Trimodality Treatment Of Very Locally Advanced Sinonasal Cancer: A National Cancer Database Analysis

Sinonasal cancer is an uncommon head and neck cancer often presenting with invasion of local neighboring structures. For very locally advanced (T4b) tumors, national guidelines include only chemotherapy and RT as part of the standard of care. However, surgical management as part of a multimodal approach to sinonasal cancer results in a survival benefit in small retrospective series. We performed a large database analysis using the National Cancer Database (NCDB) to examine patterns of care and outcomes in T4b sinonasal cancer. Patients with sinonasal squamous cell carcinoma treated between 2004 and 2016, with pT4b and cT4b non-metastatic disease, were included in the analysis. Patients were stratified between those who underwent trimodality therapy including surgical resection, chemotherapy and radiation (trimodality cohort) and those who received chemotherapy and RT without surgical resection (chemoradiotherapy cohort). Untreated patients, and patients treated with single modality therapy (i.e., surgery, RT or chemotherapy alone) or surgery with only chemotherapy or RT were excluded. Kaplan-Meier methods and univariate/multivariate Cox regression analysis were performed on the full cohorts and propensity-matched cohorts, with age, year of diagnosis, sex, race, Charlson-Deyo comorbidity score, facility type, site, and nodal status included in analysis. Among 615 patients included in analysis, 238 (38.7%) were in the trimodality cohort, while 377 (61.3%) were in the chemoradiotherapy cohort. Median follow-up time was 20.2 months. Baseline characteristics were generally well-balanced between the two cohorts, with trimodality therapy used significantly more in patients diagnosed after 2010 (p = 0.006). Kaplan-Meier estimate of overall survival at 2 years for patients in the trimodality cohort was 61% (95% CI: 55-68%), while for patients in the chemoradiotherapy cohort was 46% (95% CI: 41-52%). On univariate and multivariate analysis, a significant survival difference was associated with the inclusion of surgery in management (univariate: HR 0.681, p < 0.001, multivariate: HR 0.728, p = 0.005). The survival advantage in the surgical cohort remained consistent on propensity score matched analysis (238 patients in each cohort; HR 0.726, p = 0.008). Patients who underwent trimodality treatment for very locally advanced sinonasal cancer (stage T4b) experienced longer survival. This data suggests that there might be a role for incorporating surgery in treatment of advanced stage disease patients, even when curative intent surgical extirpation with microscopic negative margins is not likely. These findings should be further investigated in clinical trial settings.