摘要
目的探究人工智能迭代重建(AIIR)算法在腹部低剂量增强CT中的临床应用价值。方法前瞻性纳入2023年1-8月于我院行腹部增强CT检查的患者31例,男/女(24/7);平均年龄(60.5±10.8)岁。同时采集常规剂量(100 KVp,采用管电流调制技术,剂量等级3级,参考管电流213 mAs)和低剂量(80 KVp,采用管电流调制技术,剂量等级1级,参考管电流47 mAs)门静脉期腹部增强CT图像并记录辐射剂量参数。常规剂量(SD)CT采用迭代(Karl)重建,低剂量(RD)CT采用Karl和AIIR算法1、3、5级强度重建。测量5组图像门静脉、肝实质、肌肉、脂肪的CT值、噪声指数、信噪比(SNR)和对比度噪声比(CNR),主观评价5组图像的整体图像质量和“蜡质感”。选择长径大于5 mm的肝脏病变共61个,测量病变的SNR及CNR并进行病变显著性主观评价。结果SD腹部增强CT的有效剂量(ED)为(7.06±2.50)mSv,RD腹部增强CT的ED为(1.18±0.54)mSv,RD组辐射剂量较SD组降低约83.3%。5组图像中肝实质、肌肉、脂肪的CT值无差异(P>0.05),门静脉CT值在SD Karl组与RDKarl组之间、RDKarl组与RDAIIR各组之间存在差异(P<0.005)。RDAIIR重建图像的噪声指数、SNR和CNR值均显著优于RD和SD Karl重建图像(P<0.005),RD AIIR 5级重建在整体图像质量和“蜡质感”方面与SD Karl重建相当,符合临床诊断要求。RDAIIR各组病变的SNR和CNR值均显著高于RDKarl组(P<0.005),RD AIIR 5组病变显著性主观评分与SD Karl组相当,高于RDKarl、RD AIIR 1、RD AIIR 3组(P<0.005)。结论AIIR算法可有效降低腹部低剂量增强CT的图像噪声,改善图像质量,提高病变显著性。
Objective To explore the clinical application value of artificial intelligence iterative reconstruction(AIIR)algorithm in low-dose abdominal contrast-enhanced CT.Methods A total of 31 patients(Male/female,24/7;Average Age,60.5±10.8 years)who underwent abdominal contrast-enhanced CT examination in our hospital from January to August 2023 were prospectively enrolled.Conventional dose(100 KVp,tube current modulation technology,dose level 3,reference tube current 213 mAs)and low-dose(80 KVp,tube current modulation technology,dose level 1,reference tube current 47 mAs)abdominal contrast-enhanced CT images of portal venous phase were collected and radiation dose parameters were recorded at the same time.Conventional dose(SD)CT was reconstructed by Karl iterative reconstruction and low-dose(RD)CT was reconstructed by Karl and AIIR algorithm with grade 1,3,and 5.CT values,noise index,signal to noise ratio(SNR)and contrast to noise ratio(CNR)of portal vein,liver,muscle and fat of the five groups of images were measured.The overall image quality and“wax texture”of the five groups of images were subjectively evaluated.A total of 61 liver lesions with a long diameter greater than 5 mm were selected,the SNR and CNR of the lesions were measured and the significance of the lesions was subjectively evaluated.Results The effective dose(ED)of the SD group was(7.06±2.50)mSv and the ED of the RD group was(1.18±0.54)mSv,the radiation dose of the RD group was reduced by about 83.3%compared with the SD group.There were no significant differences in CT values of liver,muscle and fat among the five groups(P>0.05).The CT values of portal vein were significantly different between SD Karl group and RD Karl group,as well as between RD Karl group and RD AIIR groups(P<0.005).The SD,SNR and CNR values of RD AIIR reconstruction images were significantly better than those of RD and SD Karl reconstruction image(P<0.005).The RD AIIR 5 reconstruction was equivalent to the SD Karl reconstruction in terms of the overall image quality and“wax
作者
张宝平
李傲
李宇航
朱书萌
田倩
赵文哲
杨健
金超
Zhang Baoping;Li Ao;Li Yuhang;Zhu Shumeng;Tian Qian;Zhao Wenzhe;Yang Jian;Jin Chao(Department of Radiology,The First Affiliated Hospital of Xi’an Jiaotong University,Xi’an Shaanxi 710061,China;Shaanxi Engineering Research Center of Computational Imaging and Medical Intelligence,Xi’an Shaanxi 710061,China)
出处
《遵义医科大学学报》
2024年第5期487-494,501,共9页
Journal of Zunyi Medical University
基金
国家自然科学基金数学天元重点专项(NO:12226007)。
关键词
腹部增强CT
低剂量CT
人工智能迭代重建
肝脏
abdominal contrast-enhanced CT
low-dose CT
artificial intelligence iterative reconstruction
liver
