多模态超声联合细针穿刺特征构建列线图预测甲状腺乳头状癌颈部淋巴结颈部淋巴结转移转移

Multimodal ultrasound combined with fine needle puncture features to construct a column chart for predicting cervical lymph node metastasis in papillary thyroid carcinoma

目的分析多模态超声联合细针穿刺特征构建列线图模型预测甲状腺乳头状癌颈部淋巴结转移的应用价值。方法收集我院86例甲状腺乳头状癌患者资料,以术后颈部淋巴结病理结果为标准分为转移组和非转移组,单因素及多因素Logistic筛选预测甲状腺癌颈部淋巴结转移的独立影响因素以构建列线图。校准曲线、Hosmer-Lemeshow检验及决策曲线分析评价列线图的校准度及临床应用价值。结果性别、弹性评分、结节边缘、结节最大径、血流分级、A-SD、弹性比Shell/A、穿刺位移为筛选后甲状腺癌淋巴结转移的独立影响因素(P<0.05),列线图曲线下面积为0.975(95%CI:0.944~1.000),敏感度为90.01%,特异度为100%,准确率为95.39%;列线图=5.885×弹性评分-3.201×性别+0.158×结节边缘+6.718×血流分级+0.37×结节最大径+0.481×A-SD+0.901×穿刺位移-40.165×弹性比-Shell/A。HL检验、校准曲线显示模型具有良好的校准;决策曲线分析表明在相同的预测概率下列线图的临床效能最大。结论构建列线图预测甲状腺癌淋巴结转移有助于临床诊断和治疗。

Objective The application value of analyzing the characteristics of multimodal ultrasound combined with fine needle aspiration and constructing a nomogram model to predict cervical lymph node metastasis in papillary thyroid carcinoma.Methods We collected data from 86 patients with papillary thyroid carcinoma at our hospital.These patients were categorized into metastatic and non-metastatic groups based on their postoperative cervical lymph node pathology results.Through univariate and multivariate logistic regression analyses,we identified independent predictors for cervical lymph node metastasis in thyroid cancer.Subsequently,a nomogram was constructed to predict the risk of cervical lymph node metastasis in thyroid cancer.Results Gender,elasticity score,nodule margin,maximum nodule diameter,blood flow grade,ASD,the elasticity ratio(Shell/A),and puncture displacement were identified as independent influencing factors for lymph node metastasis in thyroid cancer post-screening(P<0.05).The area under the ROC curve was 0.975(95%CI:0.944-1.000),with a sensitivity of 90.01%,specificity of 100%,and an accuracy of 95.39%;The equation for the ROC curve was 5.885×elasticity score-3.201×gender+0.158×nodule margin+6.718×blood flow grade+0.37×maximum nodule diameter+0.481×A-SD+0.901×puncture displacement-40.165×elasticity ratio Shell/A.The HL test and calibration curve indicated that the model exhibits good calibration;The decision curve analysis demonstrated that the ROC curve provided the greatest clinical efficacy at the same prediction probability.Conclusion Constructing a column chart to predict lymph node metastasis in thyroid cancer is helpful for clinical diagnosis and treatment.