Questionnaire
Elective Neck Dissection in T1N0 sinonasal SCCs
In total, 17 (89.5%) and 18 (94.7%) European head and neck cancer centers would not perform END in T1N0 SCCs of the paranasal sinuses or nasal cavity, respectively. Answers and recommendations did not significantly differ among centers from different countries (p=0.964). Only two German centers (10.5%) would perform END, comprising stage I-IV in one center and only level II in another (Figure 1A).
Elective Neck Dissection in T2N0 sinonasal SCCs
Similarly, END would not be performed by 13 (68.4%) and 14 (73.7%) respondents for T2N0 SCCs of the paranasal sinuses or nasal cavity, respectively. Again, approaches did not significantly differ among countries (p=0.792). However, END would be performed in 2 Austrian and 4 German centres (31.6%). Those centers would do particularly level I (5/19; 26.3%) and level II (5/19; 26.3%) ND followed by level III (3/19; 15.8%) and IV (1/19; 5.3%), respectively (Figure 1B).
Applied imaging tools to decide on the necessity of Elective Neck Dissection
Ultrasound, CT-scan, MRI or PET-CT/MRI are performed by 10 (52.6%), 10 (52.6%), 8 (42.1%), and 6 (31.6%) centers to obtain additional information regarding lymph node status. Six centers (31.6%) applied one imaging technique (ultrasound: n=1; CT-scan: n=3; PET-CT/MRI: n=2), while the majority of centers combined two (n=10; 52.6%) or three different techniques (n=3; 15.8%). The number of applied radiological studies did not significantly differ between institutes who perform END (2.0 ± 0.0) compared to those who did not (1.79 ± 0.8; p=0.336).
Correlation and Analysis of Clinical Data
Recurrence and Risk Score
Local, regional and distant recurrences were noticed in 7 (12.1%), 10 (17.2%) and 1 (1.7%) patient, respectively. The mean and median time between diagnosis and recurrence was 50.9 and 43.0 months. Regional recurrence was significantly more often observed in tumors originating from nasal vestibule (20.0%) and nasal septum (22.7%; p=0.013). In addition, we noticed a trend towards local failure in T1 tumors and towards regional failure in T2 tumors (p=0.116; Table 1). Although difference failed to reach statistical significance, none of the 9 electively neck-dissected patients (15.5%) experienced recurrence (p=0.188). Adjuvant radiotherapy was applied in 11 patients and significantly more often in patients with incomplete tumor resection (80% vs. 13.5%; p=0.004) and more likely in T2 than T1 tumors (28.6% vs 13.5%; p=0.181).
Based on our data, we supposed that T-classification (T2 > T1) and tumor site (septum and vestibule > other origin) seem to represent risk factors for occurrence of regional and/or distant recurrence. Therefore, we created a risk score based on T-classification (T1 tumor = 0 points (=pts.); T2 tumor = 1 pts.) and tumor site (nasal septum or vestibule = 1 pts.; other origin = 0 pts.) to better estimate on whether or not patients may benefit from END. After risk stratification of all patients according to our proposed risk score, we had 9 patients with low risk (0 pts.), 35 patients with moderate risk (1 pt.) and 14 patients with high risk (2 pts.) for regional recurrence (Figure 2).
Nomogram
Based on our data, we further calculated a nomogram to predict the risk of regional recurrence in all patients based on T-Classification and anatomic subsite (Figure 3). Anatomic origin (0=other subsite; 1= nasal septum or vestibule) was quantified with a higher impact on regional recurrence than T-classification (1=T1; 2=T2). As exemplified, a T1 SCCs originating from the nasal septum or vestibule had a 13-14% risk for regional recurrence (Figure 3).
Outcome Analyses
Subsequently, we performed log-rank test and univariate cox-regression analyses to assess the effect of different clinical variables and our risk-scoring system on freedom from regional and distant recurrence (FFR) and cancer specific survival (CSS).
The 5-year FFR was 49.9% in T2 tumors compared to 88.1% in T1 tumors, which was significantly different (p=0.016). Moreover, elder patients and those who have not undergone END showed a worse but not significantly different FFR (p=0.052; p=0.086). Applying our risk score, patients with high-risk score, indicating T2 tumors originating from nasal septum or vestibule, showed significantly worse FFR (p<0.001). Indeed, even at univariate cox-regression analysis, high-risk scores showed a 9.52-fold higher risk for regional and or distant recurrence (HR 9.52; p=0.002; Table 3).
Similarly, the 5-year CSS was 96.4% in T1 tumors compared to 58.3% in T2 sinonasal tumors (p=0.009). Noteworthy, 5-year CSS was almost halved in patients with high-risk scores (44.4% vs. 92.8%) compared to patients with moderate or low risk (p=0.017). Occurrence of regional recurrence was also associated with significantly worse CSS (40.0% vs. 95.7%; p=0.005). High-risk scores (HR 6.62; p=0.039) and regional recurrences (HR 8.13; p=0.016) represented significantly worse prognostic factors for CSS (Table 4).