T2 mapping成像纹理分析鉴别宫颈癌病理分型和鳞癌病理分级
Texture analysis of pathological classification of cervical cancer and pathological grading of squamous cell carcinoma by T2 mapping imaging
宋庆玲 1刘爱连 1田士峰 1孟醒 2马长军 1陈丽华 1王楠 1宋清伟1
作者信息
- 1. 大连医科大学附属第一医院放射科,辽宁 大连 116011
- 2. 大连妇女儿童医疗中心放射科,辽宁 大连 116033
- 折叠
摘要
目的:探讨T2 mapping成像纹理分析鉴别宫颈癌病理分型和鳞癌病理分级的价值.方法:回顾性分析 54 例宫颈鳞癌(高级别 37 例、中-低级别 17 例)、19 例腺癌患者的资料,术前均行T2 mapping序列在内的 3.0T MRI检查.两名观察者分别独立提取T2 mapping全肿瘤纹理参数.比较鳞癌与腺癌之间及高级别与中-低级别鳞癌之间参数的差异性,采用Logistic回归构建联合参数.采用受试者工作特征(Receiver operating characteristic,ROC)曲线评估单一纹理参数及联合参数鉴别宫颈癌病理分型及鳞癌病理分级的效能.采用DeLong检验对比分析单一纹理参数和联合参数曲线下面积(Area under the curve,AUC)的差异.结果:从每例患者肿瘤中提取 88 个纹理参数,最终筛选出 3 个参数用于鉴别病理分型,8 个参数用于鉴别鳞癌病理分级(P均<0.05).鉴别宫颈癌病理分型及鳞癌病理分级的联合参数的AUC分别为 0.963、0.966.联合参数的AUC较单一参数的AUC显著提高.结论:基于T2 mapping成像的纹理分析对于宫颈癌病理分型和鳞癌病理分级的鉴别具有一定的价值.
Abstract
Objective:To explore the value of texture analysis of T2 mapping imaging in differentiating pathological classi-fication of cervical cancer and pathological grading of squamous cell carcinoma.Methods:Fifty-four cases of cervical squa-mous cell carcinoma(high grade 37 cases,medium-low grade 17 cases)and 19 cases of adenocarcinoma were analyzed retro-spectively.All patients were scanned by 3.0T MRI including T2 mapping sequence preoperatively.Two observers independently extracted the whole tumor texture parameters of T2 mapping imaging.The differences in related parameters were assessed be-tween cervical squamous cell carcinoma and adenocarcinoma and between high grade and middle-low grade squamous cell carcinoma.Logistic regression was employed to establish the combined parameters.The diagnostic performance of a single tex-ture parameter and the combined parameters in distinguishing pathological classification and pathological grading was evaluated by receiver operating characteristic(ROC)curve.DeLong test was used to compare the difference of area under the curve(AUC)between a single texture parameter and the combined parameter.Results:Eighty-eight texture parameters were extracted from each patient's tumor.Three and eight parameters were selected to distinguish the pathological classification and pathological grading of cervical squamous cell carcinoma respectively(all P<0.05).The AUC of the combined parameters for distinguishing pathological classification and grading were 0.963 and 0.966,respectively.The combined parameters had significantly higher performance than a single parameter.Conclusion:Texture analysis based on T2 mapping imaging is valuable in the differenti-ation of pathological classification of cervical cancer and pathological grading of squamous cell carcinoma.
关键词
宫颈肿瘤/磁共振成像Key words
Uterine Cervical Neoplasms/Magnetic Resonance Imaging引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
万方数据