Objective To explore the value of binary logistic regression analysis in the differential diagnosis of focal organizing pneumonia(FOP)and peripheral lung cancer(PLC)using multi-slice spiral CT(MSCT).Methods 53 patients with focal organizing pneumonia confirmed by pathology were selected as the study group,and 61 patients with peripheral lung cancer were selected as the control group.The CT signs of the two groups were analyzed and compared,and a regression model was constructed by using statistically significant features.The area under the curve,sensitivity,and specificity were calculated by receiver operating characteristic(ROC)curves.Results Binary multivariate logistic regression analysis showed that unclear boundaries,long thorn,short thorn,and air bronchogram were independent predictive factors for distinguishing FOP from PLC;The area under the ROC curve of the model is 0.945,with a sensitivity of 79.20%and a specificity of 98.40%.Conclusion Unclear boundaries,long thorn,short thorn,and air bronchogram signs are helpful for the differential diagnosis of FOP and PLC.The predictive model can improve the diagnostic efficiency and provide important auxiliary information for clinical practice.
维普
万方数据