肿瘤患者肌少症的危险因素分析及模型建立

Risk factors analysis and model establishment of sarcopenia in tumor patients

目的:分析肿瘤患者中肌少症的危险因素并构建风险预测模型。方法:收集我院放疗科2022年6月至2023年6月136例肿瘤患者,基于CT图像诊断肌少症。将其分为肌少症组57例、非肌少症组79例。通过单因素分析筛查出相关变量,Logistic回归排除混杂因素并得出风险预测模型,根据ROC曲线下面积(AUC)评估模型的灵敏度、特异度及截断值,并使用Bootstrap法对模型进行内部验证。结果:肿瘤患者中,肌少症相关独立预测因子分别为性别(OR=0.037,95%CI:0.006~0.213,P<0.001)、BMI(OR=0.811,95%CI:0.675~0.975,P=0.026)、握力值(OR=0.792,95%CI:0.693~0.904,P<0.001)和步行速度(OR=0,95%CI:0~0.014,P<0.001),得出风险预测模型为Logit(P)=23.304-3.304×性别-0.209×BMI+0.234×握力值-8.873×步行速度,AUC模型=0.943(95%CI:0.908~0.978,P<0.001),该模型灵敏度和特异度分别为88%和86%,截断值为0.037。结论:本研究所构建的模型可对肿瘤患者进行肌少症的风险预测,且具有无创、简便的优点,为临床医生诊断以及患者自检提供帮助和便利。

Objective:To analyze the risk factors of sarcopenia in tumor patients and construct a risk prediction model.Methods:136 tumor patients from the radiotherapy department of our hospital from June 2022 to June 2023 were collected and diagnosed with sarcopenia based on CT images.They were divided into a sarcopenia group of 57 cases and a non-sarcopenia group of 79 cases.Relevant variables were screened through univariate analysis,and a risk prediction model was obtained through Logistic regression.The sensitivity,specificity,and cutoff values of the model were evaluated based on the area under the ROC curve(AUC),and the model was internally validated using the Bootstrap method.Results:Among tumor patients,the independent predictors of sarcopenia were gender(OR=0.037,95%CI:0.006~0.213,P<0.001),BMI(OR=0.811,95%CI:0.675~0.975,P=0.026),grip strength(OR=0.792,95%CI:0.693~0.904,P<0.001),and walking speed(OR=0,95%CI:0~0.014,P<0.001).The risk prediction model was obtained as follows:Logit(P)=23.304-3.304×gender-0.209×BMI+0.234×grip strength-8.873×walking speed.The AUC model was 0.943(95%CI:0.908~0.978,P<0.001),with sensitivity and specificity of 88%and 86%,respectively,and a cutoff value of 0.037.Conclusion:The model constructed in this study can predict the risk of sarcopenia in tumor patients,and has the advantages of non-invasive and simple,providing assistance and convenience for clinical doctors'diagnosis and patient self examination.