Background: National Comprehensive Cancer Network (NCCN) guidelines for treatment of early-stage glottic cancer is limited to either surgery or radiation alone. However, post-operative radiation (PORT) continues to be administered to a significant subset of patients for unclear indications. The primary objective of this study was to identify factors associated with the use of PORT.
Methods: A retrospective analysis of the National Cancer Database (NCDB) was performed for primary cT1-T2N0M0 glottic squamous cell carcinoma (SCC) patients who underwent primary surgery during the years 2004-2014. Multivariate logistic regression was performed to identify independent predictors of PORT.
Results: 7109 patients were identified. The majority of tumors were cT1 (N = 5913, 84.2%) compared to cT2 (N = 1106, 15.8%). Patients were surgically treated with either local excision (N = 6370, 90.1%) or partial laryngectomy (N = 649, 9.2%). 2852 (40.6%) patients received surgery alone while 4167 (59.4%) underwent surgery and PORT. Patients between these two cohorts did not significantly vary by age, race, or comorbidity status (p > 0.05). Margins were negative in 67.8% (1270 of 1874) of PORT patients compared to 89.4% in surgery alone (p < 0.001). In multivariable regression, cT2 tumors (adjusted odds ratio [aOR], 2.27; [1.8-2.87 95% confidence interval]), positive margins (aOR 3.56 [2.92-4.33]), treatment at an Academic/Research institution (aOR 1.34 [1.10-1.64]) and more aggressive initial surgery (aOR 0.15 [1.11-0.20]) were independent predictors of PORT. 53 of 1030 hospitals analyzed across the US (5.1%) comprised the top-volume quartile, of which 76.6% were Academic/Research institutions. The top-volume quartile comprised 39.3% of those treated with surgery alone, compared to 15.1% of those receiving PORT (p < 0.001). Compared to the 1st quartile, treatment at 2nd quartile (aOR 3.23), 3rd quartile (aOR 5.70), and lowest quartile (aOR 5.40) institutions had significantly higher odds of receiving PORT in this model. Patient age, comorbidity, and insurance status were not predictive of PORT. In a multivariable Cox proportional-hazards model, receipt of PORT was not predictive of survival (p = 0.41). Five-year overall survival (OS) for those receiving surgery was 78% versus 77% for surgery with PORT (p = 0.168).
Conclusion: A majority of patients with early-stage glottic SCC continue to receive PORT after primary surgery despite tumor-free margins and lack of survival benefit. The highest odds of receiving PORT in multivariable analysis included treatment at lower-volume facilities, positive margins, and cT2 disease. This study highlights a critical need to re-evaluate the use of PORT in patients with early-stage glottic SCC.