Enhanced CT combined with histogram analysis for differential diagnosis of Warthin tumor and polymorphic adenoma of parotid gland / 中国医学影像技术

ABSTRACT

Objective: To investigate the value of enhanced CT combined with histogram analysis in differential diagnosis of Warthin tumor and pleomorphic adenoma(PA) of parotid. Methods: Maxillofacial plain and enhanced CT were performed on 37 patients with Warthin tumor (Warthin tumor group) and 28 patients with PA (PA group). CT values of plain, arterial phase (AP) and venous phase (VP) CT of lesions were measured, and the enhanced patterns and pure enhanced values were recorded. Then the parameters of histogram analysis, including the mean, standard deviation (StdDev), minimum (Min), maximum (Max), Median, skewness (Skew) and kurtosis (Kurt) were calculated. The differences of all parameters were compared between groups. ROC curves of parameters being statistically different between groups were drawn for differential diagnosis of Warthin tumor and PA of parotid gland, and the corresponding diagnostic efficacy were analyzed. Results: There were statistically differences of CT values in each phase, AP pure enhanced values and enhanced patterns between groups (all P<0.05), and MeanAP, MinAP, MaxAP, MedianAP, SkewAP, MeanVP, StdDevVP, MinVP, MedianVP and SkewVP also showed significant differences (all P<0.05). AUC and Youden indexes of enhanced pattern, AP pure enhanced value, MeanAP, MinAP, MaxAP, MedianAP and SkewAP were all higher than those of VP histogram parameters in differential diagnosis of PA and Warthin tumor. SkewAP had the highest Youden index (0.71), with AUC of 0.88. The diagnostic efficacy enhanced pattern+SkewAP was the most effective, with AUC, Youden index, sensitivity and specificity of 0.94, 0.80, 83.58% and 96.43%, respectively. Conclusion: Enhanced CT combined with histogram analysis based on AP images were helpful to differential diagnosis of Warthin tumor and polymorphic adenoma of parotid. The diagnosis performance of combined parameters was better than that of single parameter.