摘要
目的 基于人工神经网络模型,探讨电动吸切器腺样体切除术治疗慢性鼻窦炎的疗效及术后出血的影响因素。方法 回顾性分析2019年9月-2022年9月该院收治的343例慢性鼻窦炎患者的临床资料,将采用电动吸切器腺样体切除术治疗的患者设为观察组(275例),将采用腺样体刮除术治疗的患者设为对照组(68例)。按照随机数表法,将观察组患者以4∶1的比例,分为训练集(220例)和验证集(55例);根据术后是否出血,将训练集患者分为术后出血组(23例)和术后未出血组(197例)。分析电动吸切器腺样体切除术治疗慢性鼻窦炎的疗效,以及术后并发症发生情况。比较术后出血组和术后未出血组患者的临床资料,并分析患者术后出血的影响因素。通过Python 3.6软件,构建人工神经网络模型,并对模型的效能进行验证。结果 观察组治疗慢性鼻窦炎的总有效率明显高于对照组,差异有统计学意义(χ^(2)=4.46,P=0.035)。观察组术后出血率、腺样体残留率和并发症总发生率明显低于对照组,差异均有统计学意义(χ^(2)=8.90,P=0.003;χ^(2)=8.43,P=0.004;χ^(2)=15.42,P=0.000)。年龄≤7岁、Lund-Mackay评分> 14分、手术操作经验不足、合并临近器官炎症、凝血功能异常、手术切割过深、术中止血不彻底、术中副损伤、使用肾上腺素和术后感染,是慢性鼻窦炎患者电动吸切器腺样体切除术后出血的独立危险因素(P <0.05)。绘制训练集和验证集的受试者操作特征曲线(ROC curve),曲线下面积(AUC)分别为0.842 (95%CI:0.786~0.898)和0.786 (95%CI:0.719~0.853),特异度分别为92.64%和90.17%,灵敏度分别为92.55%和90.06%;C指数分别为0.859 (95%CI:0.702~0.943)和0.872 (95%CI:0.723~0.981),两集的拟合效果良好。结论 电动吸切器腺样体切除术治疗慢性鼻窦炎的疗效显著,且术后出血等并发症发生率低,是治疗慢性鼻窦炎安全、有效的方式。人工神经网络模型可为临床研究术后出血的影响因素提供帮助。
Objective To discuss the efficacy of adenoidectomy with electric aspirator in the treatment of chronic sinusitis and the influencing factors of postoperative bleeding.Methods Retrospective analysis of the clinical data of 343 patients with chronic sinusitis from September 2019 to September 2022.The patients treated with adenoidectomy with electric aspirator were set as the observation group(275 cases).The patients treated with adenoidectomy were set as control group(68 cases).The patients in the observation group were divided into training set(220 cases)and verification set(55 cases)at a ratio of 4∶1 according to the random number table method.The patients in the training set were divided into postoperative bleeding group(23 cases)and postoperative non bleeding group(197 cases)according to whether there was postoperative bleeding or not.To analyze the curative effect and postoperative complications of adenoidectomy with electric aspirator for chronic sinusitis.The clinical data of patients with postoperative bleeding and those without postoperative bleeding were compared,and the influencing factors of postoperative bleeding were analyzed.Building artificial neural network model by Python 3.6 software,and verifing the effectiveness of the model.Results The total effective rate of observation group was higher than that of the control group,with a statistically significant difference(χ^(2)=4.46,P=0.035).Postoperative bleeding rate,adenoid residual rate and total incidence of complications of observation groups were lower than those of the control group,with statistically significant differences(χ^(2)=8.90,P=0.003;χ^(2)=8.43,P=0.004;χ^(2)=15.42,P=0.000).Age≤7 years old,Lund-Mackay score>14 points,insufficient surgical experience,inflammation of adjacent organs,abnormal coagulation function,deep surgical cutting,incomplete hemostasis during operation,side injury during operation,use of adrenaline,and postoperative infection were independent risk factors for bleeding after adenoidectomy with electric aspirator(P<0.05).The area under the curve(AUC)of the receiver operator characteristic curve(ROC)of the training set and the validation set were 0.842(95%CI:0.786~0.898)and 0.786(95%CI:0.719~0.853),respectively;The specificities of the curve were 92.64%and 90.17%,respectively;The sensitivities of the curve were 92.55%and 90.06%respectively;The C-index values were 0.859(95%CI:0.702~0.943)and 0.872(95%CI:0.723~0.981),respectively.The fitting responses of the two sets were good.Conclusion The effect of adenoidectomy with electric aspirator on chronic sinusitis is significant.The incidence of complications such as postoperative bleeding is low.It is a safe and effective way to treat chronic sinusitis.Artificial neural network models can provide assistance for clinical research on the influencing factors of postoperative bleeding in patients.
作者
管俊华
孙饶希
Guan Junhua;Sun Raoxi(Department of Otolaryngology,the First People’sHospital of Shuangliu,Chengdu,Sichuan 610200,China;Department of Otolaryngology,Chengdu Second People’s Hospital,Chengdu,Sichuan 610000,China)
出处
《中国内镜杂志》
2023年第11期53-60,共8页
China Journal of Endoscopy
关键词
电动吸切器腺样体切除术
慢性鼻窦炎
术后出血
影响因素
人工神经网络模型
adenoidectomy with electric aspirator
chronic sinusitis
postoperative bleeding
influencing factors
artificial neural network model
