基于原发灶的磁共振ADC图、T2WI影像组学模型预测前列腺癌骨转移的价值  

The value of ADC map and T2WI radiomics analysis of the primary tumor for prediction of bone metastases in prostate cancer

作　　者：陈凤喜 冯俊榜 成杰 王健 韩祺 CHEN Fengxi;FENG Junbang;CHENG Jie;WANG Jian;HAN Qi(Department of Radiology,the Frist Affiliated Hospital of Army Medical University,Chongqing 400038,China;Department of Medical Imaging,Chongqing University Central Hospital/Chongqing Emergency Medical Center,Chongqing 400014,China)

机构地区：[1]陆军军医大学第一附属医院放射科,重庆400038 [2]重庆市急救医疗中心/重庆市第四人民医院医学影像科,重庆400014

出　　处：《磁共振成像》2023年第2期61-67,共7页Chinese Journal of Magnetic Resonance Imaging

基　　金：国家重点研发计划专项(编号:2016YFC0107101);2022年重庆市临床重点专科建设项目(编号:CQZDZK007)。

摘　　要：目的探讨基于前列腺癌(prostate cancer,PCa)原发灶的表观扩散系数(apparent diffusion coefficient,ADC)图、T2加权成像(T2 weighted imaging,T2WI)非增强序列影像组学模型预测PCa骨转移的价值。材料与方法回顾性分析来自2个中心且经穿刺或手术病理证实的178例PCa患者(不伴骨转移者115例,伴骨转移者63例)的MRI(ADC图和T2WI)资料。来自中心1的患者按照7:3的比例随机分为训练组(n=97)和测试组(n=43),来自中心2的患者(n=38)作为外部验证组。图像重采样后,分别在ADC和T2WI图像上勾画肿瘤区域并进行特征提取。经过一致性检验后,采用单因素分析和最小绝对收缩和选择算子(least absolute shrinkage and selection operator,LASSO)算法进行影像组学特征筛选。采用logistic回归分析构建影像组学预测模型,使用受试者工作特征(receiver operating characteristic,ROC)曲线下面积(area under the curve,AUC)评估模型的预测效能。使用DeLong检验对模型进行比较,校准曲线对模型进行评价。结果从ADC、T2WI及ADC+T2WI联合序列中各提取10、3和5个影像组学特征分别进行建模。测试组的AUC分别为0.83(95%CI:0.71~0.95)、0.78(95%CI:0.62~0.93)和0.81(95%CI:0.67~0.95);外部验证组的AUC分别为0.82(95%CI:0.67~0.97)、0.69(95%CI:0.51~0.86)和0.84(95%CI:0.72~0.97)。DeLong检验显示,在外部验证组中基于ADC+T2WI联合序列影像组学模型优于基于T2WI单序列影像组学模型(P=0.02),余无差异。结论基于原发灶的ADC图较T2WI单序列影像组学模型对PCa骨转移具有更高的预测效能,而ADC+T2WI联合影像组学模型并没有表现出更佳的效能,这可能便于临床早期预测PCa骨转移风险。Objective:To investigate the value of prediction model based on apparent diffusion coefficient(ADC)map and T2 weighted imaging(T2WI)unenhanced sequence radiomics of primary tumor in evaluating the bone metastasis in prostate cancer(PCa).Materials and Methods:Unenhanced sequence MRI data(ADC map and T2WI)of 178 PCa patients(115/63 without/with bone metastases)confirmed by puncture or surgical pathology from two centers were retrospectively included.Patients from Center 1 were randomized in a 7∶3 ratio into a training group(n=97)and a test group(n=43),and patients from Center 2(n=38)served as an external validation group.After image resampling,regions of interests(ROIs)were outlined and features were extracted on ADC and T2WI,respectively.Univariate analysis and least absolute shrinkage and selection operator(LASSO)were used for feature screening after consistency testing.Logistic regression analysis was used to construct the radiomics model,and the predictive efficacy of the model was evaluated using the area under the receiver operating characteristic(ROC)curve.The DeLong test was used to compare the models,and the calibration curve was used to evaluate the models.Results:Ten and 3 imaging histological features were extracted from ADC and T2WI sequences,respectively,and 5 imaging histological features were extracted from the combined ADC+T2WI sequence for model building.The AUCs of the test group were 0.83(95%CI:0.71-0.95),0.78(95%CI:0.62-0.93),and 0.81(95%CI:0.67-0.95),respectively,and those of the external validation group were 0.82(95%CI:0.67-0.97),0.69(95%CI:0.51-0.86),and 0.84(95%CI:0.72-0.97),respectively.The DeLong test showed significant difference between ADC+T2WI sequence radiomics model and T2WI single sequence radiomics model in external validation group(P=0.02),others showed no significant difference.Conclusions:The ADC map single sequence radiomics models have relatively high predictive efficacy for the bone metastasis of PCa,and showed similar predictive effective compared with the ADC+T2WI combi

关 键 词：前列腺癌 骨转移 磁共振成像 扩散加权成像 表观扩散系数 影像组学 

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]