Survival Impact of Transoral Robotic Surgery in Early-Stage Oropharyngeal Cancer

Abstract

Transoral robotic surgery (TORS) has multiple theoretical advantages over conventional surgery for early-stage oropharyngeal squamous cell carcinoma (OPSCC). Since FDA-approval in December 2009, TORS has become a standard of care treatment for early-stage OPSCC, but there is limited data comparing TORS to conventional surgery with respect to peri-operative mortality, surgical outcomes, and long-term oncologic outcomes. To this end, we examined the patterns of uptake of TORS and its impact on surgical outcomes and survival using real-world data from a large national cancer registry. Patients with T1-2 OPSCC undergoing definitive surgery from 2010 to 2015 were identified in the National Cancer Database. Univariate and multivariate Cox regression models were constructed to evaluate the association between TORS and survival, adjusting for patient- and disease-related covariates. Survival with robotic and non-robotic surgery was also compared in three unrelated disease sites where both approaches are widely employed: Gleason 7-10 prostate cancer, stage I endometrial cancer, and stage IA2 or IB1 cervical cancer, to assess for systemic selection bias regarding surgical approach in national registry data. Of the 9,745 patients who met inclusion criteria, 2,694 (27.6%) underwent TORS. There was a significant increase in the utilization of TORS, from 18.3% to 35.5% of all surgeries for T1-2 OPSCC from 2010 to 2015 (P = 0.003). TORS was associated with similar 90-day peri-operative mortality as conventional surgery (1.4% vs. 1.0%, P = 0.11) and length of hospital stay (4.28 days vs 4.42 days, P = 0.49). In terms of surgical outcomes, TORS was associated with lower rates of positive surgical margins (12.5% vs 20.3%, P<0.001), and accordingly lower utilization of adjuvant chemoradiation (28.6% vs. 35.7%, P<0.001). In the subset of 4,071 patients with known HPV-status, TORS was associated with improved overall survival (OS) compared to non-robotic surgery in multivariable Cox regression (hazard ratio [HR], 0.747; 95% CI, 0.612-0.913, P = 0.004), even when excluding facilities that did not offer TORS. 5-year OS was 84.8% vs. 80.3% among patients undergoing TORS versus non-robotic surgery in propensity score-matched cohorts (P<0.001). By contrast, robotic surgery was not associated with improved OS in other cancers, such as prostate (HR 0.893; 95% CI 0.772-1.032; P = 0.13), endometrial (HR 0.948; 95% CI 0.879-1.022; P = 0.16), and cervical cancer (HR 1.282; 95% CI 0.943-1.743, P = 0.11). The utilization of TORS has rapidly increased during the first 6 years of adoption. TORS was associated with higher rates of margin negative surgery and improved OS compared to non-robotic surgery in early-stage OPSCC. The association between improved OS and robotic surgery was unique to early-stage OPSCC, and was not observed in prostate, endometrial or cervical cancer. It is unclear if these results are due to technical differences, surgical acumen, or other factors.