感兴趣区选取方法对不同病理分级肝细胞癌ADC测量值的影响

The influence of different settings of region of interest on the measurements of ADC value in hepatocellular carcinoma with different pathological grades

目的:探讨不同ROI设置方法对不同病理分级肝细胞癌ADC值测量结果的影响和观察者间的一致性。方法:44例经病理确诊的肝细胞癌(HCC)患者在自由呼吸状态下行肝脏DWI扫描,两位有经验的放射科医师分别采用整体容积法(A组)、单层面最大面积法(B组)和小圆形ROI法(C组),对47个肝细胞癌病灶进行ADC值测量。对三种测量方法获得的不同分化程度HCC的ADC值以及观察者间的组内相关系数(ICC)进行对比分析。结果:对高分化、中分化和低分化HCC的ADC值测量结果,小圆形ROI法分别为(1.41±0.26)×10^(-3)、(1.29±0.33)×10^(-3)和(1.23±0.44)×10^(-3) mm^2/s;单层最大面积法分别为(1.58±0.19)×10^(-3)、(1.40±0.38)×10^(-3)和(1.32±0.39)×10^(-3) mm^2/s;整体容积法分别为(1.59±0.18)×10^(-3)、(1.42±0.29)×10^(-3)和(1.28±0.30)×10^(-3) mm^2/s。A和B组中测得的总体ADC值高于C组,差异有统计学意义(t=-4.522,P=0.000,t=-4.616,P=0.000);而A与B组间的差异无统计学意义(t=-1.029,P=0.309)。B和C组中测量的不同级别HCC的ADC值的差异无统计学意义(P>0.05);A组中测量的不同级别HCC的ADC值的差异有统计学意义(P<0.05),其中以高分化组与低分化组间的差异有统计学意义(t=2.774,P=0.011)。三种ROI设置方法在不同观察者间均存在较好的一致性,A、B、C组内相关系数分别为0.920、0.759和0.843。C组中ADC值与肝癌的病理分级间无显著相关性(r=0.250,P=0.09);而A和B组中ADC值与肝癌的病理分级间呈正相关(r=0.374,P=0.01;r=0.366,P=0.01)。结论:ROI的大小、位置对肝细胞癌ADC测量值和不同观察者间的一致性具有一定影响,整体容积法测量的ADC值有助于高分化与低分化肝细胞癌的鉴别,而且在不同观察者间具有较好的一致性。

Objective: The aim of this study was to investigate the effect of ROI setting on the measurements of ADC value of hepatocellular carcinoma with different pathological grades and the Consistency of different observers. Methods: Forty-four subjects with surgical pathology-proven hepatocellular carcinoma receiving DWI examination before surgery were enrolled in this study.ADC values of 47 lesions were measured by two experienced radiologists based on three ROI protocols:whole tissue volume of the lesion (WTV),maximum areas on single slice (SS max ) and several small round areas (SRAs).The three protocols were compared with each other for difference in ADC values and inter-observer variability (using intraclass correlation coefficient,ICC). Results: The ADC values of highly,moderately and poorly differentiated hepatocellular carcinoma measured by three different ROI protocols were as follows:(1.41±0.26)×10^-3 ,(1.29±0.33)×10^-3 and (1.23±0.44)×10^-3 mm ^2/s in SRAs group;(1.58±0.19)×10^-3 ,(1.40±0.38)×10^-3 and (1.32±0.39)×10^-3 mm ^2/s in SS max group;(1.59±0.18)×10^-3 ,(1.42±0.29)×10^-3 and (1.28±0.30)×10^-3 mm ^2/s in WTV group.The ADCs in SRAs group were lower than those in WTV group and SS max group.The difference between SS max and WTV group and SRAs group was statistically significant ( t =-4.522,P =0.000;t =-4.616, P =0.000),but with no significant difference between SS max group and WTV group ( t =-1.029,P =0.309).In WTV group,the ADCs of the poorly differentiated hepatocellular carcinomas were significantly lower than those of the well differentiated hepatocellular carcinomas ( P <0.05).The difference between SS max group and SRAs group was not statistically significant ( P >0.05).Good consistency of the two observers was showed in the three groups with ICC of 0.920 (WTV),0.843 (SS max ) and 0.759 (SRAs).A weak positive correlation was found between the ADC values and histological grades of HCCs in SS max groups and WTV groups ( r =0.374,P =0.01;r =0.366,P =0.01). Conclusion: The size and location of ROI have some influence on ADC measurement and variation among observers.WTV is helpful in differentiating well-differentiated from poorly differentiated hepatocellular carcinoma and good consistency of the two observers was showed by WTV.