Department of Ultrasonic Diagnosis, Shenzhen Maternity and Child Healthcare Hospital, Cheeloo College of Medicine, Shandong University, Shenzhen, Guangdong, 518000, China.
Department of Ultrasonic Diagnosis, The Second Affiliated Hospital and Yuying Children's Hospital, Wenzhou Medical University, Wenzhou, Zhejiang, 325000, China.
Med Phys. 2022 Jan;49(1):382-392. doi: 10.1002/mp.15332. Epub 2021 Nov 29.
The prevalence of thyroid diseases has been increasing year by year. In this study, we established and validated a deep learning method (Cascade region-based convolutional neural network, R-CNN) based on ultrasound videos for automatic detection and segmentation of the thyroid gland and its surrounding tissues in order to reduce the workload of radiologists and improve the detection and diagnosis rate of thyroid disease.
Seventy-one patients with normal thyroid ultrasound were included. The ultrasound videos of 59 patients were used as the training dataset, the data of 12 patients were used as the validation dataset, and in addition, the data of 9 patents were used as the testing dataset. Ultrasound videos of thyroid examination, including five standard sections (left and right lobe transverse scan, central isthmus transverse scan, left and right lobe longitudinal scan), were collected from all patients. The radiologists labeled the neck tissues, including anterior cervical muscle, cricoid cartilage, trachea, thyroid gland, endothyroid vessels, carotid artery, internal jugular vein, and esophagus. A large dataset was constructed to train and test the deep learning method. The performance was evaluated using the COCO metrics AP, AP50, and AP75. We compared the Cascade R-CNN with a state-of-the-art method CenterMask in the test dataset.
We annotated 166817, 34364, and 29227 regions in training, validation and testing samples. The model could achieve a good detection performance for the thyroid left lobe, right lobe, isthmus, muscles, trachea, carotid artery, and jugular vein; the AP of these tissues were 86.5%, 87.5%, 89.1%, 96.1%, 96.6%, 97.7%, and 91.8%, respectively. In addition, the model showed good segmentation performance for the muscles, trachea, and carotid artery; the AP of these tissues were 96%, 96.6%, and 97.8%, respectively. For the left lobe, right lobe, isthmus, esophagus, and jugular vein, AP was ≥86%. However, the segmentation results for the cricoid cartilage and endothyroid vessels were not high (AP of 53.9% and 48.5%, respectively). For fair comparison, the performance of Cascade R-CNN is better than that of CenterMask for detection and segmentation tasks. The difference was statistically significant (p < 0.05).
The new method could successfully detect and segment the thyroid gland and its surrounding tissues.
甲状腺疾病的患病率逐年上升。在这项研究中，我们建立并验证了一种基于超声视频的深度学习方法（级联区域卷积神经网络，R-CNN），用于自动检测和分割甲状腺及其周围组织，以减少放射科医生的工作量，提高甲状腺疾病的检测和诊断率。
纳入 71 例甲状腺超声正常的患者。59 例患者的超声视频作为训练数据集，12 例患者的数据作为验证数据集，另外 9 例患者的数据作为测试数据集。对所有患者进行甲状腺检查的超声视频采集，包括 5 个标准切面（左右叶横切面、峡部横切面、左右叶纵切面）。放射科医生对颈部组织（包括前颈肌、环状软骨、气管、甲状腺、甲状腺内血管、颈总动脉、颈内静脉、食管）进行标注。构建了一个大型数据集来训练和测试深度学习方法。使用 COCO 指标 AP、AP50 和 AP75 评估性能。在测试数据集上，我们将 Cascade R-CNN 与最先进的方法 CenterMask 进行了比较。
我们在训练、验证和测试样本中分别标注了 166817、34364 和 29227 个区域。该模型对甲状腺左叶、右叶、峡部、肌肉、气管、颈动脉和颈内静脉的检测性能较好；这些组织的 AP 分别为 86.5%、87.5%、89.1%、96.1%、96.6%、97.7%和 91.8%。此外，该模型对肌肉、气管和颈动脉的分割性能较好；这些组织的 AP 分别为 96%、96.6%和 97.8%。对于左叶、右叶、峡部、食管和颈内静脉，AP 均≥86%。然而，环状软骨和甲状腺内血管的分割结果并不高（AP 分别为 53.9%和 48.5%）。为了公平比较，Cascade R-CNN 的性能优于 CenterMask 用于检测和分割任务。差异具有统计学意义（p<0.05）。
该新方法能够成功检测和分割甲状腺及其周围组织。
