| Background:Solitary pulmonary nodule(SPN)generally refer to isolated,round or circular,opaque,and<3 cm in diameter lesions existed in the lungs.The nodules are completely surrounded by lung parenchyma,without mediastinum or hilar lymph node enlargement,pleural effusion and atelectasis[1].With the rapid development of medical imaging technology,the detection rate of SPN has been continuously improved.How to make a definite diagnosis of benign and malignant SPN has always been a challenge for respiratory medicine physicians,thoracic surgery physicians and radiologists.If we can diagnose and treat SPN early,we will greatly improve the overall survival rate of lung cancer patients and improves the prognosis[2,3]Therefore,the correct judgment of benign and malignant lung nodules has become the key to clinical diagnosis and treatment,the ideal state is to make benign nodules to avoid unnecessary invasive examinations and sugery,while malignant nodules can be diagnosed at an early stage and treatment,so that patients achieve maximum benefit.At present,the diagnosis of SPN without relying on pathological examination mainly depends on the comprehensive analysis of the patient's clinical and imaging data.This has a very high requirement on the doctor's theoretical level and clinical experience.Therefore,in order to reduce the interference of human factors and improve the diagnostic accuracy,relevant models have been established based on clinical data and image features to predict the probability of SPN malignancy and guide physicians to choose the next step treatment method[4,5].Domestic and foreign scholars have established models for the diagnosis of benign and malignant solitary pulmonary nodules.However,there are no uniform standards and the risk factors adopted are not the same.The accuracy of mathematical model prediction is also different.Most SPN prediction models are established by the patient's general clinical data combined with imaging features.A very small number of models have been included in the study of lung tumor signage properties.At present,there are few models included Platelet related markers in the study.The tumor marker is an important method for lung cancer screening and early and differential diagnosis.Platelet related markers mentioned in the relevant literature can identify benign and malignant tumors,although its specificity is not high.However,there is still hope that it will become a marker for determining benign and malignant SPNs,and both tumor markers and Platelet related markers are more easily obtained in clinical practice.Therefore,this study intends to establish a diagnostic model of SPN and include tumor markers and platelet markers at the data collection stage.Collecting the data from general clinical of patients,scientific imaging methods and body fluid collection,in multidimensional dimensions to advance and observe the platelet markers,cloud it be used as an independent risk factor to predict the benign and malignant lung nodules?Objectives:1.Through the univariate and multivariate analyses performed on clinical data,imaging data,and tumor markers of patients with pulmonary nodules.Risk factors for benign and malignant lung nodules were screened out and a predictive model for the likelihood of pulmonary nodules was established.To reduce human factors in clinical work and provide suggestions for further treatment of solitary pulmonary nodules;2.To explore the value of platelet-related indicators in the diagnosis model of solitary pulmonary nodules;3.To investigate the role of imaging features of pulmonary nodules in the diagnosis model of solitary pulmonary nodules;4.Establish Logistic regression equation,analyze the data of solitary pulmonary nodules,select suitable predictive values for the analysis of benign and malignant SPN,and select the best treatment plan for patients.Methods:1.Clinical data:We collected 246 SPN patients who had a clear pathological diagnosis after surgical resection or biopsy from January 2014 to December 2017.Inclusion criteria were as follows:1.1 Within the range of ?3 cm in diameter for single-segment and circular lesions,nodules were completely surrounded by pulmonary parenchyma,without mediastinum or hilar lymph node enlargement,pleural effusion and atelectasis;1.2 CT-guided puncture,fiberoptic bronchoscopic biopsy,or surgery and a clear pathological diagnosis of the case;1.3 A complete clinical records and CT imaging data.Collecting the patient's gender,age,smoking history,smoking index,past cancer history,family cancer history;2.Blood data:Peripheral venous blood was collected from patients with SPN,and five tumor markers of serum CEA,CA125,CA153,Cyfra-21,and NSE were performed according to laboratory requirements;Detection of four platelet markers:PLT,MPV,P-LCR,and PDW(all completed by the First Affiliated Hospital of Kunming Medical University);3.Imaging data:Using multi-slice spiral CT plain scan or enhanced scan,collecting patient data are as follows:1,isolated nodule location;2,the largest diameter of nodules;3,whether nodular boundary is clear or not;4,whether there is no vascular,calcified,bronchial vasodilatation sign,pleural depression sign;5,nodule burr length,nodule depth of leaf.4.SPSS 17.0 statistical software package was used for statistical analysis.Measured data with normal distribution were expressed as mean ± standard deviation.Independent sample t-test was used for comparison between the two groups;Non-normal distribution measurement data was Z((P25,P75)and non-parametric tests were used for comparison between groups;x2 test was used for the analysis of count data;logistic regression was used to analyze the value of assignments,screened for independent risk factors,and a regression equation was constructed to predict the likelihood of solitary pulmonary nodules malignancy model.The ROC curve was drawn to select the appropriate critical value of benign and malignant probabilities,and the sensitivity,specificity,positive predictive value and negative predictive value were calculated.The difference was statistically significant at P<0.05.Results:1.Find out through single factor analysis:The patient's smoking index(X2 = 4.034,P = 0.045),age(t =-3.636,P<0.001),burr(X2 = 3.864,P = 0.049),nodule diameter(t =-2.43,P = 0.017),P-LCR(t---2.591,P=0.010),PDW(t--3.975,P<0.001),CEA(z=-3.202,P=0.001),NSE(z=-2.226,P=0.026))There are statistical differences between the benign and malignant lung nodules.2.Multivariate logistic regression analysis shows:The patient's age(OR 1.04,P=0.006),nodule diameter(OR 1.56,P=0.008),burr(OR 1.98,P=0.012),NSE(OR 1.12,P=0.015),P-LCR(OR 1.12,P=0.003)and PDW(OR 1.32,P=0.001)were statistically significant differences between benign and malignant lung nodules and were:independent risk factors for benign and malignant lung nodules.3.Establish a predictive model for malignant lung nodules:Model:Malignancy prediction(P)=ex/(1+ex),X=-2.98+(0.04xage)+(0.08xnodule diameter)+(0.31×burr symptom)+(0.11×NSE)+(0.16 x P-LCR)+(0.27 xPDW).e is the natural logarithm.4.Calculate the malignancy prediction rate of each patient.Use the pathological result as the gold standard to draw the ROC curve.Calculate that the area under the ROC curve is 0.72(0.66-0.78),and select the appropriate critical value of 0.81 as benign and malignant pulmonary nodules.The standard value of the judgment,the model predicts sensitivity of 74%,specificity of 62%,positive predictive value of 71%,negative predictive value of 69%.The result is acceptable.Conclusions:According to multivariate analysis,the patient's age,nodule diameter,burr sign,NSE,P-LCR,and PDW were independent risk factors for malignant SPN.In addition,the prediction model established by Logistic regression analysis,it is helpful to predict the good and bad performance of SPN to reduce the error of human factors,and it has good clinical diagnostic efficiency. |
PDF Full Text Request