DOI: 10.36347/sjams.2018.v06i03.041

Pages: 1021-1027

The Parotid tumors represent less than 3% of all cervical tumors. They are benign in 80% of cases. The objective of this study is to evaluate the contribution of 3 imaging methods: ultrasound, CT and MRI in the differentiation between malignant and benign tumor of the parotid. This is a retrospective study of 49 cases of parotid tumors collected over 3 years from 2014 to 2016. The parameters studied were age, sex, reason for consultation, data from clinical examination, imaging, surgery and histological data. Each of the radiological criteria used to determine the nature of the tumor was analyzed and correlated with the histology data. For the analytical study, the chi-two test was used and the p-value was calculated (significant if p <0.05) and then the sensitivity, specificity and diagnostic efficiency were calculated for each modality. The average age of our patients was 44 years. The sex ratio is a 0.58 with a clear female predominance. The pattern of consultation was parotid swelling in all cases. The right side with 30 cases and the left with 19 cases, with cervical adenopathy in 4 cases. The average time for consultation was 59 months. The definitive histology was benign in 39 patients and malignant in 10 cases. On ultrasound, fuzzy boundaries and irregular contours predict malignancy. At CT, the criteria that lead to malignancy are: blurred boundary, area of necrosis, and extends to neighboring tissues. At MRI, irregular contours, capsular effraction, extension to neighboring tissues and presence of adenopathy are the criteria that lead to malignancy. Radiation-histological correlation showed a sensitivity of 66%, 42% and 88%, and a specificity of 90%, 82% and 82% for ultrasound, CT and MRI, respectively. Preoperative knowledge of the nature of the tumor guides the surgeon in his surgical decision. Therefore, imaging has become a real diagnostic tool for surgeons, and MRI represents the exam of choice, especially with the emergence of new dynamic techniques.