Objectives
The objectives of our study were to review our experience in management of regional lymph nodes, in patients with carcinoma of the parotid gland, identify clinico-pathological factors predictive of neck metastases, identify neck levels pathologically positive for metastases following neck dissection and report neck recurrence rates.

Materials and Methods:
This was a single institution retrospective cohort study. We identified 266 patients with previously untreated carcinomas of the parotid gland between the years 1985- 2009. Three patients were M1 at presentation and were excluded from analysis, leaving 263 patients for the study. Patient, treatment and tumor characteristics were collected by retrospective review of patient charts. Patients were stratified by neck management into 3 groups: Observation (Obs), elective neck dissection (END) and therapeutic neck dissection (TND). The pathological positivity of each neck level was quantified for the END and TND groups. Clinico-pathological characteristics of the END group and TND group vs Obs group were compared using the Chi square test of association. Neck recurrence free survival was determined for each group using Kaplan Meier statistics.

Results:
There were 136 males and 127 females (median age 62 years). Of the 263 patients, 232 were cN0 and 31 cN+. Of the cN0 patients, 158 were selected to have neck observation and 74 END. All cN+ patients had TND. Of the END group, occult neck metastases were detected in 26 (35%) patients. The % positivity per neck level was 19.2% level I (5/26), 84.6% level II (22/26), 61.5% level III (16/26), 26.9% level IV (7/26), 15.4% level V (4/26). Of the TND group, pathological positivity was found in 86.1% patients. The % positivity per neck level was 59.3% level I (16/27), 85.2% level II (23/27), 85.2 % level III (23/27), 59.3% level IV (16/27), 44.4% level V (12/27). Compared to the observation group, the END and TND groups were more likely to be over 65 yrs of age and have clinical stage T3/4 disease. Pathology showed the END and TND groups had more aggressive histology with a greater percentage high grade, vascular invasion, perineural invasion, positive margins and pT stage. The majority of patients who were pN+ in the END group (93%) and TND group (92%) had postoperative radiation (PORT). Figure 1 outlines the management of the neck in our patient groups. The Obs group and END group (pN0) had an excellent 5yr NRFS of 98.7% and 97.3% respectively. Patients who had pathologically positive neck on END or TND had a NRFS of 88.6%.

Conclusion
Patients who are cN0 who present with clinical stage T3/4 disease or high grade histology have a high rate of occult metastases involving neck levels I-IV. Patients who are pN+ managed by END or TND followed by adjuvant PORT achieve a satisfactory NRFS of 88.6%.