胆囊息肉良恶性病变结合增强CT的临床鉴别研究  

Enhanced CT Combined with Clinical Data to Distinguish Benign and Malignant Polypoid Lesions of Gallbladder

作　　者：王冬梅 王悦华[1] 王晓娟[2] WANG Dong-mei;WANG Yue-hua;WANG Xiao-juan(Department of General Surgery,Xuanwu Hospital,Capital Medical University,Tongzhou 100053,Beijing,China;Beijing Chaoyang Hospital,Capital Medical University,Chaoyang 100100,Beijing,China)

机构地区：[1]首都医科大学宣武医院普外科,北京通州100053 [2]首都医科大学附属北京朝阳医院,北京朝阳100100

出　　处：《中国CT和MRI杂志》2025年第3期140-142,共3页Chinese Journal of CT and MRI

基　　金：2023首都卫生发展全科医学与社区卫生科研专项(首发2023-2Y-006)。

摘　　要：目的良恶性胆囊息肉样病变(PLG)在胆囊切除术前利用影像组学进行鉴别诊断。方法回顾分析经手术病理证实的145例(增加数据量2000例)≥1 cm胆囊息肉样病变增强CT资料,70%案例用于训练数据集,30%案例用于测试集。多变量Logistic回归用于生成放射组学模型,绘制ROC曲线和判决曲线,以评估区分性能。结果放射组学标志物和临床标志物的ROC曲线下面积分别为0.924和0.861。放射组学诊断准确率为88.6%,特异度为88.0%,敏感度为89.5%。联合应用AUC值为0.931,特异度为84.0%,灵敏度为89.5%。两种单一模型和联合模型的AUC值差异无统计学意义。结论基于CT图像的放射组学检查有助于鉴别≥1cm大小的良性和恶性胆囊息肉。Objective To establish and validate a radiomic-based method for the preoperative differential diagnosis of benign and malignant gallbladder polypoid lesions(PLG).Methods The enhanced CT data of 145 gallbladder polyposis≥1 cm confirmed by surgery and pathology were retrospectively analyzed.70%cases were used in training dataset and 30%cases were used in test dataset.Multivariate Logistic regression was used to generate a radiomic model and plot the ROC curve and decision curve to evaluate the differentiation performance.Results The area under ROC curve of radiomic and clinical markers was 0.924 and 0.861,respectively.The diagnostic accuracy,specificity and sensitivity of radiomics were 88.6%,88.0%and 89.5%respectively.The AUC value of combined application was 0.931,the specificity was 84.0%and the sensitivity was 89.5%.There was no significant difference between the AUC values of the two single models and the combined models.Conclusion Radiomic examination based on CT images is helpful in differentiating benign and malignant gallbladder polyps≥1 cm in size.

关 键 词：良恶性胆囊息肉样病变 放射组学 多变量Logistic回归 

分 类 号：R575.6[医药卫生—消化系统]