单指数、双指数、拉伸指数DWI模型鉴别前列腺癌与基质型前列腺增生的价值研究  

The value of mono-exponential,bi-exponential and stretched exponential DWI models in identifying prostate cancer and stromal prostate hyperplasia

作　　者：吴慧[1] 吴静[1] 蔚纳 钱洛丹 蔚晓玉 贾宇珊 高阳[1] 牛广明[1] WU Hui;WU Jing;YU Na;QIAN Luodan;YU Xiaoyu;JIA Yushan;GAO Yang;NIU Guangming(Department of Radiology,Affiliated Hospital of Inner Mongolia Medical University,Hohhot 010050,China)

机构地区：[1]内蒙古医科大学附属医院影像科,呼和浩特010050

出　　处：《磁共振成像》2020年第7期546-551,共6页Chinese Journal of Magnetic Resonance Imaging

基　　金：内蒙古自治区自然科学基金项目[编号:2017MS(LH)0837]。

摘　　要：目的比较单指数、双指数及拉伸指数DWI模型在鉴别前列腺癌(prostate cancer,PCa)与基质型前列腺增生(stromal prostate hyperplasia,SH)中的价值。材料与方法回顾性分析经直肠超声引导下穿刺活检证实的PCa 24例(28个病灶),基质型前列腺增生20例(22个病灶)的临床资料。所有患者均行T2WI、T1WI、DWI、多b值DWI扫描。根据穿刺病理结果确定癌或增生的部位,与DWI及多b值DWI图像进行匹配,测量两组的单指数模型参数:ADC值,双指数模型参数:D,D*,f值及拉伸指数模型参数:DDC、α值,采用独立样本t检验进行差异性比较;受试者ROC曲线分析各参数鉴别诊断效能,计算其诊断阈值;采用Sperman或Pearson进行相关分析。结果(1)PCa组的ADC、D、DDC、α(0.72±0.14×10^-3 mm^2/s,0.57±0.11×10^-3 mm^2/s,0.71±0.14×10^-3 mm^2/s,0.63±0.03)均低于SH组(1.14±0.13×10^-3 mm^2/s,0.94±0.12×10^-3 mm^2/s,1.30±0.17×10^-3 mm^2/s,0.76±0.05),差异有统计学意义;PCa组的D*(17.96±13.39×10^-3 mm^2/s)高于SH组(10.28±4.96×10^-3 mm^2/s),差异有统计学意义;PCa组的f值(0.30±0.08)略高于SH组(0.29±0.05),差异无统计学意义。(2)在诊断PCa时DDC、D、ADC、α均具有较高诊断效能,曲线下面积AUC分别为:0.9961、0.9957、0.9903、0.9573;分别的诊断阈值为:0.97×10^-3 mm^2/s、0.70×10^-3 mm^2/s、0.86×10^-3 mm^2/s、0.69。(3)无论是PCa组还是SH组ADC、D、DDC两两之间均具有较好相关性,PCa组ADC与DDC相关性最强,r=0.852;SH组ADC与D相关性最强,r=0.751。结论拉伸指数模型DDC、α值,双指数模型D值和单指数模型ADC在鉴别PCa与SH中均具有较高效能,MRI在PCa的诊断中具有明显优越性,DWI不同模型成为其信息补充,双指数模型同时可以提供灌注信息,拉伸指数模型可以提供扩散异质性信息,但双指数模型及拉伸指数模型诊断效能尚未高于传统的单指数模型,所以临床仍推荐应用单指数模型DWI。Objective:To compare the value of various diffusion parameters obtained from monoexponential,biexponential,and stretched exponential diffusion-weighted imaging(DWI)models in identification of prostate cancer and prostate hyperplasia.Materials and Methods:Retrospective analysis of puncture biopsy confirmed prostate cancer(PCa)28 cases,stromal prostate hyperplasia(SH)22 cases of clinical data.All patients underwent T2WI,T1WI,DWI,multi-b DWI scan.According to the results of puncture pathology to determine the location of cancer or hyperplasia,with DWI and multi-b value DWI images to match.ADC,D,D*,f and DDC,αvalues were measured in PCa group and SH group.Differences were compared using independent t test,receiver operating characteristic(ROC)to differentiate the diagnostic efficacy,Correlation analysis using Sperman or Pearson.Results:(1)ADC,D,DDC,αwere lower in the PCa group(0.72±0.14×10^-3 mm^2/s,0.57±0.11×10^-3 mm^2/s,0.71±0.14×10^-3 mm^2/s,0.63±0.03)than in the SH group(1.14±0.13×10^-3 mm^2/s,0.94±0.12×10^-3 mm^2/s,1.30±0.17×10^-3 mm^2/s,0.76±0.05)the difference was statistically significant.D*was higher in the PCa group(17.96±13.39×10^-3 mm^2/s)than in the SH group(10.28±4.96×10^-3 mm^2/s),the difference was statistically significant.f was slightly higher in the PCa group(0.30±0.08)than in the SH group(0.29±0.05),the difference was no statistically significant.(2)DDC,D,ADC,αall have higher diagnostic performance,the area under the curve(AUC)values were 0.9961,0.9957,0.9903,0.9573 respectively;the diagnostic threshold were 0.97×10^-3,0.70×10^-3,0.86×10^-3,0.69.(3)ADC,D,DDC have good correlation both in PCa and SH groups.The PCa group had the strongest correlation with ADC and DDC,the coefficient r=0.852;the correlation between ADC and D was the strongest in SH group,the correlation coefficient r=0.751.Conclusions:DDC,αcalculated from stretched exponential model,D calculated from bi-exponential model and ADC calculated from mono-exponential model in the identifying prostate cancer and s

关 键 词：前列腺肿瘤 前列腺增生 磁共振成像 诊断 鉴别 

分 类 号：R445.2[医药卫生—影像医学与核医学] R737.25[医药卫生—诊断学]