构建食管癌(EC)患者围放疗期营养不良风险的列线图预测模型,并评估模型的预测效能。 方法 选择 2019 年 3 月至 2022 年 3 月在徐州医科大学附属淮安医院行放疗的 194 例 EC 患者,于放疗结束后 1 d 以主观整体营养状况评估 表(PG-SGA)并结合血清白蛋白水平进行营养不良筛查,根据是否存在营养不良分成营养不良组(PG-SGA 评分≥4 分且血 清白蛋白<35 g / L)和营养良好组(PG-SGA 评分<4 分或血清白蛋白≥35 g / L)。 利用多因素 Logistic 回归分析确定 EC 患者围 放疗期营养不良风险的影响因素,并基于预测因素构建列线图预测模型,随后用 ROC 曲线下面积(AUC)和 H-L 拟合优度对 模型的区分度和精准度进行验证。 结果 83 例 EC 患者出现营养不良,营养不良发生率为 42. 78%;单因素分析显示,营养不良 组与营养良好组年龄、消瘦、家庭人均月收入、是否接受手术、肿瘤分期、伴随消化道症状数量、同步化疗、吞咽障碍、抑郁等 9 项因素比较有显著差异(P<0. 05);Logistic 回归分析结果显示,EC 患者围放疗期营养不良发生的危险因素有年龄≥60 岁、消 瘦、肿瘤分期>Ⅱ期、伴随消化道症状数量≥2 个、同步化疗、吞咽障碍等 6 项(P<0. 05),EC 患者围放疗期营养不良评估回归公 式 Logit(P)= -11. 563+0. 958×年龄+1. 167×消瘦+0. 915×肿瘤分期+0. 903×伴随消化道症状数量+1. 306×同步化疗+0. 839×吞 咽障碍;将上述 6 项危险因素引入 R 软件建立列线图模型,AUC 为 0. 786>0. 75(95%CI = 0. 722~ 0. 850),校准曲线和理想曲线 拟合较好,且拟合优度 H-L 检验 χ 2 = 7. 114,P= 0. 524,表明列线图模型预测 EC 患者围放疗期发生营养不良的区分度及精准 度均较好。 结论 基于年龄≥60 岁、消瘦、肿瘤分期>Ⅱ期、伴随消化道症状数量≥2 个、同步化疗、吞咽障碍等 6 项因素构建的 列线图预测模型对 EC 患者围放疗期营养不良风险具有良好的预测作用,可为肿瘤科医护人员早期对食管癌患者围放疗期采 取预防性护理干预提供借鉴。

To construct a nomogram prediction model for the risk of malnutrition in patients with esophageal cancer EC during periradiotherapy and to evaluate the predictive performance of the model. Method From March 2019 to March 2022 194 EC patients who underwent radiotherapy in our hospital were taken One day after radiotherapy scored Patient-generated Subjective Global Assessment PG-SGA and serum albumin level were used for malnutrition screening. According to the presence or absence of malnutrition they were grouped into a malnutrition group PG-SGA score ≥4 and serum albumin <35 g / L and a good nutrition group PG-SGA score <4 or serum albumin ≥35 g / L . Multivariate Logistic regression analysis was applied to determine the influencing factors of malnutrition risk in EC patients during periradiotherapy and a nomogram prediction model was constructed based on the predictive factors and then the discriminativeness and precision of the model were verified by the area under the ROC curve AUC and H - L goodness of fit. Result Malnutrition occurred in 83 EC patients and the incidence of malnutrition was 42. 78% univariate analysis showed that there were significant differences between the malnutrition group and the good nutrition group in 9 factors including age thin monthly household income per capita whether to undergo surgery tumor stage the number of accompanying gastrointestinal symptoms concurrent chemotherapy dysphagia and depression P < 0. 05 Logistic regression analysis showed that the risk factors of malnutrition in EC patients during periradiotherapy were age ≥ 60 years old weight loss tumor stage > stage Ⅱ number of accompanying gastrointestinal symptoms ≥ 2 concurrent chemotherapy and dysphagia P < 0. 05 regression formula of malnutrition assessment in EC patients during periradiotherapy Logit P = -11. 563+0. 958×age+1. 167×weight loss+ 0. 915 × tumor stage + 0. 903 × number of accompanying gastrointestinal symptoms + 1. 306 × concurrent chemotherapy + 0. 839 × dysphagia the above 6 risk factors were introduced into R software to establish a nomogram model the AUC was 0. 786 > 0. 75 95%CI = 0. 722-0. 850 the calibration curve and the ideal curve fit well and goodness-of-fit H-L test χ 2 = 7. 114 P = 0. 524 indicating that the nomogram model has good discrimination and accuracy in predicting malnutrition in EC patients during periradiotherap. Conclusion The nomogram prediction model constructed based on six factors including age ≥60 years weight loss tumor stage >Ⅱ number of accompanying gastrointestinal symptoms ≥2 concurrent chemotherapy and dysphagia has a good effect on the risk of malnutrition in EC patients during periradiotherapy The predictive role of esophageal cancer can provide reference for oncology medical staff to take preventive nursing intervention in the periradiotherapy period of esophageal cancer patients in a timely manner.