摘要
骨肿瘤种类繁多,分类复杂,且部分肿瘤影像学表现相似,如一些良性骨肿瘤可能表现出类似恶性骨肿瘤的侵袭性特征,使得基于传统影像学的骨肿瘤诊断和鉴别诊断面临很大挑战。近年来,影像组学和人工智能技术的快速发展,大大提高了病变图像识别能力,广泛用于各种肿瘤的分类和表征。文章主要就影像组学在骨肿瘤诊断、预后预测、基因分析预测方面的研究进展进行综述,并提出目前所面临的挑战及未来发展方向。
Bone tumors have a wide variety and complex classification,even some of them have similar imaging manifestations.For example,a few benign bone tumors may show invasive characteristics similar to the malignant bone tumors,which makes the diagnosis and differential diagnosis of bone tumors based on traditional imaging facing great challenges.In recent years,the rapid development of radiomics and artificial intelligence(AI)technology has greatly improved the ability of lesion image recognition,which is widely used in the classification and characterization of various tumors.This article mainly reviews the advances in radiomics of bone tumor diagnosis,prognosis prediction and genetic analysis and prediction,and puts forward the current challenges and future development direction.
作者
杨震(综述)
杨明
戚荣丰(审校)
YANG Zhen;YANG Ming;QI Rongfeng(Department of Ultrasound,Nanjing Jinling Hospital,Affiliated Hospital of Medical School,Nanjing University/General Hospital of Eastern Theater Command,PLA,Nanjing 210002,Jiangsu,China;Department of Medical Imaging,Nanjing Children's Hospital,Nanjing 210000,Jiangsu,China;Department of Radiology,Drum Tower Hospital,Clinical College of Nanjing Medical University,Nanjing 210008,Jiangsu,China)
出处
《医学研究与战创伤救治》
CAS
北大核心
2024年第5期538-542,共5页
Journal of Medical Research & Combat Trauma Care
基金
江苏省自然科学基金(BK20221554)
南京师范大学“心理健康与认知科学”省重点实验室开放课题(23MHCSKF01)。
关键词
骨肿瘤
影像组学
特征提取
机器学习
bone tumor
radiomics
feature extraction
machine learning
