摘要
MRI作为乳腺疾病的重要影像检查方法之一,在乳腺癌早期检出和疗效预测中具有重要意义。目前,乳腺癌肿瘤大小的评估主要依据二维图像中包含的肿瘤直径、形态等信息,存在一定局限性,可重复性较低,预测准确性有待进一步提高。基于MRI动态增强测量肿瘤的体积等三维信息可以为乳腺癌病程的判断及新辅助化疗疗效的评估提供重要依据。目前,对乳腺癌肿瘤三维信息的测量依赖影像医师的经验且耗时较长。为提升测量准确度并降低时间成本,人工智能技术在乳腺MRI领域有着广阔的研究前景。鉴于此,本文就人工智能特别是深度学习技术在自动测量乳腺癌肿瘤体积中的研究和应用情况进行梳理,主要涉及图像分割、形态三维重建、可视化和容积测量几方面。本文为临床医生深入了解人工智能技术如何帮助乳腺肿瘤的自动化高精度测量提供精准材料,为信息技术人员理解如何将人工智能技术应用于乳腺肿瘤测量提供思路。
As one of the important imaging methods for breast diseases,MRI is of great significance in the early detection of breast cancer and prediction of outcome.At present,the assessment of breast cancer tumor size is mainly based on the information of tumor diameter and morphology contained in two-dimensional images,which has certain limitations and low reproducibility,and the prediction accuracy needs to be further improved.Based on dynamic contrast enhanced MRI measurement of tumor volume and other three-dimensional information can provide an important basis for determining the course of breast cancer and evaluating the efficacy of neoadjuvant chemotherapy.However,the measurement of 3D information of breast cancer tumors relies on the experience of the physicians and takes a long time.To improve the measurement accuracy and reduce the time cost,artificial intelligence technology has a promising research prospect in the field of breast MRI.In view of this,we compared the research and applications of artificial intelligence,especially deep learning techniques,in automated breast cancer tumor volume measurement,mainly in the areas of image segmentation,morphological 3D reconstruction,visualization and volume measurement.This paper provides precise material for clinicians to gain insight into how AI techniques can help in automated and high-precision measurement of breast tumors,and provides ideas for information technology personnel to understand how AI techniques can be applied to breast tumor measurement.
作者
徐京瑶
刘晓民
张新峰
郭伟
王飞
李相生
XU Jingyao;LIU Xiaomin;ZHANG Xinfeng;GUO Wei;WANG Fei;LI Xiangsheng(Faculty of Information Technology,Beijing University of Technology,Beijing 100124,China;Department of Medical Imaging,Air Force Medical Center,People Military Army,Beijing 100142,China)
出处
《磁共振成像》
CAS
CSCD
北大核心
2023年第9期148-153,共6页
Chinese Journal of Magnetic Resonance Imaging
关键词
乳腺癌
乳腺肿瘤体积
深度学习
图像分割
三维重建技术
磁共振成像
breast cancer
breast tumor volume
deep learning
image segmentation
3D reconstruction techniques
magnetic resonance imaging
