目的：构建初次行甲状腺癌手术患者术后声带麻痹情况的预测模型，预测声带麻痹发生机率。 方法：回顾性收集我院2012年1月-2022年2月394例甲状腺癌患者的临床资料，根据术后有无声带麻痹分为声带麻痹组及无声带麻痹组，根据时间顺序，将2012年1月至2021年5月358例作为建模组，2021年6月至2022年2月36例作为验证组，收集及分析两组临床患者基本资料，运用单因素及多因素分析，选择最合适的自变量构建Logistic回归模型，分析甲状腺癌术后声带麻痹的影响因素，内部通过十折交叉验证，外部验证通过比较建模组及验证组的区分度、校准度及临床有效性，绘制受试者工作特征（ROC）曲线、校准曲线及临床决策曲线，评估Logistic回归模型预测价值，计算约登指数、灵敏度、特异度、预测概率P值，根据P=1/1+e-y，计算出Y值，最后绘制nomogram图。 结果：394例患者均一次完成甲状腺癌手术切除术，术后41例患者发生声带麻痹，声带麻痹发生率为10.4%。Logistic回归模型可知喉返神经入喉处是否粘连（OR=11.804，95%CI为3.078~45.273）、术前Tg（OR=0.021，95%CI为0.002~0.202）、是否贴近喉返神经（OR=20.984，95%CI为2.058~214.007）、手术时间（OR=2.768，95%CI为1.122~6.829）是甲状腺癌术后声带麻痹的独立预测因素。十折交叉验证显示ROC曲线下面积为0.7284，建模组ROC曲线下面积为0.7943（95%可信区间为0.716~0.872），验证组ROC曲线下面积为0.7722（95%可信区间为0.526~1）；Hosmer-Lemeshow拟合优度检测显示模型拟合较好，建模组Chi-Square=1.1，P=0.9816>0.05，验证组Chi-Square=3.87，P=0.5677>0.05；约登指数最大为0.188，此时灵敏度为0.57，特异度为0.62，ROC曲线下面积为0.59，P=0.216，为最佳临界值，根据P=1/1+e-y，此时Y=-1.25。 结论：本回归模型预测准确度较好，对甲状腺癌术后声带麻痹的发生具有一定的参考意义。
ABSTRACT Objective: To construct a prediction model of vocal cord paralysis in patients undergoing thyroid cancer surgery for the first time, and to predict the incidence of vocal cord paralysis. Methods:The clinical data of 394 patients with thyroid cancer in our hospital from January 2012 to February 2022 were collected retrospectively. According to the postoperative silent zone paralysis, they were divided into vocal cord paralysis group and silent zone paralysis group. According to the time sequence, 358 patients from January 2012 to may 2021 were used as the modeling group and 36 patients from June 2021 to February 2022 were used as the validation group. The basic data of clinical patients in the two groups were collected and analyzed by univariate and multivariate analysis, The most appropriate independent variables were selected to construct a logistic regression model to analyze the influencing factors of vocal cord paralysis after thyroid cancer surgery. The internal was verified by ten fold cross validation, and the external validation was verified by comparing the differentiation, calibration and clinical effectiveness of the modeling group and the validation group. The receiver operating characteristic (ROC) curve, calibration curve and clinical decision-making curve were drawn to evaluate the predictive value of the logistic regression model, and the Jordan index, sensitivity, specificity Predict the probability p value, calculate the Y value according to p = 1 / 1 + e-y, and finally draw the nomogram diagram. Results: All 394 patients completed surgical resection of thyroid cancer at one time. 41 patients developed vocal cord paralysis after operation. The incidence of vocal cord paralysis was 10.4%. Logistic regression model showed that whether the recurrent laryngeal nerve entered the larynx (OR = 11.804, 95% CI 3.078 ~ 45.273), preoperative TG (OR = 0.021, 95% CI 0.002 ~ 0.202), whether it was close to the recurrent laryngeal nerve (OR = 20.984, 95% CI 2.058 ~ 214.007), and operation time (OR = 2.768, 95% CI 1.122 ~ 6.829) were independent predictors of vocal cord paralysis after thyroid cancer surgery. Ten fold cross validation showed that the area under the ROC curve was 0.7284, the area under the ROC curve in the modeling group was 0.7943 (95% confidence interval was 0.716 ~ 0.872), and the area under the ROC curve in the validation group was 0.7722 (95% confidence interval was 0.526 ~ 1); Hosmer lemeshow goodness of fit test showed that the model fitted well, chi square = 1.1, p = 0.9816 > 0.05 in the modeling group and chi square = 3.87, p = 0.5677 > 0.05 in the validation group; The maximum youden's?index is 0.188, the sensitivity is 0.57, the specificity is 0.62, the area under the ROC curve is 0.59,p = 0.216, which is the best critical value. According to p = 1 / 1 + e-y, y = -1.25. Conclusion: The prediction accuracy of this regression model is good, and it has a certain reference significance for the occurrence of vocal cord paralysis after thyroid cancer surgery.