摘要
目的探讨1.5T MR条件下T_2WI联合DWI及PI-RADS对前列腺癌(prostate cancer,PCa)的临床诊断价值。方法回顾性分析93例前列腺疾病患者的MRI资料,其中前列腺癌43例,前列腺增生42例,炎症8例,所有病例均行T_2WI、DWI扫描。根据PI-RADS对上述93例MRI资料分别按T_2WI组和T_2WI+DWI组进行影像诊断评分,将影像评分与病理结果对照后作受试者工作特征曲线(ROC)分析。结果 43例PCa,其中T_2WI漏诊7例,T_2WI+DWI漏诊3例;非癌病例50例,其中T_2WI误诊为癌27例,T_2WI+DWI误诊11例。各种检查方案检出前列腺癌的敏感性、特异性、准确率和曲线下面积(AZ)分别为:T_2WI组,83.72%、46.00%、63.44%和0.777;T_2WI+DWI组,93.02%、78.00%、84.95%和0.929。T_2WI组和T_2WI+DWI组的Az值存在显著统计学差异(Z值为3.635,P<0.001)。结论 T_2WI联合DWI配合PI-RADS能够显著提高临床对前列腺癌的诊断特异性和准确率,适合作为前列腺癌MRI筛查的基本诊断序列。
Objective To evaluate the diagnostic value of traditional MRI( T2WI) combined with DWI for prostate cancer detection using 1. 5 T MR scanner with the application of PI-RADS. Methods Within ninety-three patients with MRI data,fortythree were confirmed pathologically as cancerous and fifty were noncancerous. All cases were performed traditional MRI( T2WI)and DWI prior to needle biopsy. Thus,the power of diagnosis for prostate cancer using PI-RADS and two imaging protocols as follows were subjected to receiver operating characteristic( ROC) curve analysis: T2WI alone,T2WI and DWI. Results Of fortythree prostate cancer,seven were missed diagnosis by T2WI alone,three were missed diagnosis by T2WI and DWI. Of fifty noncancerous,twenty-seven were misdiagnosis by T2WI alone,eleven were misdiagnosis by T2WI and DWI. The sensitivity,specificity,accuracy,and area under the ROC curve( Az) for region-based analysis were 83. 72%,46. 00%,63. 44% and 0. 777 in T2WI alone; 93. 02%,78. 00%,84. 95% and 0. 929 in T2WI and DWI. ROC analysis revealed that significant differences of Az have been found between protocols one and two( Z = 3. 635,P〈0. 001). Conclusion T2WI and DWI in combination with PIRADS can improve the diagnostic efficiency significantly for the detection of prostate cancer prominently. T2WI with DWI is an available imaging protocol for prostate cancer screening.
出处
《医学影像学杂志》
2016年第5期881-884,共4页
Journal of Medical Imaging
基金
2014江苏省苏州市科技发展计划项目(编号:SYS201475)
关键词
前列腺癌
磁共振成像
Prostate cancer
Magnetic resonance imaging