双参数MRI影像组学对前列腺癌Gleason分级的诊断价值  

Diagnostic value of radiomics based on biparametric prostate MRI imaging in Gleason classification of prostate cancer

作　　者：张洪涛 胡泽宇 王海屹[1] 王博[3] 白旭 叶慧义[1] Zhang Hongtao;Hu Zeyu;Wang Haiyi;Wang Bo;Bai Xu;Ye Huiyi(Department of Radiology, the First Medical Center of PLA General Hospital, Beijing 100853, China;College of Microelectronics,Xidian University,Xi' an 710071,China;Tsinghua University,Beijing 100084, China;Department of Radiology, South Area of the Fifth Medical Center of PLA General Hospital, Beijing 100071, China)

机构地区：[1]解放军总医院第一医学中心放射科,北京100853 [2]西安电子科技大学微电子学院,西安710071 [3]清华大学,北京100084 [4]解放军总医院第五医学中心南院区放射科,北京100071

出　　处：《中华放射学杂志》2019年第10期849-852,共4页Chinese Journal of Radiology

基　　金：国家自然科学基金面上项目(81771785);保健专项课题(14BJZ02).

摘　　要：目的探讨双参数MRI影像组学区分前列腺癌Gleason高、低危组分级的价值。方法回顾性分析2015年10月至2018年12月解放军总医院第一医学中心,MRI检查2个月内获得前列腺根治性切除术病理结果的316例前列腺癌患者。高危组(Gleason评分≥4+3)182例,低危组(Gleason评分≤3+4)134例。患者均行高分辨率横轴面T2WI和横轴面DWI扫描(b=0、1 000、2 000、3 000 s/mm^2)。利用3D-Slicer软件手动勾画病灶,使用Radiomics方法提取特征,进行Spearman非参数相关性检验,然后进行降维。通过前列腺癌作为输入,对构造的神经网络进行训练,并且输入测试集以获得模型的预测能力。最后使用10次交叉验证和100次数组洗牌来提高预测的准确性和模型的泛化能力。结果对316例患者的前列腺癌靶病灶提取了106维特征。将不显著相关的29维特征剔除后,余下的77维特征送入PCA降维,得到了保留99%原始特征信息的21维新特征空间。T2WI以及b=1 000、2 000、3 000 s/mm^2 DWI区分测试集高、低危组前列腺癌的ROC下面积分别为0.712、0.689、0.689和0.691。结论双参数MRI中提取特征并利用神经网络能够准确自动区分前列腺癌病理的Gleason高危组和低危组。Objective To explore the value of radiomics in stratifying the Gleason score (GS) of prostate cancer based on vast image features from biparametric MRI. Methods Three hundred and sixteen patients were enrolled in this study from October, 2015 to December, 2018 and their results of surgical pathology were obtained. The lesions were manually depicted by 3D-Slicer. Then, 106-dimensional features extracted by radiomics were used to conduct Spearman non-parametric correlation test with the high and low risk stratification of GS. The constructed Neural Network was trained with the features after dimension reduction by principal component analysis as the input. Then, the testing set was fed in to get the predictive capability of the model. In the end, 10-fold cross-validation and shuffle of 100 times were used to test the accuracy of the prediction and the generalization ability of the model. Results Seventy seven-dimensional features with significant correlation were found at the level of P valued=0.05 (two-tailed). After dimensional features were reduced, 21 dimensional new feature spaces with 99% original feature information were obtained. The results on the testing data after the 10-fold validation and shuffle were AUC=0.712 with T2WI, AUC=0.689 with DWI (b=1 000 s/mm^2), AUC=0.689 with DWI (b=2 000 s/mm^2) and AUC=0.691 with DWI (b=3 000 s/mm^2). Conclusion The neural network after extracting features from biparametric MRI images can accurately and automatically distinguish the high risk and low risk groups of Gleason grade of prostatic cancer.

关 键 词：前列腺肿瘤 磁共振成像 影像组学 

分 类 号：R737.25[医药卫生—肿瘤] R445.2[医药卫生—临床医学]