Construction and validation of hemoptysis risk prediction model for CT-guided lung biopsy patients
Objective To investigate incidence and main risk factors of hemoptysis in patients with pulmonary nodules undergoing routine CT-guided transthoracic needle biopsy(CT-TNB)procedures,and establish risk prediction model for guiding clinical practice.Methods From April 2020 to April 2023,a total of 306 patients with pulmonary nodules screened by conventional CT were enrolled,which included 163 males and 143 females,aged 51-72 years old with mean age of 63.8 years old;diameter of nodules was 1-4 cm with mean diameter of 2.7 cm;lesions located on pulmo sinister in 1 12 cases and pulmo dexter in 197;28 cases located on pulmonary hilar,125 on pulmonary uperior segment,42 pulmonary middle segment and 61 on pulmonary inferior segment;systolic blood pressure was 117-136 mmHg with mean systolic blood pressure of 128.9 mmHg;diastolic blood pressure was 68-89 mmHg with mean diastolic blood pressure of 79.0 mmHg;heart rate was 67-96 beats/minute with mean heart rate was 82.1 beats/minute;blood oxygen saturation was 95％-100％with mean oxygen saturation of 97.9％;chronic obstructive pulmonary disease(COPD)in 65 cases;prothrombin time(PT)was 10.6-14.8 seconds with mean PT was 12.4 seconds;D-dimer was 0.2-0.7 mg/L with mean value of 0.5 mg/L;platelet count was(187.3-297.4)× 109/L with mean platelet count of 239.9 × 109/L;C-reactive protein(CRP)was 3.9-9.3 mg/L with mean CRP of 6.5 mg/L;fasting blood glucose was 3.2-10.1 mmol/L with mean of 5.7 mmol/L;total cholesterol was 3.5-7.7 mmol/L with mean of 5.5 mmol/L;squamous cell carcinoma(SCC)antigen was 0.51-0.92 µg/L with mean of 0.68 μg/L;carcinoembryonic antigen(CEA)was 2.4-3.8 ng/L with mean of 2.9 ng/L.CT-TNB was used for differential diagnosis of benign and malignant.According to incidence of hemoptysis after operation,all of the patients were divided into hemoptysis group and non-hemoptysis group according to their postoperative situation.The 306 patients were randomly divided into training set(n=204)and validation set(n=102)at ratio of 2∶1.The demographic data,blood biochemical indexes,nodule characteristics and operative information of hemoptysis group and non-hemoptysis group were compared.The risk factors were screened by LASSO regression and multivariate Logistic regression model,and nomogram model was established.Results There were 65 patients(31.9％,65/204)with hemoptysis in training set and 29(28.4％,29/102)in validation set;and there was no significant difference in incidence of hemoptysis between 2 groups(P＞0.05).There was no significant difference in clinical data between hemoptysis group and non-hemoptysis group(P＞0.05),while there was significant difference in distribution of the maximum diameter of pulmonary nodules between 2 groups(x2=11.101,P=0.004).The univariate analysis results showed that the puncture distance of hemoptysis group was significantly longer than that of non-hemoptysis group,the number of ground glass nodules,benign nodules and nodules with diameter＜3 cm increased,and puncture time was prolonged(P＜0.05).The regression analysis results showed that puncture distance[odds ratio(OR)=1.768,95％confidence interval(CI)=1.325-2.124,P＜0.001]and puncture time(OR=1.235,5％CI=1.002-1.524,P=0.024)was risk factor for hemoptysis after CT-TNB,while solid nodules(OR=0.336,95％CI=0.115-0.857,P=0.009)and cavity nodules(OR=0.105,95％CI=0.023-0.428,P=0.033),malignant nodules(OR=0.502,95％CI=0.312-0.866,P=0.011),and the diameter of nodules≥3 cm(OR=0.857,95％CI=0.524-0.968,P＜0.001)were protective factors for hemoptysis.The R-software was used to establish nomogram based on the correspondence weights of each factor,with total score of 180 scores.The receiver operating characteristic(ROC)curve analysis results showed that the area under curve(AUC)of nomogram for predicting hemoptysis in training set and validation set was 0.903 and 0.867,respectively(P＜0.001),suggesting that the prediction accuracy of nomogram was better.The calibration curve and decision curve also showed that the predicted consistency and clinical net benefit of nomogram were better.Conclusion It is demonstrated that there is certain risk of hemoptysis in the identification of benign and malignant of pulmonary nodules using CT-TNB,and the developed nomogram model based on main risk factors has important application potential for guiding clinical accurate screening with high-risk patients of hemoptysis.
NETL
NSTL
万方数据
