不同感兴趣区选位对健康实验兔肾超声造影效果分析的影响

目的探讨不同感兴趣区（ROI）的选位对健康实验兔肾超声造影分析效果的影响,并分析其变化规律,探索其最佳ROI取样部位。 方法采用GE LOGIQ-E9超声诊断仪,9L-D探头,频率6~9 MHz。用低机械指数实时谐波超声造影观察20只健康实验兔。使用TIC-Analysis（GE）软件包,选择直径3 mm圆形的ROI取样框,分别置于肾皮质外层、肾皮质深层、肾髓质、肾门部主肾动脉4个不同部位,进行肾脏血流灌注定量分析,测量超声造影灌注参数值到达时间（AT）、达峰时间（TTP）、峰值强度（PI）、曲线下面积（AUC）。分析左、右肾同一取样部位ROI定量结果,比较左、右肾血流灌注的差异。分析同侧肾脏不同取样部位ROI定量结果,比较肾内不同位置对定量结果的影响。采用组内相关系数（ICC）及Bland-Altman法分析ROI置于右肾皮质外层及深层2名观察者定量结果,比较2个ROI取样部位在2名观察者间其测量一致性的优劣。 结果所有健康实验兔均顺利完成超声造影检查,同一肾脏不同ROI取样部位所生成的时间-强度曲线（TIC）曲线形态相似。左、右肾同一ROI设置所测定量结果差异均无统计学意义（P值均>0.05）。右肾相同形态ROI取样框置于不同位置所测超声造影定量参数AT、TTP、PI、AUC差异有统计学意义（AT：F=5.61,P<0.05;TTP：F=4.28,P<0.05;PI：F=145.88,P<0.05;AUC: F=216.72,P<0.05）。行两两比较SNK-q检验,取样框置于肾皮质外层及肾皮质深层所测定量参数比较,PI差异有统计学意义（q=4.43,P<0.05）;取样框置于肾皮质外层及肾髓质所测定量参数比较,PI及AUC差异有统计学意义（q=12.69,P<0.05;q=13.39,P<0.05）;取样框置于肾皮质深层及肾髓质所测定量参数比较,PI及AUC差异有统计学意义（q=8.41,P<0.05;q=13.13,P<0.05）。取样框置于肾皮质外层及肾门部主肾动脉内所测定量参数比较,PI及AUC差异有统计学意义（q=-8.14,P<0.05;q=-11.67,P<0.05）。取样框置于肾皮质深层及肾门部主肾动脉内所测定量参数比较,PI及AUC差异有统计学意义（q=-12.25,P<0.05;q=-12.78,P<0.05）。取样框置于肾髓质及肾门部主肾动脉内所测定量参数比较,AT、TTP、PI及AUC差异有统计学意义（AT：q=3.91,P<0.05;TTP：q=3.00,P<0.05;PI：q=-19.58,P<0.05;AUC：q=27.12,P<0.05）。取样框置于右肾皮质外层时,2名医师测量PI的一致性非常好,ICC=0.978,95%可信区间为（0.946,0.991）;使用Bland-Altman法分析,2名医师测量PI差值均值为0.42,95%可信区间为（-0.89,1.73）。ROI置于右肾皮质深层时,2名医师测量PI的一致性尚可,ICC=0.698,95%可信区间为（0.238,0.881）;使用Bland-Altman法分析,2名医师测量PI差值均值为1.1,95%可信区间为（-3.3,5.5）。 结论在超声造影定量分析过程中ROI取样框放置于不同部位可对定量分析结果产生影响。ROI取样框置于与声束垂直的右肾皮质外层是兔肾脏血流灌注定量分析较理想ROI设置,实验过程中需保持同一深度、位置,以保证定量分析结果具有较好的可比性和可重复性。

Impact of the location of region of interest on the contrast enhanced ultrasound of rabbit kidney: an animal study

ObjectiveTo study the relationship between different sampling site of region of interest (ROI) and effects of contrast-enhanced ultrasound (CEUS) in normal rabbit kidney and to investigate the optimal location of ROI. MethodsCEUS was performed in twenty normal rabbits using GE LOGIQ E9 machine and a 6-9 MHz probe, after i.v. bolus injection of 1.0 ml/kg contrast agent (SonoVue^?, Bracco, Italy). The qualitative analysis of the corresponding compartment was assessed using low mechanical index (MI) true agent detection contrast mode. Corresponding parameters of the time-intensity curve (TIC): arrival time (AT), time to peak (TTP), Peak intensity (PI) and area under the curve (AUC) were measured using time intensity curve analysis software package (GE). Quantification results of the same sampling site of ROI in the left and right kidney were compared and statistically analyzed. Quantification results of different sampling site of ROI in the right kidney were compared and statistically analyzed. Quantification results of the set of ROI in superficial and deep renal cortex which were measured by two different observers were analyzed by Bland-Altman analysis. ResultsCEUS was performed successfully in all of the rabbits. The TICs of different ROIs in one kidney had a similar trend. The TICs of the same ROI in left and right kidneys had no statistical significance (P>0.05). The corresponding parameters of different ROIs in right kidney had significant difference (AT: F=5.61, P<0.05; TTP: F=4.28, P<0.05; PI: F=145.88, P<0.05; AUC: F=216.72, P<0.05). SNK-q test was used for multiple comparisons. Superficial renal cortex vs deep renal cortex, PI had statistical significance (PI: q=4.43, P<0.05); superficial renal cortex vs renal medulla, PI and AUC had statistical significance (PI: q= 12.69, P<0.05; AUC: q=13.39, P<0.05); deep renal cortex vs renal medulla, PI and AUC had statistical significance (PI: q= 8.41, P<0.05; AUC: q=13.13, P<0.05); superficial renal cortex vs renal hilus, PI and AUC had statistical significance (PI: q=-8.14, P<0.05; AUC: q=-11.67, P<0.05); deep renal cortex vs renal hilus, PI and AUC had statistical significance (PI: q=-12.25, P<0.05; AUC: q=-12.78, P<0.05); renal medulla vs renal hilus, AT, TTP, PI and AUC had statistical significance (AT: q=3.91, P<0.05; TTP: q=3.00, P<0.05; PI: q=-19.58, P<0.05; AUC: q=27.12, P<0.05). For the ROI in superficial renal cortex in right kidney, the ICC for PI measured by 2 observers was 0.987 (95%CI: 0.946 to 0.991), indicating a very strong agreement. Bland-Altman pot showed that the mean difference of PI measured by 2 observers was 0.42 (95%CI: -0.89 to 1.73). For the ROI in deep renal cortex in right kidney, the ICC for PI measured by 2 observers was 0.698 (95%CI: 0.238 to 0.881), indicating a common agreement. Bland-Altman pot showed that the mean difference of PI measured by 2 observers was 1.1 ( 95%CI: -3.3 to 5.5). ConclusionDifferent sampling site of ROIs can affect the results of quantitative analysis. It is important to realize the relationship between the sampling site of ROI and contrast-enhancement effects in microcirculatory quantification. The superficial renal cortex appears to be the optimal location of ROI for tissue perfusion studying in healthy rabbit kidney.