Construction of a diagnostic model for smear-negative active pulmonary tuberculosis based on multi-slice CT imaging features

LI Xin; ZHU Xiaogang; LUO Minxia

doi:10.12122/j.issn.1674-4500.2024.01.15

Abstract：Objective To analyze the multi-slice CT(MSCT)features of patients with smear-negative active pulmonary tuberculosis and construct a diagnostic model.Methods A total of 1016 patients with pulmonary diseases who were admitted to our hospital were enrolled from January 2020 to January 2023,and they were divided into smear-negative pulmonary tuberculosis group(n=478)and non-tuberculous pulmonary disease group(n=538,including 200 cases of lung cancer and 338 cases of pneumonia)according to clinical diagnosis.All patients received MSCT examination,and their MSCT imaging features were analyzed.Logistic regression analysis was conducted to identify the signs related to the diagnosis of smear-negative active pulmonary tuberculosis,and a diagnostic model for smear-negative active pulmonary tuberculosis was constructed.The diagnostic efficiency of the model was evaluated with AUC.Results There were statistically significant differences in MSCT signs such as tree in bud,centrilobular nodule,cavity,calcification and lobulation between the two groups(P<0.05).Logistic regression analysis found that tree in bud sign,centrilobular nodule and cavity were independent risk factors for smear-negative active pulmonary tuberculosis(P<0.05).The equation of the combined diagnosis model constructed based on logistic regression analysis was as follow:Log(P)=-1.256+1.455×tree in bud sign+0.982×centrilobular nodule+1.023×cavity.The AUC,sensitivity and specificity of this model were 0.825,93.94%and 70.97%,respectively.Conclusion The diagnostic model constructed based on MSCT imaging features is of high value in diagnosis of smear-negative active pulmonary tuberculosis and can provide a reliable basis for clinical diagnosis and treatment of the disease.