摘要
目的探讨组织声学结构定量(ASQ)技术在肝纤维化(HF)分期诊断中的应用价值。方法收集肝脏穿刺活检明确的HF患者87例和健康志愿者50例,分别对其进行腹部常规超声及ASQ技术检查;收集相应二维及ASQ图像,导出图像后脱机分析,图像选定感兴趣区(ROI),得出χ2直方图;软件计算红色曲线众数(Redmode)、红色曲线均值(Redave)、红色曲线标准差(Redsd)、蓝色曲线众数(Bluemode)、蓝色曲线均值(Blueave)、蓝色曲线标准差(Bluesd)及红蓝曲线下面积比(FD ratio)。各参数与病理HF分期进行相关性分析,进而绘制ROC曲线并获得各期界值。结果除Bluesd外其余参数值均与HF分期呈正相关性,其中Redmode(rs=0.853,P<0.05)及Redave(rs=0.686,P<0.05)与肝纤维化分期呈正相关性。ROC曲线提示:S0 vs S1期,仅Redmode、Redave参数曲线偏左上角,曲线下面积(AUC)分别为0.763、0.624,与AUC=0.5比较,差异有统计学意义(P<0.05),灵敏度及特异度较高,可得出临界值,差异有统计学意义(P<0.05);S1 vs S2期,除Bluesd外所有参数均位于左上角,余参数AUC均大于0.5,与AUC=0.5比较,差异有统计学意义(P<0.05),灵敏度、特异度较高,可得出临界值,差异有统计学意义(P<0.05);S2 vs S3期,所有参数均偏左上角,与AUC=0.5比较,差异有统计学意义(P<0.05),灵敏度、特异度较高,可得出临界值,差异有统计学意义(P<0.05)。结论 ASQ技术是一项无创性HF检查技术,在HF的准确分期标准制定方面有较高的临床应用价值及前景。
Objective To discuss the value of diagnosis of acoustic structure quantification technology in grading liver fibrosis. Methods 50 healthy volunteers and 87 patients with liver fibrosis were examined by B-Mode ultra- sound with ASQ software. Images were collectted, ROIs in each image were selected. Then ROIs were analyzed by ASQ software offline so as to get their X^2 histograms, then, X^2 histograms were calculated and Mode, Ave, SD and FD Ratio of red and blue curves were obtained. After that, the correlation between ASQ parameters and the patho- logical results of liver fibrosis was evaluated. At last, the ROC curves were drawn and the cut-off points of each grade were calculated. Results Expect for Bluesd, the other six parameters of ASQ were positively correlated to liver fibrosis grade, in which, Redmode （ rs = 0. 853, P 〈 0. 05 ） and Redave （ rs = 0. 686, P 〈 0. 05 ） were most chosely correlated to histological fibrosis grade. In ROC curve, SO vs S1 ： only the curves of Redmode and Redave were in the upper left corner. The AUCs of each were 0. 763 and 0. 624,compared with AUC = O. 5, statistically sig- nificant differences were found between them （ P 〈 0. 05 ）. The sensitivities and specificities were relatively high. The values of cut-offs were obtained, all of them had statistical significance（P 〈 0.05）. S1 vs S2： Except for the curve of Bluesd,curves of other six parameters were in the upper left comer. The AUCs of each were all larger than 0. 5, compared with AUC = 0.5, statistical significant differences were found between them （P 〈 0.05）. The sensi- tivities and specificities were relatively high. The values of cut-offs were obtained, all of them had statistical signifi- cance（P 〈 0. 05）. S2 vs S3 ：curves of all parameters were in the upper left corner. The AUCs were all larger than 0. 5, compared with AUC = 0.5, statistical significant differences were found between them （P 〈 0. 05）. The sensi- tivities and specificities were relatively high. The values
出处
《安徽医科大学学报》
CAS
北大核心
2016年第1期118-121,共4页
Acta Universitatis Medicinalis Anhui
基金
安徽高校省级自然科学研究项目(编号:KJ2013A166)
