摘要
目的评估多参数MRI纹理分析法对前列腺癌的诊断价值及其鉴别低危与中高危前列腺癌的诊断效能。方法回顾性分析经穿刺活检或手术病理证实的61例前列腺癌病人,中位年龄66岁(42～89岁),前列腺特异性抗原(PSA)范围3.4～300 ng/mL,并进行Gleason(GS)评分。其中低危(GS≤6分)病人10例,中高危(GS≥7分)病人51例。所有病人均行前列腺常规MRI及扩散加权成像(DWI)检查,基于T2WI勾画三维兴趣区(ROI),使用纹理分析软件(Matlab)提取ROI的直方图纹理特征参数(熵、偏度、峰度和方差),并计算平均ADC值。采用单因素方差分析评估T2WI和ADC图上各直方图纹理参数及平均ADC值在低危前列腺癌、中高危前列腺癌及正常外周区之间的差异。采用受试者操作特征(ROC)曲线计算曲线下面积(AUC),评估各参数的诊断效能,并评价定量纹理参数对低危和中高危前列腺癌的鉴别诊断效能。结果 ADC图及T2WI上,3组之间各诊断参数差异均有统计学意义(P<0.05)。中高危前列腺癌在ADC图上平均ADC值的AUC(0.88)最大,在T2WI图纹理分析中,方差值AUC(0.88)最大。低危和中高危前列腺癌鉴别诊断中,ADC图和T2WI的纹理参数中熵值AUC最大,分别为0.87和0.79。结论多参数MRI纹理分析可用于诊断前列腺癌,并能为鉴别诊断低危与中高危前列腺癌提供可靠的量化信息,其中熵值有助于前列腺癌病理分级。
Objective To evaluate the diagnostic value of multi-parameter MRI texture analysis in the diagnosis of prostate cancer and its diagnostic efficacy in differentiating low-risk from intermediate-high risk prostate cancer.Methods Sixty one patients(median age 66 years,range 42-89 years;prostate-specific antigen range 3.4-300 ng/mL)with prostate cancer confirmed by biopsy or pathology were retrospectively analyzed.Gleason scores(GSs)were evaluated.Among them,ten cases were at low risk(GS≤6)and 51 cases were at intermediate-high risk(GS≥7).All patients underwent routine prostate MRI and DWI,region of interesting(ROI)was drew based on T2WI images,texture features(entropy,skewness,kurtosis,and variance)were extracted from each ROI using texture analysis software(Matlab),and averaged apparent diffusion coefficient(ADC)values were calculated.One-way ANOVA was used to evaluate the difference in histogram texture parameters and average ADC values of lesions and normal peripheral zones among low-,medium-and high-risk cases.The area under the reciever operating characteristic(ROC)curve(AUC)was used to evaluate the diagnostic value of each parameter and assess the diagnostic efficacy of quantitative texture parameters in differentiating low-risk from intermediate-high risk prostate cancer.Results On the ADC map and T2WI,the differences of diagnostic parameters among the three groups were statistically significant(P<0.05).The ADC mapping texture analysis showed that in the intermediate-high risk prostate cancer diagnosis,the average ADC value achieved the highest efficacy,AUC were 0.88.In the T2WI texture analysis,the parameter with the highest diagnostic efficacy were the variance,with AUC of 0.88.In the differential diagnosis low-risk form intermediate-high risk prostate cancer,the AUC of entropy values in the ADC and T2WI mapping were the largest,the values were 0.87 and 0.79,respectively.Conclusion Multi-parameter MR texture analysis can be used to diagnose prostate cancer and provide reliable quantitative information for differentiating low-risk from intermediate-high risk prostate cancer.Entropy is helpful to the pathological grade of prostate cancer.
作者
张沥
张鑫
王睿
宦怡
李陇超
ZHANG Li;ZHANG Xin;WANG Rui;HUAN Yi;LI Longchao(Department of MR Imaging,Shaanxi Provincial People’s Hospital,Xi’an 710000;Department of Radiology,Xijing Hospital,Affiliated to Air Force Military Medical University;Department of Radiology,Xidian Group Hospital)
出处
《国际医学放射学杂志》
北大核心
2019年第3期264-268,共5页
International Journal of Medical Radiology
基金
陕西省重点研发计划(2018SF-169)
作者简介
通信作者:李陇超,E-mail:18202990355@139.com.
