摘要
目的 构建预测早期喉癌术后吞咽障碍的列线图模型.方法 回顾性分析南阳市中心医院2020年3月至2022年5月收治的154例早期喉癌患者的临床资料,根据术后2周内是否发生吞咽障碍分成病例组和对照组.开展单因素分析,通过Logistic回归分析患者术后吞咽障碍的影响因素,利用R软件并基于预测变量构建预测吞咽障碍发生的列线图模型.以ROC曲线下面积(AUC)评估模型的预测效果,并行HL拟合优度检验.结果 154例患者术后发生吞咽障碍49例(31.82％);病例组年龄≥60岁、有吸烟史、有咽喉反流性疾病史、手术方式为常规喉裂开术、张口受限程度为3度、术后1周疼痛程度为重度的患者比例和对照组比较差异存在统计学意义(P＜0.05);Logistic回归分析显示,早期喉癌患者术后吞咽障碍的危险因素包括年龄≥60岁、手术方式为常规喉裂开术、咽喉反流史、张口受限程度为3度、术后1周疼痛度为重度等5项(P＜0.05);将上述危险因素引入R软件构建列线图模型,结果显示AUC为0.759(95％CI=0.674～0.844),拟合优度HL检验x2=10.293,P=0.173.结论 基于年龄、咽喉反流病史、手术方式、张口受限程度、术后1周疼痛度等5项因素构建的列线图模型可较好地预测早期喉癌患者术后吞咽障碍发生风险.
Abstract
Objective To construct a nomogram model for predicting dysphagia after early laryngeal cancer.Methods The clinical data of 154 patients with early laryngeal cancer who were admitted to Nanyang Central Hospital from March 2020 to May 2022 were retrospectively analyzed,and the patients were grouped into a case group and a control group according to whether dysphagia occurred 2 weeks after the operation.Univariate analysis was carried out,Logistic regression was performed to analyze the influencing factors of postoperative dysphagia in patients with vocal cord sur-gery,a nomogram model for predicting the occurrence of dysphagia was constructed using R software and based on pre-dictors.The prediction effect of the model was evaluated by the area under the ROC curve(AUC),and the HL goodness of fit test was performed.Results Among 154 patients,49 patients(31.82％)developed dysphagia after operation;there were obvious differences in the proportions of patients aged≥60 years,with a history of smoking,the surgical method was conventional laryngeal dehysis,a history of pharyngeal reflux disease,restricted mouth opening degree of 3 degrees,and severe pain at 1 week after surgery between the case group and the control group(P＜0.05);Logistic re-gression analysis showed that the risk factors for postoperative dysphagia in patients with early laryngeal cancer includ-ed age≥60 years,the surgical method was conventional laryngeal dehysis,history of laryngeal reflux,restricted mouth opening degree of 3 degrees,and severe pain at 1 week after surgery(P＜0.05);the above risk factors were introduced into R software to construct a nomogram model,the results showed that the AUC was 0.759(95％CI=0.674～0.844),and the goodness of fit HL test was x2=10.293,P=0.173.Conclusion The nomogram model constructed based on five factors including age,history of laryngopharyngeal reflux,surgical method,restriction of mouth opening,and degree of pain at 1 week after operation can better predict the risk of dysphagia in patients with early laryngeal cancer.
基金项目
河南省医学适宜技术推广项目(SYJS2022054)
NETL
NSTL
万方数据