| 摘要: |
| 目的:基于多模态超声及临床资料,构建甲状腺微小乳头状癌(PTMC)患者颈部淋巴结转移(CLNM)的预测模型。方法:选取2022年1月～2023年3月在我院诊治的经手术病理证实的PTMC患者82例,术前行常规超声、超声造影(CEUS)和超微血管成像(SMI)检查,按照有无发生CLNM分为CLNM组32例和非CLNM组50例,比较两组多模态超声资料及临床资料,并分析PTMC患者CLNM的危险因素,基于多模态超声及临床特征构建的PTMC患者CLNM的预测模型,分析其预测价值。结果:CLNM组年龄、病灶最大径、微钙化数量、增强早期被膜连续性、甲状腺浸润与非CLNM组比较,均有统计学意义(P<0.05)。经二元Logistic回归分析,年龄大是PTMC患者CLNM的保护因素(P<0.05),病灶最大径>1.0cm、微钙化数量>5个、增强早期被膜连续性中断、甲状腺浸润是PTMC患者CLNM的危险因素(P<0.05)。Hosmer-Lemeshow拟合度检验显示,χ2=6.868,P=0.551,模型拟合优度较好。ROC分析显示,基于多模态超声及临床特征构建的PTMC患者CLNM的预测模型的AUC为0.924,约登指数为0.733,敏感度、特异度分别为81.3%、92.0%,95%CI为(0.864,0.984),准确性为87.8%。结论:本研究基于多模态超声及临床特征构建的PTMC患者CLNM的预测模型,预测价值较好,能够为PTMC患者发生CLNM的预测提供辅助参考。 |
| 关键词: 多模态超声 临床资料 甲状腺微小乳头状癌 颈部淋巴结转移 |
| DOI： |
| 投稿时间：2023-10-17修订日期：2023-10-25 |
| 基金项目: |
|
| Prediction Model of Cervical Lymph Node Metastasis of Thyroid Micropapillary Carcinoma Based on Multimodal Ultrasound and Clinical Application |
| Lin fengsen |
| () |
| Abstract: |
| 【】Objective:Based on multimodal ultrasound and clinical data, a prediction model of cervical lymph node metastasis (CLNM) in patients with papillary thyroid microcarcinoma (PTMC) was established.Methods:Eighty-two PTMC patients diagnosed and treated in our hospital from January, 2022 to March, 2023 were examined by conventional ultrasound, contrast-enhanced ultrasound and microvascular imaging (SMI) before operation, and were divided into CLNM group (32 cases) and non-CLNM group (50 cases) according to the presence or absence of CLNM. The multimodal ultrasound data and clinical data of the two groups were compared, and the risk factors of CLNM in PTMC patients were analyzed. The CLNM prediction of PTMC patients was constructed based on multimodal ultrasound and clinical characteristics.Result:Compared with the non-CLNM group, CLNM group had significant differences in age, maximum diameter of focus, number of microcalcifications, enhancement of early capsule continuity and thyroid infiltration (P<0.05). According to binary Logistic regression analysis, older age is the protective factor of CLNM in PTMC patients (P<0.05), and the largest diameter of lesions > 1.0cm, the number of microcalcifications > 5, the interruption of early capsule enhancement and thyroid infiltration are the risk factors of CLNM in PTMC patients (P<0.05). Hosmer-Lemeshow test shows that χ2=6.868, P=0.551, and the model has a good goodness of fit. ROC analysis showed that the AUC, Jordan index, sensitivity and specificity of the prediction model of CLNM in PTMC patients based on multimodal ultrasound and clinical features were 0.924, 0.733, 81.3%, 92.0%, 95%CI (0.864,0.984) respectively, and the accuracy was 87.8%.Conclusion:The prediction model of CLNM in PTMC patients based on multimodal ultrasound and clinical features in this study has good prediction value and can provide auxiliary reference for the prediction of CLNM in PTMC patients. |
| Key words: |
