论文部分内容阅读
目的评价多层螺旋CT肝脏体积测量的可重复性和一致性。方法将20例上腹部螺旋CT扫描患者的扫描数据传输至工作站进行肝脏体积重组、测量。由3位有经验的放射科医师独立进行,每位观察者对患者均测量2次。观察者内或观察者间的差异用重复测量的方差分析比较,相互关系用组内相关系数(ICC)描述。观察者间的可靠性分析用Cronbach’s alpha系数描述。结果观察者A2次测得肝脏体积为(1 265.53±421.43)cm3和(1 272.30±414.73)cm3,两者间差值率为(0.80±2.20)%(P=0.15,ICC=0.998 8);观察者B 2次测量的体积为(1 299.97±403.48)cm3和(1 310.23±404.04)cm3,差值率为(0.90±1.90)%(P=0.07,ICC=0.998 3);观察者C 2次测量的体积为(1 282.18±425.91)cm3和(1 289.73±424.66)cm3,差值率为(0.70±2.60)%(P=0.19,ICC=0.998 3)。3个观察者所测肝脏体积的值分别是(1 268.91±418.07)、(1 305.10±403.59)cm3和(1 285.97±425.10)cm3(P=0.07,ICC=0.986 9,Cronbach’salpha系数=0.995 6)。结论多层螺旋CT肝脏体积测量具有较高的一致性和可重复性。
Objective To evaluate the repeatability and consistency of multi-slice spiral CT liver volume measurement. Methods Twenty patients with upper abdominal spiral CT scan were transferred to a workstation for liver volume reorganization and measurement. Independently performed by 3 experienced radiologists, each observer measured 2 times to the patient. Differences between observers or between observers were compared by means of repeated measures ANOVA and correlations were described by intraclass correlation coefficient (ICC). Interobserver reliability analysis is described using the Cronbach’s alpha coefficient. Results The difference between the two groups was (1 265.53 ± 421.43) cm3 and (1 272.30 ± 414.73) cm3, respectively. The difference between the two groups was (0.80 ± 2.20)% (P = 0.15, ICC = 0.998 8) Observer B measured twice the volume of (1 299.97 ± 403.48) cm3 and (1 310.23 ± 404.04) cm3, the difference was (0.90 ± 1.90)% (P = 0.07, ICC = 0.998 3) The volume of the second measurement was (1 282.18 ± 425.91) cm3 and (1 289.73 ± 424.66) cm3, respectively. The difference was (0.70 ± 2.60)% (P = 0.19, ICC = 0.998 3). The values of liver volume measured by 3 observers were (1 268.91 ± 418.07), (1 305.10 ± 403.59) cm3 and (1 285.97 ± 425.10) cm3 respectively (P = 0.07, ICC = 0.986 9, Cronbach’s salinity coefficient = 0.995 6). Conclusion Multi-slice spiral CT liver volume measurement has high consistency and repeatability.