Risk Factors Analysis And Clinical Prediction Model Construction Of Chronic Obstructive Pulmonary Disease Complicated With Lung Cancer

【Background】The incidence of lung cancer in patients with chronic obstructive pulmonary disease（COPD）is significantly increased.In clinical diagnosis and treatment,most of patients with COPD and lung cancer are diagnosed at an advanced stage,with poor prognosis and high mortality.This may be related to the symptoms and signs of COPD patients,which may reduce the vigilance of patients and doctors to lung cancer,resulting in delayed diagnosis of lung cancer.Therefore,improving the early diagnosis rate of COPD combined with lung cancer has important clinical significance for improving the prognosis of patients.Most of the existing studies only describe the clinical characteristics of COPD patients with lung cancer,and have not yet established an effective prediction model for COPD complicated with lung cancer.【Objective】To explore the clinical features of COPD and the risk factors of COPD patients with lung cancer,this study constructs a clinical prediction model for the risk of lung cancer in COPD patients,which can provide reference for clinicians to make decisions,improve the early diagnosis rate of lung cancer,and reduce clinical misdiagnosis and missed diagnosis.【Method】1.A total of 300 patients with COPD complicated with lung cancer and simple COPD admitted to A tertiary A hospital in Northwest China from January 2015 to October 2022 were retrospectively collected as training sets.We collected the general conditions,main clinical symptoms,serological indexes（including blood routine,coagulation series,inflammatory indexes and tumor markers）,arterial blood gas analysis,lung function indexes and chest CT data of COPD patients before the initial diagnosis of lung cancer.The clinical characteristics of COPD patients with lung cancer were explored by difference comparison,and the independent risk factors of lung cancer in COPD patients were screened by multivariate Logistic regression analysis.The prediction model of lung cancer risk in COPD patients was established and the formula of the prediction model was listed.At the same time,R software was used to make a nomogram to visualize the model.Receiver operating characteristic（ROC）curve and C-index were used to evaluate the discrimination of the model.Calibration curve and goodness of fit test were used to evaluate the calibration of the model.Decision curve analysis（DCA）was used to evaluate the clinical effectiveness of the model.The model was internally validated using repeated sampling.2.A total of 110 patients with COPD complicated with lung cancer and COPD alone admitted to A tertiary hospital B in Northwest China from January 2021 to December 2022were retrospectively collected as the validation set.This is used as an external validation of the prediction model.【Results】1.Among the 300 patients in the training set,151 patients with COPD and lung cancer were in the case group,including 136 males and 15 females,with the age ranged from 46 to86 years（an average age of 66.36±7.61 years）.A total of 149 patients with simple COPD were selected as the control group,including 121 males and 28 females,with the age ranged from 37 to 85 years（an average age of 64.36±8.73 years）.The location of lung cancer in COPD combined with lung cancer group was more common in the left upper lobe.The most common pathological type was squamous cell carcinoma（49%）,followed by adenocarcinoma（28.5%）.Squamous cell carcinoma（52.9%）was the most common in male COPD patients,while adenocarcinoma（53.3%）was the most common in female COPD patients.tumor-node-metastasis（TNM）stage IV（42%）is the most common stage,and most patients have metastasis at the time of initial diagnosis.2.Univariate analysis showed that a total of 37 predictors were significantly different between the two groups（P<0.05）,including gender,age,smoking,duration of COPD,and proportion of patients using inhaled corticosteroid（ICS）.Chest pain,hemoptysis;white blood cell count,red blood cell count,platelet count,hemoglobin concentration,neutrophil count,neutrophil percentage,lymphocyte percentage,neutrophil to lymphocyte ratio,platelet to lymphocyte ratio;fibrinogen,fibrinogen degradation products,D-dimer,prothrombin time;the proportion of elevated erythrocyte sedimentation rate（ESR）,procalcitonin,and C-reactive protein;carcinoembryonic antigen,cytokeratin 19 fragment,neuron specific enolase,carbohydrate antigen-125,squamous cell carcinoma-associated antigen;the forced expiratory volume in the first second,forced vital capacity,forced expiratory volume in the first second as a percentage of the predicted value,forced vital capacity as a percentage of the predicted value,forced expiratory volume in the first second as a percentage of forced vital capacity,residual volume,residual volume/total lung capacity（RV/TLC）,Global Initiative for Chronic Obstructive Lung Disease（GOLD）classification;emphysema.3.Multivariate logistic regression analysis showed that COPD duration,smoking,RV/TLC,ESR,ICS treatment,and GOLD grade were independent risk factors for lung cancer in COPD patients.The duration of COPD,smoking,RV/TLC≥40%,GOLDI/II,ESR≥20mm/h were independent risk factors for lung cancer in COPD patients,and ICS treatment was a protective factor for lung cancer.4.Based on the above predictors,a predictive model for the risk of lung cancer in COPD patients was established.The formula was calculated as follows:P=e^x/（1+e^x）,X=-0.877+0.569×COPD duration（<10 years=1;≥10years,<20years=2;≥20years=3）+1.459×smoking（no=0,yes=1）-0.715×ICS（no=0,yes=1）+1.149×RV/TLC（<40%=0,≥40%=1）-0.751×GOLD classification（FEV₁%pred≥80%=1;50%≤FEV₁%pred<80%=2;30%≤FEV₁%pred<50%=3;FEV₁%pred<30%=4）+1.585×ESR（ESR<20mm/h=0,≥20mm/h=1）,where e is the natural logarithm,e=2.718281828.And draw a nomogram,and evaluate and verify the model.The area under the ROC curve of the model in the training set was 0.812（95%confidence interval:0.764-0.860）,the area under the ROC curve of the external validation set was 0.798（95%confidence interval:0.711-0.884）,and the Bootstrap method was sampled 1000 times for internal validation.The consistency index（C-index）was 0.7908,which indicated that the risk prediction model had good discrimination ability.The calibration curves of the training set and the validation set model were close to 45°,and the goodness of fit test was P>0.05,suggesting that the model had good calibration.The DCA curves showed that the model curve was higher than the two extreme lines,suggesting that the model had good clinical practicability.【Conclusion】1.COPD combined with lung cancer are more common in elderly patients with a longer duration of COPD,low GOLD grade of lung function,less ICS treatment of COPD,and long-term smoking complicated with emphysema,who should be screened regularly.2.When patients with COPD have chest pain,hemoptysis or blood in sputum,and tumor markers and inflammatory markers are higher than before,they should be considered with the higher risk of lung cancer.3.The duration of COPD,smoking,RV/TLC,ESR,ICS treatment and GOLD classification were independent predictors of lung cancer risk in COPD patients.4.The clinical prediction model established in this study has good discrimination,calibration and clinical validity,which can be used as a quantitative tool to evaluate the risk of lung cancer in COPD patients.