多模态MRI定量分析诊断前列腺癌的价值研究  

Evaluating the Performance of Muti-Modal MRI Quantitative Analysis in Diagnosing Prostate Cancer

作　　者：冯潇[1] 龙莉玲[1] FENG Xiao;LONG Liling(The First Affiliated Hospital of Guangxi Medical University,Nanning 530021,P.R.China)

机构地区：[1]广西医科大学第一附属医院放射科

出　　处：《临床放射学杂志》2018年第12期2043-2047,共5页Journal of Clinical Radiology

摘　　要：目的探讨多模态MRI定量分析在诊断前列腺癌中的价值,建立综合诊断的预测模型。方法 82例患者均行前列腺MR常规形态学、扩散加权成像(DWI)及动态增强磁共振成像(DCE-MRI)扫描,在不知道病理结果的情况下,根据前列腺12分区切割模型在T2WI图像上采用5分制对各分区评分、测量其表观扩散系数(ADC)、容积转运常数(Ktrans)、速率常数(Kep)、血管外细胞外容积分数(Ve)值并获得相应的时间-信号强度(SI-T)曲线,与病理结果对照,行Logistic回归分析,分析综合诊断准确率,建立综合诊断的预测模型。结果 82例病理证实的患者共获得有效分区772个,其中外周带非癌区185个,中央带非癌区326个,癌区共261个。行Logistic回归分析,T2WI评分、ADC值、Kep值及SI-T曲线等因素对方程的改变均有显著意义(P=0.000),影响力排序为ADC >Kep >SI-T >T2WI,综合诊断灵敏度、特异度、准确度分别为98.8%、96.6%、98.1%。结论前列腺多模态MRI定量分析能提高MRI诊断前列腺癌的准确率,并可获得能客观评价肿瘤良恶性的预测方程。Objective To investigate the valueo of application of Muti-modal MRI in the quantitative diagnosis of prostate cancer.To develop a multi-parametric model suitable for prospectively identifying prostate cancer using MRI.Methods Eighty-two patients had T2-weighted,DWI,and DCE-MRI at 3.0 T MR with body coil to yield parameters T2 score,ADC,SI T curve,Ktrans,Kepand Ve.All patients'diagnosis was confirmed by histopathology through systematic transrectal prostate biopsy or radical prostatectomy.Step-wise logistic-regression modeling was performed,testing changes in deviance for significance,and used to evaluate and compare performance.ResultsAll the 772 effective sextants from 82 cases were confirmed by biopsies.The sensitivity,specificity and accuracy for the detection of prostate cancer were 98.8%,96.6%,98.1%by the combining application of the three examination methods,the influence sort for ADC >Kep >SI-T >T2WI,respectively.Conclusion Prostate cancer can be diagnosed effectively by the combining application of T2WI,DWI and DCE-MRI,and can develop a multi-parametric model suitable for prospectively identifying prostate cancer.

关 键 词：前列腺癌 多模态磁共振成像 定量分析 

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]