Imaging Differential Diagnosis Of Soft Tissue Hemangioma And Soft Tissue Schwannoma

Objective: To evaluate the imaging differential diagnosis in hemangioma and schwannoma of soft tissue.Methods: The imaging findings of 50 cases of soft tissue hemangioma and 51 cases of soft tissue schwannoma confirmed by pathology in our hospital were analyzed retrospectively.Univariate analysis was carried out for the differences of imaging signs of lesions.Multivariate regression analysis was carried out for the indicators with statistical significance,and Logistic regression model was established.Hosmer-Lemeshow test was used to determine the goodness of fit of Logistic regression model,and receiver operating characteristic(ROC)curve was used to evaluate its diagnostic efficacy.Results: There were significant differences between the two groups in terms of maximum diameter(>5 cm),fuzzy boundary,irregular shape,non-marginal calcification,thylakoid zone,internal fat,target sign,nerve access sign,capsule and banded shadow(P< 0.05).The five features entered into the Logistic model by multivariate regression analysis were: irregular morphology(OR=25.842,95%CI: 1.999-334.152),thylakoid zone(OR=0.038,95%CI: 0.002-0.847),internal fat(OR=24.340,95%CI: 1.197-495.014),target sign(OR=0.009,95%CI: 0-0.764),and capsule(OR=0.006,95%CI:00.146).Hosmer-Lemeshow test had a significance level of 0.999.The area under ROC curve was 0.986(P < 0.01).The sensitivity,specificity,positive predictive value and negative predictive value were 94.0%,94.1% and 94.0% respectively.Conclusion: Irregular morphology,thylakoid zone,internal fat,target sign and capsule are the independent characteristic features to predict hemangioma or schwannoma of soft tissue.Logistic regression model had high sensitivity and specificity in the differentiation.