Outcomes of Primary Versus Salvage Surgery for Sinonasal Malignancies: A Population-Based Analysis

There exists a lack of consensus on the optimal sequence of treatment for many sinonasal malignancies (SNMs). This study compares the overall survival (OS) outcomes for primary surgery (PS) versus salvage surgery (SS) in SNM patients across stage, histology, and primary site.

Retrospective database review.

The National Cancer Database was queried for all SNM cases treated with multimodal surgical and nonsurgical therapy between 2004 and 2015. Logistic regression identified predictors of SS. Cox proportional hazards models evaluated predictors of mortality, and Kaplan-Meier log-rank test assessed OS outcomes.

Our SNM cohort consisted of 3,011 patients (PS = 2,804; SS = 207). SS patients had significantly longer postoperative hospital stays (P = .009) and increased rates of 30-day (P < .001) and 90-day mortality (P < .001) compared to PS. On multivariate logistic regression, predictors of undergoing SS included sinonasal undifferentiated carcinoma histology (odds ratio = 2.72; 95% confidence interval [CI]: 1.16-6.66; P = .024). On multivariate Cox proportional hazards analyses among SS patients, late-stage disease (hazard ratio [HR] = 4.80; 95% CI: 1.46-15.8; P = .01) and positive surgical margins (HR = 2.31; 95% CI: 1.29-4.13; P = .005) portended significantly worse OS. In the propensity score-matched cohort controlling for stage and histology, PS had significantly improved OS compared to SS (P = .007). Compared to SS, PS also had improved OS in subgroup analyses for patients with late-stage disease (P = .026) and squamous cell carcinoma histology (P = .006).

In our SMN cohort, PS resulted in improved OS outcomes compared to SS independent of stage and histology. Consideration may be given to primary surgical resection for SMN whenever feasible, though a targeted, individualized approach is warranted.

3 Laryngoscope, 131:E710-E718, 2021.