基于超声图像的直方图分析在睾丸良恶性病变鉴别诊断中的价值  

Value of Histogram Analysis Based on Ultrasound Images in the Differential Diagnosis of Benign and Malignant Testicular Lesions

作　　者：苏唤忠 洪龙城 黄美[1] 苏毅明[1] 张佐炳[1] 张晓东[1] Su Huanzhong;Hong Longcheng;Huang Mei;Su Yiming;Zhang Zuobing;Zhang Xiaodong(Department of Ultrasound,The First Affiliated Hospital of Xiamen University,Xiamen,Fujian 361003,China)

机构地区：[1]厦门大学附属第一医院超声影像科,福建省厦门市361003

出　　处：《中国超声医学杂志》2024年第3期324-327,共4页Chinese Journal of Ultrasound in Medicine

基　　金：福建省自然科学基金(No.2023J011617)。

摘　　要：目的 探讨基于超声图像的直方图分析在睾丸良恶性病变鉴别诊断中的可行性和价值。方法 回顾性分析经病理证实的70例睾丸病变患者的超声图像,根据病理结果分为良性组(23例)和恶性组(47例)。基于超声医师手工勾画的感兴趣区进行直方图分析,计算均值、方差、偏度、峰度、不同百分位数(P_(1)、P_(10)、P_(50)、P_(90)、P_(99))。对良恶性组的直方图参数进行比较,并用多因素Logistic回归分析筛选独立影响因素。绘制受试者工作特征(ROC)曲线评价直方图参数的鉴别诊断效能。结果 良性组的方差、偏度、P_(99)明显高于恶性组(P<0.05);P_(10)、P_(50)明显低于恶性组(P>0.05)。多因素Logistic回归分析结果显示,方差及偏度是鉴别睾丸良恶性病变的独立影响因素。ROC曲线结果显示,方差、偏度、P_(10)、P_(50)、P_(99)鉴别诊断的曲线下面积(AUC)为0.693~0.839;方差及偏度联合鉴别诊断的AUC为0.910(95%CI:0.838~0.982)。结论 基于超声图像的直方图分析是鉴别睾丸良恶性病变的一个有潜能的工具,其中方差及偏度的联合诊断价值最高。Objective To investigate the value and feasibility of histogram analysis based on ultrasound images in the differential diagnosis of benign and malignant testicular lesions.Methods Ultrasound images of 70 patients with pathologically confirmed testicular lesions were retrospectively analyzed,and the patients were divided into benign group(23 cases) and malignant group(47 cases) according to the pathological results.Histogram analysis was performed based on the region of interest manually outlined by two sonographers,thereafter,mean,variance,skewness,kurtosis,percentiles(1st,10th,50th,90th,99th) were generated.Histogram parameters of benign and malignant groups were compared and the independent predictors were screened by using multivariate Logistic regression analysis.The differential diagnostic efficacy of the histogram parameters were assessed by the receiver operator characteristic(ROC) curve.Results The variance,skewness,and 99th percentiles in the benign group were significantly higher than those in the malignant group(P<0.05),and the 10th and 50th percentiles were significantly lower than those in the malignant group(P>0.05).Multivariate Logistic regression analysis showed that variance and skewness were independent factors in differentiating benign and malignant testicular lesions.ROC curves showed that the area under the curves(AUC) of variance,skewness,10th percentiles,50th percentiles,and 99th percentiles was 0.693-0.839.The AUC of combined variance and skewness was 0.910(95%CI:0.838-0.982).Conclusions Histogram analysis based on ultrasound images is apromising tool in distinguishing between benign and malignant testicular lesions,in which the joint diagnosis value of variance and skewness is the highest.

关 键 词：超声图像 直方图分析 睾丸病变 鉴别诊断 

分 类 号：R445.1[医药卫生—影像医学与核医学] R737.21[医药卫生—诊断学] R697.22[医药卫生—临床医学]