磁共振纹理分析技术在前列腺癌Gleason分级中的应用价值

Application of magnetic resonance texture analysis in Gleason grading of prostate cancer

目的探讨磁共振纹理分析方法在前列腺癌侵袭性评价中(Gleason分级)的价值。方法回顾性收集66例行前列腺癌根治术患者,根据术后病理Gleason评分(GS)分为五组:低危组:GS=6分;中低危组:GS=(3+4)分;中危组:GS=(4+3)分;中高危组:GS=8分;高危组:GS=9~10分。提取并分析磁共振T2加权图像上前列腺癌区的纹理特征,比较五组的熵值、能量、偏度及峰度等纹理值变化;分析各纹理值与Gleason分级的相关性,及其组合对前列腺癌GS的诊断价值。结果不同GS组前列腺癌区纹理值中熵值随着GS评分的增高逐渐增高,能量随着GS评分的增高呈降低趋势,差别均有统计学差异(F分别=10.83、2.80,P均<0.05),而不同分组间的偏度及峰度比较,差异均无统计学意义(F分别=1.00、0.84,P均>0.05)。不同GS组间与前列腺癌区熵值呈正相关,与能量呈负相关(r分别=0.27、-0.49,P均<0.05)。低危组与非低危组之间,联合使用熵和能量的AUC高于单独使用能量的AUC,差异有统计学意义(Z=3.50,P<0.05);低危+中低危组与其他三组之间、低危+中低危+中危组与其他两组之间,联合使用熵和能量的AUC均高于单独使用熵的AUC,差异均有统计学意义(Z分别=2.74、2.16,P均<0.05)。结论基于磁共振T2WI图像的纹理分析可以评估前列腺癌Gleason分级,联合使用熵及能量可以取得更好的诊断效能。

Objective To investigate the value of magnetic resonance texture analysis in the evaluation of prostate cancer aggressiveness.Methods Totally 66 patients with prostate cancer of pathological confirmation were retrospectively enrolled.According to postoperative pathological Gleason scores(GS),the patients were divided into five groups.The patients whose GS was 6 points,3 plus 4 points,4 plus 3 points,8 points,and 9~10 points were enrolled as low risk group,medium low risk group,medium risk group,medium high risk group,and high risk group,respectively.The texture features of prostate cancer on T2 weighted MR images were extracted and analyzed,and the changes of entropy,energy,skewness and kurtosis among five groups were compared.The receiver operating characteristic curve was used to analyze the diagnostic value of each texture value and its combination for prostate cancer GS.Results In different GS groups,the entropy value of prostate cancer area increased with the increasing of GS score,while the energy value decreased with the increasing of GS score(F=10.83,2.80,P<0.05).There was no significant difference in skewness and kurtosis among different groups(F=1.00,0.84,P>0.05).The GS was positively correlate with the entropy of prostate cancer,while negatively correlate with the energy(r=0.27,-0.49,P<0.05).The AUC of entropy combining with energy was higher than that of energy only between the low-risk group and the non-low-risk group(Z=3.50,P<0.05).The AUC of entropy combining with energy was higher than that of entropy only between low-risk,medium low-risk group and other three groups as well as between low-risk,medium low,and medium-risk group and other two groups(Z=2.74,2.16,P<0.05).Concl⁃usion Texture analysis based on T2-weighted images could evaluate Gleason grade of prostate cancer,the combination of entropy and energy can achieve better diagnostic efficiency.