基于NCCT影像组学预测自发性脑出血血肿扩大的研究进展  

Advances in predicting hematoma expansion in spontaneous cerebral hemorrhage based on NCCT radiomics

作　　者：张泽鑫 谢召勇[2] ZHANG Zexin;XIE Zhaoyong(Chifeng Clinical Medical College of Inner Mongolia Medical University,Chifeng 024000,China;Chifeng Hospital Medical image center,Chifeng 024000,China)

机构地区：[1]内蒙古医科大学赤峰临床医学院,内蒙古赤峰024000 [2]内蒙古赤峰市医院医学影像中心,内蒙古赤峰024000

出　　处：《医学影像学杂志》2025年第3期122-125,共4页Journal of Medical Imaging

摘　　要：自发性脑出血血肿扩张极大影响患者的生存及预后,已被视为一个潜在的治疗靶点。尽管目前提出的诸多检查技术和影像征象都能提示血肿扩张,但均存在局限性,因此需要发展新技术预测血肿扩张以期参与临床决策。人工智能与医学图像相结合的方法使得影像组学技术蓬勃发展。CT平扫影像组学不仅能够利用脑出血患者急性期CT检查图像,在不增加额外辐射剂量和费用的情况下,通过建立多样化模型为预测血肿扩张提供了更客观准确快速的方法。深度学习作为机器学习最有效的算法所拥有的强大运算能力及高效的输出结果,有望成为现代医学研究者预测脑出血血肿扩张最有潜力的工具。Hematoma expansion in spontaneous cerebral hemorrhage greatly affects patient survival and prognosis,which has been recognized as a potential therapeutic target.Although modern medical examination techniques and many imaging signs can indicate hematoma expansion,they all have many shortcomings.Therefore,we need to develop new techniques to predict hema-toma expansion in order to participate in clinical decision-making.The combination of artificial intelligence and medical imaging has made the rapid development of radiomics.NCCT radiomics not only uses acute phase CT images of patients with cerebral hemorrhage,but also provides a more objective,accurate and rapid method for predicting hematoma expansion by establishing diversified models without additional radiation dose and cost.The powerful computing power and efficient output of deep learning as the most effective algorithm for machine learning are expected to become the most promising tool for modern medical research-ers to predict hematoma expansion in cerebral hemorrhage.

关 键 词：自发性脑出血 血肿扩张 非增强体层计算机成像 影像组学 

分 类 号：R743.34[医药卫生—神经病学与精神病学] R445[医药卫生—临床医学]