CT影像组学模型对糖尿病足患者足底神经病变的诊断价值  

Diagnostic Value of CT Radiomics Model for the Plantar Neuropathy in Patients with Diabetic Foot

作　　者：耿冀[1] 吕喆[1] 张滨[1] 徐井旭 GENG Ji;LV Zhe;ZHANG Bin;XU Jingxu(Department of Radiology,Peking University Shougang Hospital,Beijing 100144,China;Beijing Deepwise&League of PHD Technology Co.,Ltd.,Beijing 100089,China)

机构地区：[1]北京大学首钢医院医学影像科,北京100144 [2]北京深睿博联科技有限责任公司,北京100089

出　　处：《中国医疗设备》2022年第8期132-136,共5页China Medical Devices

摘　　要：目的探讨CT影像组学模型在诊断糖尿病足患者足底神经病变中的价值。方法回顾性分析2019年7月至2020年12月于我院就诊的并发糖尿病足的29例糖尿病患者(糖尿病足组)及47例同期单侧足部创伤患者(非糖尿病组)的临床资料及足部CT影像学资料。基于足部CT图像采用深睿科研平台提取1743个影像组学特征,使用特征间线性相关检查和F检验进行特征筛选,随后采用Logistic回归分析进行模型构建。采用五折交叉验证训练模型,在训练组与验证组中应用受试者操作特征(Receiver Operating Characteristic,ROC)曲线对模型进行验证,评价影像组学特征在诊断糖尿病足患者足底神经病变中的效能。结果经过特征筛选,最终选取12个影像组学特征用于构建糖尿病足患者足底神经病变诊断模型。训练组中诊断模型的ROC曲线下面积(Area Under Curve,AUC)为0.97(95%CI:0.94~1.00),敏感度为90.20%,特异度为89.13%,诊断准确率为89.69%;验证组AUC为0.89(95%CI:0.82~0.95),敏感度为84.31%,特异度为80.43%,诊断准确率为82.47%。结论基于足部CT图像的影像组学模型对糖尿病足患者足底神经病变有较高的诊断效能。Objective To investigate the value of CT imaging omics model in the diagnosis of plantar neuropathy in diabetic foot patients.Methods The clinical and CT imaging data of 29 diabetic patients(diabetic foot group)and 47 patients(non-diabetic foot group)with unilateral foot trauma during the same period who were admitted to our hospital from July 2019 to December 2020 were retrospectively analyzed.Based on foot CT images,1743 image omics features were extracted by deepwise scientific research platform.Linear correlation between features and F-test were used for feature selection,and then Logistic regression analysis was used to construct the model.The receiver operating characteristic(ROC)curve was used to verify the model in the training group and the verification group,and the efficacy of imaging omics features in diagnosis of plantar nerve in diabetic foot was evaluated.Results After feature screening,12 imaging features were selected to construct the diagnosis model of plantar neuropathy in diabetic foot patients.In the training group,the area under the ROC curve(AUC)was 0.97(95%CI:0.94-1.00),the sensitivity and specificity were 90.20%and 89.13%,respectively,and the diagnostic accuracy of the model was 89.69%.In the verification group,the AUC was 0.89(95%CI:0.82-0.95),the sensitivity and specificity were 84.31%and 80.43%,respectively,and the diagnostic accuracy of the model was 82.47%.Conclusion The imaging group model based on foot CT images has high diagnostic efficiency for plantar neuropathy in patients with diabetes foot.

关 键 词：CT影像组学 模型 糖尿病足 足底神经病变 

分 类 号：R587.1[医药卫生—内分泌]