The Value Of Artificial Intelligence And PET/CT In Differential Diagnosis Of Benign And Malignant Pulmonary Nodules

Objective:The purpose of this study is to explore the consistency of AI 、 PET/CT and pathological results in differential diagnosis of benign and malignant pulmonary nodules,analyze the differences of the diagnostic value between AI and PET/CT,and evaluate the clinical value of different diagnostic methods.Methods:64 patients who were diagnosed as pulmonary nodules in the multidisciplinary consultation center of Northern Jiangsu People’s Hospital and treated surgically from December 2018 to December 2020 were selected as the research objects.All patients accepted PET/ CT and chest CT examination before operation,and the chest CT images of them were analyzed by AI.Analyzing the consistency of AI 、 PET/CT and pathological results,and compared with the results of HRCT imaging features and MDT diagnosis,using receiver operating characteristic(ROC)curve to analyze the diagnostic value,sensitivity,specificity,accuracy,AUC,etc.of different methods used alone or in combination.IBM SPSS 24 statistical software was used for data analysis.Mean ±standard deviation table was used for measurement data,frequency or percentage was used for count data,and chi square test was used for comparison between groups.(P <0.05)the difference was statistically significant.Results:1.Analyze the consequence of the diagnosis results of artificial intelligence and postoperative pathological results,the kappa value was 0.571.Its diagnostic value:sensitivity: 91.4%,specificity: 75.0%,accuracy: 89.4%,positive predictive value: 96.4%,negative predictive value: 54.5%,AUC:0.832,P<0.05,there was statistical significance.2.Analyze the consequence of the diagnosis results of PET/CT and postoperative pathological results,the kappa value was 0.404.Its diagnostic value: sensitivity: 82.8%,specificity: 75.0%,accuracy: 81.8%,positive predictive value: 96.0%,negative predictive value: 37.5%,AUC:0.789,P< 0.05,there was statistical significance.3.Analyze the consequence of the diagnosis results of imaging features of HRCT and postoperative pathological results,the kappa value was 0.415.Its diagnostic value:sensitivity: 87.9%,specificity: 62.5%,accuracy: 84.4%,positive predictive value: 94.4%,negative predictive value: 41.7%,AUC:0.752,P< 0.05,there was statistical significance.4.Analyze the consequence of the diagnosis results of MDT consultation results and postoperative pathological results,the kappa value was 0.798.Its diagnostic value:sensitivity: 96.6%,specificity: 87.5%,accuracy: 95.5%,positive predictive value: 98.2%,negative predictive value: 77.8%,AUC:0.920,P<0.05,there was statistical significance.5.Analyze the consequence of the diagnosis results of artificial intelligence combined with PET/CT and postoperative pathological results,the kappa value was 0.615.Its diagnostic value: sensitivity: 93.1%,specificity: 75.0%,accuracy: 90.9%,positive predictive value: 96.4%,negative predictive value: 50%,AUC:0.841,P<0.05,there was statistical significance.6.Analyze the consequence of the diagnosis results of PET/CT combined with imaging features of HRCT and postoperative pathological results,the kappa value was 0.486.Its diagnostic value: sensitivity: 91.4%,specificity: 62.5%,accuracy: 87.9%,positive predictive value: 94.6%,negative predictive value: 50%,AUC:0.769,P<0.05,there was statistical significance.7.Analyze the consequence of the diagnosis results of combination of artificial intelligence,PET/CT and imaging features of HRCT and postoperative pathological results,the kappa value was 0.681.Its diagnostic value: sensitivity: 98.3%,specificity:62.5%,accuracy: 93.9%,positive predictive value: 95.0%,negative predictive value:83.3%,AUC:0.804,P< 0.05,there was statistical significance.Conclusion:1.The AI analysis of chest CT images has a good consistency with postoperative pathological results in the diagnosis of pulmonary nodules,and has a higher value in the than PET/CT and imaging features.However,compared with MDT,MDT is more valuable.2.The results of PET/CT in the diagnosis of pulmonary nodules are consistent with postoperative pathology,and compared with imaging features,both have advantages in the diagnosis of malignant pulmonary nodules.But compared with AI and MDT,AI and MDT have better diagnostic value.3.MDT diagnosis was the most effective diagnostic method,followed by AI combined with PET/CT and AI and AI combined with PET/CT combined with imaging features,followed by PET/CT,PET/CT combined with imaging features and imaging features used alone.The highest diagnostic sensitivity was AI combined with PET/CT combined with imaging features,followed by MDT diagnosis,AI combined with PET/CT,followed by AI,PET/CT combined with imaging features and imaging features,PET/CT used alone.The highest diagnostic specificity was MDT diagnosis,followed by AI combined with PET/CT and AI,PET/CT alone,followed by AI combined with PET/CT combined with imaging features,imaging features combined with PET/CT and imaging features alone.