人工智能影像分析技术在降低头颈CT血管造影辐射剂量中的应用价值  

Value of artificial intelligence image analysis in reducing radiation dose of head and neck CT angiography

作　　者：罗沛霖[1] 李伟[1] 洪璇阳 钟文 戴伟创 LUO Peilin;LI Wei;HONG Xuanyang;ZHONG Wen;DAI Weichuang(Guangdong Armed Police Corps Hospital,Guangzhou 510000,China;Cancer Hospital Affiliated to Guangzhou Medical University,Guangzhou 510000,China)

机构地区：[1]武警广东总队医院,广东广州510000 [2]广州医科大学附属肿瘤医院,广东广州510000

出　　处：《临床医药实践》2025年第5期363-366,共4页Proceeding of Clinical Medicine

摘　　要：目的:探讨人工智能影像分析技术在不同辐射剂量头颈CT血管造影中的具体表现及应用价值,探索合适的低辐射剂量扫描方案。方法:选取2022年4月-2023年8月武警广东总队医院拟诊断为脑血管病变的患者90例,进行头颈CT血管造影检查。按照随机数字表法把患者分为A,B和C组,每组30例。各组的管电压均为80 kV,A组辐射剂量为每层15 mGy,B组为每层10 mGy,C组为每层5 mGy。对图像进行客观及主观分析,客观分析包括测量各主要血管的CT值、本底噪声(SD)、信噪比(SNR)及对比度噪声(CNR),主观分析包括对人工智能处理后的头颈血管的容积再现图像、最大密度投影图像及轴位图像进行评分。结果:三组头颈血管内CT值比较,差异无统计学意义(P>0.05)。A组SD及SNR与B组比较差异无统计学意义(P>0.05);C组SD及SNR与A组和B组比较差异有统计学意义(P<0.01)。三组SD和SNR比较差异有统计学意义(P<0.01);三组CNR比较差异无统计学意义(P>0.05)。三组对头颈血管的解剖结构显示以及颅底骨质的去除情况(平均分)比较,差异无统计学意义(P>0.05)。三组血管结构均能显示,图像质量主观评分均为优良以上。C组比A组和B组辐射剂量分别降低了66.6%和33.3%。结论:通过人工智能影像分析技术,能够在极低辐射剂量条件下较好地完成对头颈血管的解剖结构三维显示,适用于需要低剂量头颈CT血管造影检查或灌注成像的患者。Objective:To study the specific manifestations and application value of artificial intelligence image analysis technology in head and neck CT angiography with different radiation doses,explore appropriate low-radiation-dose scanning protocols.Methods:Ninety patients who were suspected of having cerebrovascular diseases from April 2022 to August 2023 were selected to undergo head and neck CT angiography.These patients were divided into three groups of 30 cases each according to the random number table method.The tube voltage of each group was 80 kV.The radiation dose(CTDI)in group A was 15 mGy per slice(CTDI),10 mGy per slice in group B,and 5 mGy per slice in group C.Both objective and subjective analyses were performed on the images.The objective analysis included measuring the CT values of major blood vessels,background noise(SD),signal-to-noise ratio(SNR),and contrast-to-noise ratio(CNR).The subjective analysis involved scoring the VR,MIP processing,and axial images after being processed by artificial intelligence image analysis technology.Results:There was no significant difference in the mean value of intravascular CT among the three groups(P>0.05).There was no significant difference in SD and SNR between group A and group B(P>0.05),but there was significant difference in SD and SNR between group C and group A and B(P<0.01).There was statistical difference in SD and SNR among the three groups(P<0.01),but there was no statistical difference in CNR among the three groups(P>0.05).There was no significant difference between the three groups in displaying the anatomical structure of the head and neck vessels and removing the skull base bone(P>0.05).The vascular structure of each group could be displayed,and the subjective scores of image quality were excellent or above.The radiation dose of group C was reduced by 66.6%and 33.3%respectively compared with that of group A and group B.Conclusion:Through artificial intelligence image analysis technology,it is possible to achieve a better three-dimensional display of th

关 键 词：体层摄影术 X线计算机 血管造影 头部 颈部 人工智能 影像分析 

分 类 号：R445.3[医药卫生—影像医学与核医学]