口腔癌患者行根治切除术后中重度营养不良的综合预测模型构建

摘要
图/表
参考文献(35)
相关文章 (15)

全文: PDF (2635 KB) HTML (1 KB)
输出: BibTeX | EndNote (RIS)

摘要目的分析口腔癌患者行根治切除术后中重度营养不良相关影响因素,并构建综合预测模型。方法选取 2022 年 1 月至 2024 年 1 月期间在首都医科大学附属北京口腔医院诊治的 195 例口腔癌根治切除术患者作为研究对象,术后 2 周评估口腔癌患者的营养状况,按照患者主观整体评估(PG-SGA)评分分为中重度营养不良组(117 例,PG-SGA 评分≥4 分)和无或轻度营养不良组(78 例,PG-SGA 评分<4 分),分析口腔癌患者行根治切除术后中重度营养不良的危险因素,并通过危险因素构建回归模型及决策树模型对术后中重度营养不良进行综合评估。结果经二元 Logisic 回归分析,临床分期Ⅲ期(OR = 6. 114,95%CI = 1. 667 ~ 22. 423)、临床分期Ⅳ期(OR = 6. 537,95% CI = 1. 656 ~ 25. 797)、口腔治疗史(OR = 3. 387,95% CI = 1. 388~ 8. 263)、喜欢嚼槟榔(OR = 3. 783,95%CI = 1. 550 ~ 9. 232)、气管切开(OR = 3. 120,95%CI = 1. 196 ~ 8. 139)、术后放疗 (OR= 21. 625,95%CI = 6. 457 ~ 72. 421)、术后抑郁(OR = 3. 832,95%CI = 1. 506 ~ 9. 751)、术后睡眠障碍(OR = 3. 940,95%CI = 1. 600~ 9. 704)、术后吞咽障碍(OR= 3. 474,95%CI = 1. 427~ 8. 462)是口腔癌患者行根治切除术后中重度营养不良的危险因素 (P<0. 05)。模型的 Hosmer-Lemeshow 拟合度检验显示模型拟合优度较好( χ 2 = 5. 613,P = 0. 691);受试者操作特征(ROC)曲线分析得出,模型预测口腔癌患者行根治切除术后中重度营养不良的 ROC 曲线下面积为 0. 904(95%CI = 0. 862~ 0. 945),敏感度为 81. 2%,特异度为 83. 3%,约登指数为 0. 645,实际应用准确性为 82. 1%。以是否发生术后中重度营养不良为因变量,以危险因素为自变量构建口腔癌患者行根治切除术后中重度营养不良的决策树风险模型,误判率为 21. 0%,预测准确率为 79. 0%。结论通过临床分期、口腔治疗史、喜欢嚼槟榔、气管切开及术后放疗、抑郁、睡眠障碍、吞咽障碍构建的 Logistic 回归模型及决策树模型对口腔癌患者行根治切除术后中重度营养不良具有较高的预测价值,可为临床减少口腔癌根治术患者术后中重度营养不良提供辅助参考。

	服务

	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	凌晓彤
	李德龙
	陈伟
	付佳

关键词 ： Logistic 回归模型, 决策树模型, 口腔癌, 根治切除术, 营养不良, 抑郁, 睡眠障碍, 吞咽障碍

Abstract：Objective To analyze the influencing factors of moderate and severe malnutrition in oral cancer patients after radical resection and to build a comprehensive prediction model. Method A total of 195 patients who underwent radical resection of oral cancer treated in Capital Medical University Affiliated Beijing Stomatological Hospital from January 2022 to January 2024 as the research subjects. Two weeks after surgery they were divided into a moderate and severe malnutrition group of 117 cases and a non malnutrition group of 78 cases according to their nutritional status. The risk factors for moderate and severe malnutrition in oral cancer patients after radical resection were analyzed and a regression model and decision tree model were constructed to comprehensively evaluate postoperative moderate and severe malnutrition. Result According to binary Logistic regression analysis clinical stage Ⅲ OR = 6. 114 95%CI = 1. 667-22. 423 clinical stage Ⅳ OR = 6. 537 95%CI = 1. 656-25. 797 oral treatment history OR = 3. 387 95%CI = 1. 388- 8. 263 love of chewing betel nut OR = 3. 783 95% CI 1. 550- 9. 232 tracheotomy OR = 3. 120 95% CI = 1. 196-8. 139 postoperative radiotherapy OR = 21. 625 95%CI 6. 457-72. 421 postoperative depression OR = 3. 832 95% CI = 1. 506 - 9. 751 postoperative sleep disorders OR = 3. 940 95% CI = 1. 600 - 9. 704 and postoperative swallowing disorders OR= 3. 474 95%CI = 1. 427-8. 462 are risk factors for moderate and severe malnutrition in patients with oral cancer undergoing radical resection P<0. 05 . The model's Hosmer Lemeshow fitting test showed a chi square of 5. 613 and a P-value of 0. 691. ROC curve analysis revealed that the AUC of the model for predicting moderate and severe malnutrition in oral cancer patients after radical resection was 0. 904 95%CI = 0. 862-0. 945 with a sensitivity of 81. 2% specificity of 83. 3% Youden index of 0. 645 and an accuracy of 82. 1% in practical applications. A decision tree risk model for postoperative moderate and severe malnutrition in oral cancer patients undergoing radical resection was constructed with the occurrence of postoperative moderate and severe malnutrition as the dependent variable and risk factors as independent variables. The misjudgment rate was 21. 0% and the prediction accuracy was 79. 0%. Conclusion Malnutrition and clinical stage oral treatment history preference for chewing betel nut tracheotomy and postoperative radiotherapy depression sleep disorders and swallowing disorders are related to patients with oral cancer undergoing radical resection. Logistic regression models and decision tree models have high predictive value for moderate and severe malnutrition in patients with oral cancer undergoing radical resection.

Key words： Logistic regression model Decision tree model Mouth cancer Radical resection surgery Innutrition Depression Sleep disorders Swallowing disorders

基金资助:北京市属医院科研培育计划项目(PX2024056)

通讯作者: 陈伟,电子邮箱 nedleschenwei@ 163. com

引用本文:

凌晓彤,李德龙,陈伟,付佳. 口腔癌患者行根治切除术后中重度营养不良的综合预测模型构建[J]. 肿瘤代谢与营养电子杂志, 2025, 12(2): 221-230.
Ling Xiaotong, Li Delong, Chen Wei, Fu Jia. Construction of a comprehensive prediction model for moderate and severe malnutrition in patients with oral cancer after radical resection. Electron J Metab Nutr Cancer, 2025, 12(2): 221-230.