深度学习重建算法在肝脏肿瘤患者CT检查中的应用进展  

Advancements and Applications of Deep Learning Image Reconstruction Algorithm in Computed Tomography Examinations for Patients with Liver Tumors

作　　者：冯俏 全江海 程留慧[2] FENG Qiao;QUAN Jianghai;CHENG Liuhui(Henan University of Chinese Medicine,Zhengzhou 450000,Henan,P.R.China;The First Affiliated Hospital of Henan University of CM,Zhengzhou 450000,Henan,P.R.China)

机构地区：[1]河南中医药大学,河南郑州450000 [2]河南中医药大学第一附属医院,河南郑州450000

出　　处：《影像科学与光化学》2024年第4期385-391,共7页Imaging Science and Photochemistry

摘　　要：肝脏肿瘤在临床较为常见,尤其是肝癌患者,是消化系统常见的恶性肿瘤之一,已严重威胁到人们的身心健康。基于卷积神经网络的深度学习重建算法是CT重建算法的一种新兴技术,逐渐趋于成熟并且在体模、临床中已有相关实践。与滤波反投影法、迭代重建算法相比,深度学习重建算法可以降低辐射剂量、实现有效降噪,优化细微结构的显示,提升主观诊断信心等。影像组学可以高通量地提取高维影像特征,深度学习重建算法可以提高影像组学特征的临床应用可信性。本文对深度学习重建算法在肝脏肿瘤中的相关研究展开综述,旨在进一步理解深度学习重建算法的临床进展,并为后续研究积累证据。The occurrence of liver tumors is prevalent in clinical practice,particularly among patients diagnosed with liver cancer.It represents one of the most frequently encountered malignant neoplasms within the digestive system,posing a significant threat to individuals’physical and psychological well-being.The deep learning reconstruction algorithm based on convolutional neural network is an emerging technology in computed tomography reconstruction algorithms,which is gradually maturing and has been applied in both phantom studies and clinical practice.In comparison to the filtered back projection and iterative reconstruction algorithm,the deep learning image reconstruction algorithm offers some advantages,such as reduced radiation dose,effective noise reduction,optimized display of fine structures and improved subjective diagnostic confidence.Radiomics can extract high-dimensional imaging features with high throughput,and the application of deep learning image reconstruction algorithm can enhance the reliability of radiomics features in clinical setting.This article provides a review of related research on deep learning image reconstruction algorithm in liver tumors,aiming to deepen our understanding of its clinical progress and accumulate evidence for future studies.

关 键 词：深度学习 肝肿瘤 重建算法 X线计算机体层摄影 

分 类 号：R735.7[医药卫生—肿瘤]