基于SEER数据库构建高危PTMC的诊断模型与验证

Construction and validation of a diagnostic model for high-risk papillary thyroid microcarcinoma based on the SEER database

摘要:

目的 探讨高危甲状腺微小乳头状癌（papillary thyroid microcarcinoma，PTMC）的风险预测因子，构建可靠的列线图诊断模型并验证。方法 回顾性分析2004至2015年SEER数据库中手术治疗的PTMC患者（训练集）以及2020至2022年四川省人民医院甲状腺诊疗中心诊治的甲状腺微小结节患者（外部验证集）的临床病理及超声影像资料。验证集中，利用logistic回归分析高危PTMC的风险预测因子，构建列线图诊断模型并采用内部验证集和外部验证集分别进行验证。再根据术前超声影像学特征进行间接评估，探讨术前超声影像学特征预测高危PTMC的可行性与可靠性。结果 训练集纳入患者1 552例，外部验证集纳入患者516例。训练集中筛选出高危PTMC的独立危险因素为性别（男性）、年龄（≤55岁）、结节数目（多灶）、包膜侵犯，颈部淋巴结异常（P<0.05），构建出的列线图C指数为0.946。在训练集及外部验证集中，该列线图模型的预测结果与实际情况均具有较好的一致性。基于超声影像特征诊断高危PTMC的受试者工作特征（receiver operating characteristic，ROC）的曲线下面积（area under the curve，AUC）值为0.931（95%CI=0.910~0.953），与基于病理特征诊断的AUC值具有较高的一致性（κ=0.611，P<0.05）。结论 本研究构建的高危PTMC诊断模型具有较好的预测效应，利用术前超声影像学特征预测高危PTMC具有一定的临床可行性及临床应用价值。

Abstract:

Objective To investigate the risk predictors of high-risk papillary thyroid microcarcinoma（PTMC） and to construct and validate a reliable diagnostic nomogram model.Methods A retrospective analysis was performed on the clinicopathologic and ultrasound imaging data of patients with PTMC from the SEER database who underwent surgical treatment from 2004 to 2015（training set） and patients with thyroid micronodules treated in the Thyroid Treatment Center of Sichuan Provincial People’s Hospital from 2020 to 2022（external validation set）. In the validation set，the risk predictors of high-risk PTMC were analyzed using logistic regression； a diagnostic nomogram model was constructed，and an internal validation set and an external validation set were used for internal and external validation，respectively. An indirect assessment was performed according to the preoperative ultrasound imaging characteristics to examine the feasibility and reliability of preoperative ultrasound imaging characteristics in predicting high-risk PTMC.Results The training set included 1552 patients，and the external validation set included 516 patients. The independent risk factors for high-risk PTMC were sex（male），age（no more than 55），number of nodules（multikitchen），capsular invasion，and abnormal cervical lymph nodes（P<0.05）. The C-index of the constructed nomogram was 0.946. In the training set and the external validation set，the prediction results of the nomogram model were in good agreement with the actual results. The area under the ROC curve for diagnosing high-risk PTMC based on the ultrasound imaging characteristics was 0.931（95%CI=0.910-0.953），which had a high consistency with the diagnosis based on pathological characteristics（κ=0.611，P<0.05）.Conclusions The diagnostic model of high-risk PTMC constructed in this study has good predictive validity. The prediction of high-risk PTMC based on preoperative ultrasound imaging characteristics is clinically feasible and holds promise for clinical application.