Beschreibung:
<jats:p><jats:italic>Aim</jats:italic>. To create a predictive score for the discrimination between benign and malignant parotid tumors using elastographic parameters and to compare its sensitivity and specificity with standard ultrasound.<jats:italic> Methods</jats:italic>. A total of 124 patients with parotid gland lesions for whom surgery was planned were examined using conventional ultrasound, Doppler examination, and shear wave elastography. Results of the examinations were compared with those ones of histology.<jats:italic> Results</jats:italic>. There were 96 benign and 28 malignant lesions in our cohort. Blurred tumor margin alone proved to be an excellent predictor of malignancy with the sensitivity of 79% and specificity of 97%. Enlarged cervical lymph nodes, tumor vascularisation, microcalcifications presence, homogeneous echogenicity, and bilateral occurrence also discriminated between benign and malignant tumors. However, their inclusion in a predictive model did not improve its performance. Elastographic parameters (the stiffness maxima and minima ratio being the best) also exhibited significant differences between benign and malignant tumors, but again, their inclusion did not significantly improve the predictive power of the blurred margin classifier.<jats:italic> Conclusion</jats:italic>. Even though elastography satisfactorily distinguishes benign from malignant lesions on its own, it hardly provides any additional value in evaluation of biological character of parotid gland tumors when used as an adjunct to regular ultrasound examination.</jats:p>