IQon光谱CT碘密度图影像组学对甲状腺乳头状癌颈部淋巴结转移的诊断价值

Radiomics analysis of iodine maps derived from IQon spectral CT for the diagnosing metastatic cervical lymph nodes in patients with papillary thyroid cancer

目的:探讨基于IQon光谱CT碘密度图的影像组学对甲状腺乳头状癌(PTC)患者颈部淋巴结转移的诊断价值。方法:回顾性分析2019年10月—2022年10月山西医科大学第一医院术前行颈部光谱CT增强扫描的89例PTC患者的267枚淋巴结,按照7∶3的比例分为训练集(108枚非转移淋巴结和78枚转移淋巴结)和验证集(50枚非转移淋巴结和31枚转移淋巴结)。采用ITK-SNAP软件在光谱CT动静脉期碘密度图上手动勾画淋巴结的三维感兴趣容积,使用FAE软件提取每枚淋巴结的组学特征,通过LASSO回归进行组学特征筛选。分别单独使用CT图像特征、碘密度图影像组学特征及两者联合构建Logistic逻辑回归模型。采用ROC曲线及曲线下面积(AUC)评估3个模型诊断效能。结果:模型1由淋巴结短径、强化程度、钙化和囊变4个特征构成。LASSO回归共筛选出9个组学特征并建立了影像组学标签(模型2)。模型3由淋巴结强化程度、钙化、囊变和影像组学标签构成。在训练集和验证集,模型3均具有最优的诊断效能(AUC为0.94、0.91)。结论:基于IQon光谱CT碘密度图的影像组学分析对PTC患者颈部淋巴结转移具有较好的诊断效能。

Objective:To investigate the value of radiomics analysis of iodine maps derived from IQon spectral CT for preoperative diagnosing cervical lymph nodes metastasis in patients with papillary thyroid cancer(PTC).Methods:A retrospective analysis was conducted on 89 patients with PTC who underwent preoperative neck spectral CT enhanced scans at the First Hospital of Shanxi Medical University from October 2019 to October 2022.A total of 267 lymph nodes were assessed,they were divided into a training set(108 non-metastatic lymph nodes and 78 metastatic lymph nodes)and a validation set(50 non-metastatic lymph nodes and 31 metastatic lymph nodes)in a 7:3 ratio.Manual delineation of three-dimensional regions of interest for the lymph nodes was performed on iodine density maps in the arterial and venous phases using ITK-SNAP software.Radiomic features of each lymph node were extracted using FAE software.The least absolute shrinkage and selection operator regression(LASSO)method was used for radiomic feature selection.Logistic regression was used to establish diagnostic models based on CT image feature of the cervical lymph nodes,radiomics signature,and the combined.The diagnostic performance of the three models was evaluated using ROC curves and the area under the curve(AUC).Results:Model 1 consisted of lymph node short axis,degree of enhancement,calcification,and cystic changes.LASSO regression selected nine radiomic features and established radiomics signature(Model 2).Model 3 consisted of cervical lymph nodes degree of enhancement,calcification,cystic changes and radiomics signature.In both the training and validation sets,Model 3 showed the optimal diagnostic performance(AUC=0.94,0.91).Conclusion:Radiomics analysis of IQon spectral CT-derived iodine maps can realize effective diagnosis of cervical lymph nodes metastasis in patients with PTC.