Rationale for increasing oncological vigilance in relation to clinical findings in accessory parotid gland- observations based on 2192 cases of the Polish Salivary Network Database

The accessory parotid gland (APG, Vth level) differs in histological structure from main parotid tissue. This gives rise to the hypothesis, mirrored in clinical observations, that the representation of tumours is different than in the rest of the gland. Aim To analyse the epidemiological and histological differences of parotid tumours located in regions I-V, with particular emphasis on the distinctiveness of region V. To define the epidemiological factors that will indicate the risk of histological malignancy from clinically benign appearance. Material and Method Multicentre prospective study conducted between 2017-2021 by five Head and Neck Surgery University Departments, cooperating within the Polish Salivary Network Database 1929 patients (1048 women and 881 men) were included. The age, gender, patient occupation, place of inhabitation, tumour size, clinical features of malignancy, histology, facial nerve (FN) paresis were analysed for superficial lobe (I_II), deep lobe (III_IV) and with special regard to the tumours affecting region V. Results 28 tumours were located exclusively in region V (1.45% total) and 72 tumours in region V with extension to neighbouring regions (3,7% total), characterized by significantly younger age and less frequent in retirees. In I-IV regions approximately 90% of tumours were benign, pleomorphic adenoma (PA) and Whartin tumour (WT) predominance. In region V PA exceed 75% but TW were casuistic (2/28). Incidence of malignancies in region V was 40% but clinical signs of malignancy were evident only in tumours>4 cm or in the presence of FN paresis. In 19% of patients with a benign appearance, imaging revealed malignancy, however in 38% of patients, the result remained false negative both in terms of clinical and radiological features of malignancy. Logistic regression models in 28 patients with tumour located exclusively in region V vs 1901 other patients and the in 100 patients with V extension vs 1829 other patients showed no clinical symptoms of malignancy bind with final malignant tumor histology as a single variable or in combination with other variables. The logistic regression models obtained in this study shows strong linkage between tumor location and predictors (age, male gender and tumor diameter) and aimed to act as a good classifier. Conclusion Despite the very clear image of the mid-cheek tumour, easily accessible in palpation and ultrasound examination, it is necessary to improve the oncological vigilance and preoperative patient preparation.