摘要
Abstract
Objective:To explore the application value of multiple parameters provided by AI software in the differential diagnosis of three types of sub centimeter solid nodules(metastatic nodules,inflammatory nodules and pneumoconiosis nodules)in the lungs.Methods:Thin-section CT data were collected,and a total of 386 nodules belonging to three categories were screened according to the enrollment criteria.The following parameters provided by the AI system were recorded:maximum di-ameter(long axis),minimum diameter(short axis),maximum CT value,minimum CT value,average CT value,skewness,kurtosis,volume fraction(solid component ratio),and volume.The ratio of maximum diameter to minimum diameter was manually calcu-lated.Statistical differences were compared between the metastatic nodule group and the combined group of inflammatory and pneumoconiosis nodules.Significant influencing factors identified through univariate analysis were used to establish a Logistic regression model.ROC curves were plotted for variables found to be significant in the regression model to evaluate their clin-ical diagnostic value.Results:Statistically significant differences were observed in maximum length,minimum length,minimum CT value,solid component ratio,volume,and the ratio of maximum length to minimum length(F values:45.509,66.204,3.345,7.452,25.309,6.833;P values:＜0.001,＜0.001,0.036,＜0.001,＜0.001,0.001).Statistically significant differences were also found in maximum length,minimum length,volume,and the ratio of maximum length to minimum length among all nodules(P values:＜0.001 for all).When comparing metastatic nodules with pneumoconiosis nodules,statistically significant differences were observed in maximum length,minimum length,solid component ratio,and volume(P values:＜0.001 for all).However,no statistically significant differences were found when comparing metastatic nodules with inflammatory nodules or inflammatory nodules with pneumoconiosis nodules in any of the parameters.Logistic regression analysis suggested that minimum length was an independent influencing factor for the diagnosis of sub-centimeter metastatic nodules.Conclusion:The ability of skewness and kurtosis to determine whether a sub centimeter nodule is a metastatic nodule is limited.The diagnosis of pulmonary sub centimeter solid nodules can be improved by combining clinical features with a history of primary tumor.关键词
肺肿瘤/体层摄影术,X线计算机Key words
Lung Neoplasms/Tomography,X-ray Computed分类
医药卫生
