The Predictive Value Of MRI Multiparameter Quantitative Analysis For Malignant Transformation Of Sinonasal Inverted Papilloma

Objective:To analyze the clinical and imaging features of sinonasal inverted papilloma(SNIP)and sinonasal inverted papilloma-transformed squamous cell carcinoma(SNIP-SCC),and to evaluate the predictive value of each feature and related combination model for malignant transformation of SNIP.Methods:The clinical and imaging data of 83 patients with SNIP and 21 patients with SNIP-SCC were retrospectively analyzed,including age,gender,lesion volume,T2 WI signal intensity ratio(T2W SIR),apparent diffusion coefficient ratio(ADCr)and wash in index(WII).The parameters with statistical differences were obtained by univariate analysis.Multivariate Logistic regression was used to obtain the independent risk factors for predicting malignant transformation of SNIP.The receiver operating characteristic(ROC)curve was used to evaluate the ability of each risk factor and its combination model to predict the malignant transformation of SNIP.Results:Age,ADCr,WII2 and WII3 were independent risk factors for predicting malignant transformation of SNIP,and the area under the ROC curve(AUC)was0.767,0.862,0.727 and 0.704,respectively.The AUC of the comprehensive model of MRI parameters including ADCr,WII1 and WII2 was 0.910,the sensitivity was 76.19%,the specificity was 90.36%,and the accuracy was 87.50%.The AUC,sensitivity,specificity and accuracy of the comprehensive model of age,ADCr,WII2 and WII3 for predicting malignant transformation were 0.929,85.71%,92.77% and 91.35%,respectively.Conclusion:Compared with the comprehensive model of age,ADCr,WII and other independent risk factors and MRI parameters,the comprehensive model of quantitative parameters composed of age and MRI parameters can better predict the malignant transformation of SNIP,and provide important clinical and imaging evidence for accurate preoperative evaluation of the malignant transformation of SNIP.