Cancers of the nasal cavities and paranasal sinuses represent approximately 3% of all head and neck cancers.1 Their outcome remains poor, as only 40%–60% of patients are alive 5 years after first diagnosis.2-4 The aim of our study was to describe the incidence of neck metastasis at the time of diagnosis of sinonasal cancers, as well as risk factors for nodal involvement. We conducted a retrospective study of all patients with tumors arising from nasal cavity and/or paranasal sinuses treated at a single tertiary referral center from 2010 to 2020. Prophylactic neck treatment consisted of elective nodal irradiation (ENI) or dissection (END). The choice between ENI and END mostly depended on the type of treatment to the primary tumor and its histology. Collected data included demographics, comorbidities, smoking status, occupational exposure, imaging workup, cancer characteristics, treatment type and follow-up data. Data were summarized by frequency and percentage for categorical variables and by median and range for continuous variables. Associations between nodal involvement and clinicopathological characteristics were assessed using the chi-squared or Fisher's exact test for categorical variables and the Mann-Whitney test for continuous variables. A multivariable logistic regression model was performed to study factors associated with nodal involvement. Survival rates were estimated using the Kaplan-Meier method. Univariable and multivariable analyses were performed using the log rank test and the Cox model. Statistical analyses were conducted using STATA v16 software. Of the 415 screened patients, 248 patients met inclusion criteria. Excluded patients were as follows: six aged under 18 years at diagnosis, six without follow-up data, 53 diagnosed before 2010, 29 benign affections, 12 hematological malignancies, and 61 contiguous primitive tumor or sinonasal metastasis. Patient and tumor characteristics are shown in (Table 1). Among all pooled histological types, we found an overall incidence of 12.5% of initial nodal involvement (n = 31/248) that is, 30 cN+ and 1 cN0/pN+ (occult nodal disease). The prevalence of nodal involvement at diagnosis for each histology was: 36.4% for adenocarcinoma (ACC), 27.8% for sinonasal undifferentiated carcinoma (SNUC), 15.8% for mucosal melanoma, 12.2% for squamous cell carcinoma (SCC), 8.3% for esthesioneuroblastoma (ENB) and 6.1% for adenocarcinoma. No significant association was found between histological type and presence of neck metastasis. Among all subsites analyzed, only the orbital extension showed a statistical trend for a higher risk of neck metastasis (50.0% vs. 33.8%, p = 0.084) on univariate analysis. No variable was independently associated with nodal involvement in multivariable analysis, taking into account the following variables: patient's age, histology, clinical T staging, and presence of orbit invasion. Among the nine patients with pathologically-confirmed neck metastases, level Ib and IIa were the most commonly involved (16.1% and 19.4% respectively), followed by level V (9.7%), IIb (6.5%), III (6.5%), IV (3.2%), and Ia (0%). In total, 31.5% (n = 78/248) of patients developed disease recurrence which was initially local (70.5%), regional (25.6%), and/or distant (39.7%). Among 20 patients who presented with neck recurrence, three were initially cN+ or pN+ and eight patients had received initial neck treatment (neck dissection and/or radiotherapy). Median follow-up was 29.3 months (95% confidence interval [CI], 24.3–36.9). The 2-year overall survival (OS) and event-free survival (EFS) were 84.1% (95% CI, 77.9–88.7) and 61.3% (95% CI, 53.9–67.8), respectively. Among patients who had no neck metastasis at diagnosis, the presence of a prophylactic neck treatment was not associated with an improvement of OS (p = 0.485), or EFS (p = 0.518). Nodal involvement was significantly associated with a shortened OS (hazard ratio [HR] 2.35; 95% CI, 1.18–4.66; p = 0.012), and so were midline crossing tumors and T3-T4 stages compared to T1-T2 (HR 2.87; 95% CI, 1.54–5.35; p < 0.001 and HR = 5.95; 95% CI, 1.84–19.26; p < 0.001). Only nodal involvement and T3-T4 stages were associated with a shortened EFS (HR 1.85; 95% CI, 1.13–3.02; p = 0.013 and HR 2.65; 95% CI, 1.50–4.68; p < 0.001). We conducted a multivariable analysis of EFS and OS, which identified as independent prognostic factors: nodal involvement, T3-T4 stages and midline crossing tumors (Table 2). Neck treatment showed a statistical trend for improved OS. Incidence of cervical lymph node metastasis in sinonasal cancers at diagnosis varies across studies with a range of 3%–33%, according to T-stage, locations, and histologies.3, 5, 6 SCC have previously been associated with a high incidence of nodal involvement (up to 33%), but did not show a significantly higher incidence of primary neck involvement (p = 0.904) in our study.3, 5 Advanced T-stages (T3-T4) have previously been associated with a higher risk of nodal involvement at diagnosis.2 In our N+ population, 83.9% of patients had advanced T-stages at diagnosis, but the rate was not significantly different to that of the N0 population, probably due to the heterogeneity of histologies included. Other risk factors for nodal involvement have been described (invasion of the dura, infratemporal fossa, orbit, hard palate, or rhinopharynx).3 In our cohort, only the local extension to the orbital structures seemed associated with nodal involvement. Some authors reported the presence of neck metastasis at presentation as associated with poor prognosis and indicating tumor extension beyond the sinonasal cavity,4, 7, 8 yet local recurrence continues to be the main cause of treatment failure.1 Similarly, retropharyngeal lymph node involvement at diagnosis has also been described as a significant prognostic factor for decreased OS and locoregional control.9 Of note, we did not study the presence of retropharyngeal nodes in our cohort. Other studies about sinonasal malignancies reported levels I and II as most common sites of neck metastases regardless of histological type.3, 5, 10 Our results are consistent with these findings. When required, elective nodal treatment should include levels Ib, II, III, and retropharyngeal nodes, regardless of histological type.9, 11 In this study, the rate of nodal involvement varied according to a number of demographic, pathological, and anatomical factors, however without reaching statistical significance. Nodal disease at presentation was significantly associated with worse EFS and OS. Our findings support previous considerations regarding the need for an optimal initial staging, and a prophylactic cervical treatment only in selected cases that still need to be determined.