Based On The Analysis Of Correlation Factors And The Construction Of Prediction Model Of Positive Pathological Results Of Pulmonary Solid Nodules PTNB After Operation

Objective:To investigate the prognostic factors of postoperative pathology of Subsolid Nodule(SSN)Percutaneous Thoracic Needle Biopsy(PTNB)and construct an effective prediction model.Methods:The general and imaging data of patients with suspected malignant SSN who underwent PTNB surgery in our hospital from December 2016 to May 2022 were retrospectively analyzed.General data included age,gender,smoking history,and tumor history.Imaging data included maximum diameter of the lesion,maximum diameter of the solid component,average CT value,distance from the lesion to the pleura,nodule location,burr sign,vascular cluster sign,vacuole sign,lobular sign,air bronchogram,pleural stretch sign,and presence of emphysema on the affected side.According to the postoperative pathological results of PTNB,malignant tumors were defined as pathologically positive.Benign neoplasms,glandular precursors,tuberculosis,normal bronchial and lung tissue,necrotic tissue,and institutional pneumonia were defined as pathologically negative.First,potential predictive factors can be screened out using the least absolute shrinkage and selection operator(Lasso)method.Then,screening factors can be verified by random forest method.A Nomogram model was established to predict the positive pathological results.Receiver operating characteristic curve(ROC),calibration curve and clinical decision analysis curve(DCA)were used to evaluate the model.Results:A total of 318 patients with SSN were included in this study,and 210 patients(66.04%)were pathologically positive after PTNB operation.Pathological negative group 108 cases(33.96%).LASSO regression analysis showed that age ≥68.5 years old,tumor history,maximum diameter of solid component ≥15.85 mm,burr sign and vascular cluster sign were predictive factors of positive postoperative pathology in patients with SSN,and the corresponding regression coefficients were 1.039,1.593,1.599,1.917 and 1.054,respectively.Burr sign(OR:6.780;[95%CI] 3.571-13.392;p < 0.001)were the best predictors of efficacy,and the OR values of age ≥68.5 years old,tumor history,maximum diameter of practicing component ≥15.85 mm,and vascular cluster sign were 2.827([95%CI] 1.442-5.792;p=0.003),4.920([95%CI] 2.102-12.526;p< 0.001),4.948([95%CI] 2.744-9.212;p< 0.001),2.868([95%CI] 1.605-5.209;p< 0.001).The AUC of a Nomogram model with five variables included was 0.836([95%CI] 0.77-0.884).The calculated concordance index(C-index)and corrected C-index were 0.836 and 0.826,respectively.The Bootstrap internal verification and the drawing of calibration curve and DCA curve also showed that the model had good prediction efficiency,accuracy and high clinical application value.Conclusion:Postoperative pathological results of PTNB were highly correlated with age,tumor history,maximum diameter of solid component,burr sign,and vascular cluster sign.Based on these factors,a Nomogram model of early lung malignant tumor was established to provide another reference for improving the accuracy of diagnosis of early lung malignant tumor.