摘要
肿瘤诊治策略需要整合影像学及多种临床数据.尽管医学检测技术及治疗手段有较大进步,由于胰腺肿瘤的多样性、患者间个体差异及对治疗的反应不同,对胰腺肿瘤的诊断、表征和监测仍存在巨大挑战.影像学是临床上肿瘤评估最常用手段,其主要依赖于医师对医学影像图像的视觉判断,而影像图像的解释又可以通过计算机分析来加强.未来人工智能(AI)有望在胰腺肿瘤定性解释方面取得突破,包括监测肿瘤随时间的进展,判断肿瘤病理、基因和生物学行为,预测临床预后等.AI也可以改变影像检查及分析的工作流程,提高工作效率.
The diagnosis and treatment strategy of tumor depends on the integration of imaging and multiple clinical data.Although medical detection techniques and treatment methods have made great progress,the diagnosis,characterization and evaluation of pancreatic tumors still have great challenges due to the heterogeneity of pancreatic tumors,individual differences among patients,and different responses to treatment.Imaging is the most commonly used method in the clinical tumor evaluation.It mainly depends on the doctor's visual judgment of medical images.The interpretation of images can be strengthened by computer analysis.In future,artificial intelligence(AI)is expected to make breakthroughs in the qualitative interpretation of pancreatic tumors,including screening tumors over time,determining tumor pathology,genes,and biological behavior,and predicting clinical prognosis.AI can also change the workflow of image examination and analysis to improve work efficiency.
作者
马超
边云
陆建平
Ma Chao;Bian Yun;Lu Jianping(Department of Radiology,Changhai Hospital,Navy Medical University,Shanghai 200433,China)
出处
《中华胰腺病杂志》
CAS
2019年第5期343-346,共4页
Chinese Journal of Pancreatology
基金
国家自然科学基金(81601468,81701689)
第二军医大学精准医学转化应用研究专项(2017JZ42)
上海市科技创新行动计划(17411952200)
国家临床重点专科军队建设项目(医学影像科).
关键词
人工智能
胰腺肿瘤
影像学
影像组学
深度学习
Artificial intelligence
Pancreatic neoplasms
Imaging
Radiomics
Deep learning
