Application Value of Thoracic LDCT Artificial Intelligence Combined with Physician Diagnosis in Identify of Benign and Malignant Pulmonary Nodules
Objective:To explore the application value of thoracic low dose CT(LDCT)artificial intelligence combined with physician diagnosis in identify of benign and malignant pulmonary nodules.Method:A total of 225 patients with pulmonary nodules admitted to the Huizhou Third People's Hospital from January 2022 to September 2023 were selected.Thoracic LDCT scan was performed in all patients.The benign and malignant pulmonary nodules detection condition of gold standard,thoracic LDCT physician diagnosis,thoracic LDCT artificial intelligence combined with physician diagnosis was analyzed.The diagnostic value of thoracic LDCT physicians diagnosis and thoracic LDCT artificial intelligence combined with physicians diagnosis was compared.The consistency of thoracic LDCT physician diagnosis,thoracic LDCT artificial intelligence combined physician diagnosis and gold standard was analyzed.The thoracic LDCT imaging features of benign and malignant pulmonary nodules were compared.The malignant diagnosis accuracy of different components of pulmonary nodules thoracic LDCT artificial intelligence combined with physicians diagnose was analyzed.Result:Among 225 patients,139 malignant cases detected by gold standard.There were 134 malignant cases detected by LDCT physicians diagnosis,accounting for 96.40%(134/139).Thoracic LDCT artificial intelligence combined with physicians diagnosis detected 138 malignant cases,accounting for 99.28%(138/139).The Kappa test showed that the Kappa values of the thoracic LDCT physician diagnosis,thoracic LDCT artificial intelligence combined physician diagnosis and gold standard were 0.860 and 0.972 respectively(P<0.001).The sensitivity,accuracy and negative predictive value of thoracic LDCT artificial intelligence combined with physicians diagnosis were higher than those of thoracic LDCT physicians diagnosis,and the differences were statistically significant(P<0.05).The proportions of short burr sign,pleural depression sign,vascular cluster sign,lobular sign and vacuole sign in patients with malignant pulmonary nodules were higher than those in patients with benign pulmonary nodules,and the differences were statistically significant(P<0.05).Among the 225 patients,117 cases were solid nodules and 54 cases were malignant nodules.There were 60 cases of partial solid nodules and 49 cases of malignant nodules.There were 48 cases of pure glass nodules and 36 cases of malignant nodules.The accuracy rate of malignant in solid nodules by thoracic LDCT artificial intelligence combined with physicians diagnosis was 98.15%(53/54),the malignant accuracy of partial solid nodules was 100.00%(49/49),the accuracy of malignant diagnosis of pure ground glass nodules was 97.22%(35/36).Conclusion:Thoracic LDCT artificial intelligence combined with physicians diagnosis can improve the value of differential diagnosis of benign and malignant pulmonary nodules,which can contribute to the early formulation of accurate treatment plan and improve patient prognosis.
万方数据
维普
