Study On Construction The NLR-based Logistic Regression Prediction Model Of Respiratory Failure In Lung Cancer Patients Based On MIMIC-Ⅲ Database

Background :Lung cancer is one of the most important causes of all cancer-related deaths,and the prognosis for advanced lung cancer is poor.The occurrence of tumor progression and complications was the direct cause of death.neutrophil lymphocyte ratio(NLR)in peripheral blood is a simple and easily accessible biomarker that reflects the immune activation and inflammation of the body and serves as an indicator of the early stage of pathophysiology.NLR takes precedence over clinical changes.The purpose of this study is to explore the risk factors related to respiratory failure in lung cancer patients,calculate the incidence of high-risk groups through prediction models,so as to identify lung cancer groups with high incidence of respiratory failure at an early stage,make targeted and timely intervention,choose the treatment plan with great overall benefit,and prevent the deterioration of the disease.Objective:The predictive significance of NLR for early identification of respiratory failure in lung cancer patients,and the logistic regression prediction model of lung cancer patients,which is constructed in the form of nomogram.Methods:The data of basic characteristics,blood draw,vital signs,respiratory failure,endotracheal intubation and death were extracted from MIMIC-Ⅲ database,the optimal cut-off value of NLR was calculated by the ROC curve analysis of NLR and outcome indicators,the cases were divided into low NLR and high NLR groups,compare the indicators between the two groups,and logistic regression was used to select independent risk factors related to the occurrence of respiratory failure,and then a nomogram prediction model was constructed.Results:Among the 994 patients,tracheal intubation rate was significantly higher in the high NLR group than in the low NLR group(30.3% vs 23.0%),and respiratory failure rate in the high NLR group was significantly higher than in the low NLR group(44.7% vs 28.5%).Results of multivariate logistic regression analysis showed that:NLR(OR=1.012,95%CI:1.002～1.023,P=0.022)、wbc(OR=1.049,95%CI:1.007～1.053,P=0.01)、alb(OR=0.569,95%CI:0.448～0.721,P<0.001)、COPD(OR=2.094,95%CI:1.322～3.318,P=0.002)、BP(OR=0.984,95%CI:0.972～0.996,P=0.009)、PH(OR=0.064,95%CI:0.015～0.273,P<0.001)was associated with the occurrence of respiratory failure,Was an independent influencing factor of respiratory failure(p<0.05),The contributing factors were NLR,wbc,COPD,The protective factors were alb,BP,and PH.The area under the ROC curve is 0.711.The model uses internal resampling 100 times,so that the calibration prediction curve matches the ideal curve well,suggesting that the model is accurate.The clinical efficacy analysis suggested that the model has good clinical utility.Conclusions:For lung cancer patients with elevated NLR,elevated white blood cell count,decreased albumin,low blood pressure,low blood gas PH,and patients with basic diseases of chronic obstructive pneumonia,the probability of respiratory failure is higher.The nomogram drawn in this study can assess the risk of respiratory failure in patients with lung cancer early.