images is extremely time-consuming and requires clinical expertise, especially for segmentation that demands voxel-wise labels. On the other hand, medical images without annotations are abundant and highly accessible. To alleviate the influence of the limited number of clean labels, we propose a weakly supervised DL training approach using deformable image registration (DIR)-based annotations, leveraging the abundance of unlabeled data. We generate pseudo-contours by utilizing DIR to propagate atlas contours onto abundant unlabeled images and train a robust DL-based segmentation model. With 10 labeled TCIA dataset and 50 unlabeled CT scans from our institution, our model achieved Dice similarity coefficient of 87.9%, 73.4%, 73.4%, 63.2% and 61.0% on mandible, left & right parotid glands and left & right submandibular glands of TCIA test set and competitive performance on our institutional clinical dataset and a third party (PDDCA) dataset. Experimental results demonstrated the proposed method outperformed traditional multi-atlas DIR methods and fully supervised limited data training and is promising for DL-based medical image segmentation application with limited annotated data.
深度学习（DL）自动分割技术在放射治疗应用中具有精确器官勾画的潜力，但需要大量干净的标注数据来训练稳健的模型。然而，医学图像的标注非常耗时，需要临床专业知识，特别是对于需要体素级标签的分割。另一方面，大量未标注的医学图像易于获取。为了减轻有限数量的干净标签的影响，我们提出了一种使用基于可变形图像配准（DIR）的标注的弱监督 DL 训练方法，利用丰富的未标注数据。我们通过利用 DIR 将图谱轮廓传播到大量未标注的图像上来生成伪轮廓，并训练稳健的基于 DL 的分割模型。在使用 10 个 TCIA 数据集的标注和来自我们机构的 50 个未标注 CT 扫描的情况下，我们的模型在 TCIA 测试集的下颌骨、左右腮腺和左右颌下腺上的 Dice 相似系数分别达到了 87.9%、73.4%、73.4%、63.2%和 61.0%，在我们机构的临床数据集和第三方（PDDCA）数据集上也具有竞争力。实验结果表明，该方法优于传统的多图谱 DIR 方法和全监督有限数据训练，对于基于 DL 的医学图像分割应用具有很大的应用潜力，在这种应用中，可用的标注数据非常有限。
