Risk Factor Analysis Of The Patients With Solitary Pulmonary Nodule Less Than 20mm In Diameter And Establishment Of A Prediction Model For The Probability Of Malignancy

Background:Solitary pulmonary nodule(SPN)is a frequently encountered problem in clinical practice.With the progress of imaging technique,especially the promotion of chest CT,the detection rate has significantly increased,and it is important to estimate the probability of malignancy in patients with SPN accurately for taking timely clinical intervention.In order to reduce the misdiagnosis,missed diagnosis and excessive intervention of empirical medicine,it is desirable to develop noninvasive and accurate diagnostic methods or tools of SPN.Therefore,many medical centers have analyzed the independent risk factors of malignancy in patients with SPN,and established a variety of mathematical prediction models to assist in clinical diagnosis and treatment of SPN.It is worth mentioning that a number of studies have shown that the malignant probability of SPN increases with increasing nodule diameter.The probability of malignancy is higher when the diameter of nodule is>20mm,and these nodules are easier to diagnose.Therefore,the identification between benign and malignant in nodules which are<20mm is more challenging,and has more important clinical significance on the early diagnosis and treatment.Objective:According to the clinical and imaging data of collected patients with≤20mm SPN,we analyzed independent risk factors of malignancy in patients with≤20mm SPN by univariate analysis and multivariate logistic regression analysis.Then,establish a mathematical prediction model for the probability of malignancy.Methods:A retrospective cohort study of 362 patients with<20mm SPN who underwent surgical resection with definite pathological diagnosis from June 2015 to June 2016 in Sir Run Run Shaw Hospital.The clinical and imaging data included age,gender,symptoms,history and quantity of smoking,previous history of tumor,family history of tumor,nodule location,maximum diameter,spiculation,lobulation,pleural indentation sign,border,nodule type,vacuole sign,calcification,vascular convergence.Independent risk factors were screened with multivariate logistic regression analysis,then a mathematical prediction model was built.Results:There were statistically significant differences in gender,diameter(>10mm),nodular type(ground-glass opacity),spiculation,clacification,vacuole sign and vascular convergence(all P<0.05)by the univariate analysis between benign and malignant SPN(≤20mm).Multivariate logistic regression analysis showed that gender,diameter(>10mm),nodular type(ground-glass opacity),speculation,and vacuole sign were independent predictors of malignancy in patients with<20mm SPN(P<0.05).The mathematical prediction model to estimate the probability of malignancy was:Logit(P)=ex/(1+ex),X=-1.472+(0.959×gender)+(1.002×diameter)+(1.890×speculation)+(2.879×g round-glass opacity)+(1.605×vacuole sign).And e was natural logarithm.The area under the ROC curve for our model was 0.841(95%CI:0.781-0.900).When the cut-off value was 0.801,the sensitivity was 89.4%,specificity was 69.5%,positive predictive value was 93.4%,and negative predictive value was 55.6%.Conclusion:The univariate and multivariate analysis show that gender,diameter(>10mm),nodule type(ground-glass opacity),spiculation and vacuole sign are independent predictors of malignancy in patients with<20mm SPN.The established mathematical prediction model is accurate and can be used to assist in clinical diagnosis.