ABSTRACT
Objective To screen the clinical factors affecting the malignant probability of solitary pulmonary nodule (SPN) with univariate and multivariate Logistic regression analysis,and to establish a clinical prediction model,evaluate its test effectiveness in the differential diagnosis in SPN.Methods A retrospective cohort study included 182 patients with diagnosis of SPN (group A).Clinical data included gender,age,smoking history,quitting smoking,history of tumor,serum carcinoembryonic antigen (CEA),location,diameter,density,vacuole sign,cavity,airbronchogram,calcifcation,lobulation,spiculate sign,spiculation,pleural indentation sign,vascular convergence sign,enhanced CT value,the maximum standard uptake value (SUV~x) of positron-emission tomography (PET),pathological diagnosis were collected.The independent predictors of malignancy were estimated with univariate and multivariate analysis,then the clinical prediction model to identify malignant was established.Other 45 SPN patients (group B) were used to tested value of the model.Results Univariate analysis showed that gender,age,no lower lobe location,diameter,density,airbronchogram,calcification,lobulation,spiculation,pleural indentation sign,enhanced CT value ≥ 15 HU,SUVm,≥ 2.5 of PET significantly affected judgment of SPN of benign or maligant (P ＜ 0.05).Multivariate analysis revealed that female,age,lobulation,short spiculation,pleural indentation sign was the independent predictors of malignancy in patient with SPN,solid nodule suggested benign.The clinical prediction model to identify malignant was established:P =ex/ (1 + ex),x =-3.399 +1.382 × gender + 0.056 × age + 1.377 × lobulation + 1.498 × spiculation-2.096 × solid + 1.005 × pleural indentation sign,e was natural logarithm.P =0.663 was as cut-off point,group B was used to test value of the model:the sensitivity was 86.5％,the specificity was 62.5％,the positive predictive value was 91.4％,the negative predictive value was 50.0％,the accuracy was 82.2％.Conclusions As for SPN,female,age,lobulation,short spiculation and pleural indentation sign is the independent predictor of malignancy in patient with SPN,solid nodule suggests benign.The prediction model is sufficient to estimate the malignancy of patient with SPN.