| 基于深度学习的甲状腺结节超声图像分割的临床应用 |
| |
| 引用本文： | 广旸,何文,吴佳俊,赵明昌,张雨康,万芳. 基于深度学习的甲状腺结节超声图像分割的临床应用[J]. 中华医学超声杂志(电子版), 2022, 19(3): 206-211. DOI: 10.3877/cma.j.issn.1672-6448.2022.03.004 |
| |
| 作者姓名： | 广旸 何文 吴佳俊 赵明昌 张雨康 万芳 |
| |
| 作者单位： | 1. 100070 首都医科大学附属北京天坛医院超声科2. 214028 江苏无锡，无锡祥生医疗科技股份有限公司研发中心算法部 |
| |
| 基金项目： | 国家自然科学基金青年基金(81901744); 北京市自然科学基金(7204255) |
| |
| 摘 要： | 目的探讨深度学习方法在超声甲状腺结节分割中的效果及其临床应用价值。 方法收集2018年8月至2020年10月首都医科大学附属北京天坛医院的166例甲状腺结节患者的1044张超声图像。观察使用改进自注意力机制的Unet深度学习方法和Unet基础方法在测试数据集上的分割效果。以分割结果是否接近有多年临床经验的超声医师的手动标注作为参考标准，将改进自注意力机制的Unet和Unet基础方法对甲状腺结节的分割效果进行比较，以交并比（IoU）、戴斯（Dice）相似性系数及与超声医师对甲状腺结节的手动勾勒接近程度来评价深度学习模型对甲状腺结节分割效果及临床应用价值的性能。 结果改进自注意力机制的Unet深度学习模型对甲状腺结节分割的IoU及Dice系数分别为0.815和0.839，与Unet基础方法结果（IoU为0.788，Dice系数为0.817）相比，具有更高的IoU和Dice系数值。从分割图像可以看出，基于改进自注意力机制的Unet深度学习模型对甲状腺结节整体和边缘细节上的分割效果好于Unet基础方法，更接近于超声医师的手动勾勒结果。 结论基于自注意力机制的Unet深度学习模型在甲状腺结节分割方面有着较高的性能，可提高诊断效率，并且该方法具有一定的临床应用价值。
|
| 关 键 词： | 深度学习 超声 甲状腺结节 分割 临床应用 |
| 收稿时间： | 2022-01-04 |
Clinical application of deep learning based ultrasound segmentation of thyroid nodules |
| |
| Yang Guang,Wen He,Jiajun Wu,Mingchang Zhao,Yukang Zhang,Fang Wan. Clinical application of deep learning based ultrasound segmentation of thyroid nodules[J]. Chinese Journal of Medical Ultrasound, 2022, 19(3): 206-211. DOI: 10.3877/cma.j.issn.1672-6448.2022.03.004 |
| |
| Authors: | Yang Guang Wen He Jiajun Wu Mingchang Zhao Yukang Zhang Fang Wan |
| |
| Affiliation: | 1. Department of Ultrasound, Beijing Tiantan Hospital, Capital Medical University, Beijing, 100070, China2. Algorithm Department of the Research & Development Center, CHISON Medical Technologies Co., Ltd., Wuxi 214028, China |
| |
| Abstract: | ObjectiveTo explore the clinical application value of deep learning based ultrasonic thyroid nodule segmentation. MethodsWe selected 1044 ultrasound images of 166 thyroid patients collected from Beijing Tiantan Hospital Affiliated to Capital Medical University from August 2018 to October 2020. The segmentation effect of Unet with improved self-attention mechanism and control Unet-based method was assessed using test datasets. Whether the segmentation result is close to the manual annotation by a sonographer with many years of clinical experience was used as a reference standard, and the Unet and Unet-based methods with improved self-attention mechanism were compared for the segmentation effect of thyroid nodules, using IoU (intersection and union ratio), Dice (Dice similarity coefficient), and the degree of closeness to the manual outline of thyroid nodules by the sonographer to evaluate the performance and clinical value of the deep learning model for thyroid nodule segmentation. ResultsThe IoU and Dice coefficients of thyroid nodule segmentation by Unet with improved self-attention mechanism were 0.815 and 0.839, respectively, which were higher than those of Unet (IoU=0.788, Dice=0.817). It can also be seen from the segmented images that the Unet based on the improved self-attention mechanism had a better segmentation effect on the overall and edge details of thyroid nodules than the Unet-based method, and was closer to the manual outline results of the sonographer. ConclusionUnet based on self-attention mechanism has good performance in thyroid nodule segmentation, which can improve the diagnostic efficiency, the method also has clinical application value. |
| |
| Keywords: | Deep learning Ultrasound Thyroid nodules Segmentation Clinical application |
|
| 点击此处可从《中华医学超声杂志(电子版)》浏览原始摘要信息 |
|
点击此处可从《中华医学超声杂志(电子版)》下载全文 |
|
