Objectives: To characterize patients receiving TORS for oropharyngeal tumors nationwide, compare the positive margin rate (PMR) among subsites, pathologic tumor stages, and facility types, and identify predictors of positive margins.
Methods: Retrospective review of the National Cancer Database (2010-2014). χ2 test was used to compare PMR between subgroups. Patient, tumor, and treatment factors were associated with positive margins via univariable and multivariable logistic regression All factors with p<0.2 for association with positive margins and all clinically-relevant factors were included in the multivariable model. T0 patients were excluded.
Results: 2,778 patients undergoing TORS were identified. Mean age was 59.5 years, 82.3% of patients were male, and 80.5% were treated at academic centers. 48.6% of tumors were HPV+ (34.1% unknown), 57.2% had a tonsil primary site, 39.3% base of tongue (BOT), 1.9% soft palate, and 1.5% other oropharynx. 86.9% of patients were T1/T2 and 10.3% were T3/T4. Overall PMR was 16.8%. PMR was significantly different in patients with higher T stage (pT1=13.4%, pT2=17.2%, pT3=28.3%, pT4A=42.1%, pT4B=58.3%; p<0.001). PMR was also higher in patients with BOT than tonsil primary site (19.4% vs 15.7%). Academic centers had a 15.7% PMR, as opposed to 25.2% for non-academic (p<0.001). Of demographic variables examined, age, insurance, and facility type were associated with margin status in univariable (p<0.2); the rest (sex, race, income, distance to facility) were excluded from the multivariable. On multivariable regression, high T stage was associated with increased positive margin rates (versus T1, T2: OR=1.53, CI=1.06 – 2.22; T3: 3.22 [1.88-5.52]; T4A: 5.25 [2.16-12.78]; T4B: 14.49 [2.50-83.92], p<0.001). Treatment at an academic facility was associated with decreased odds (0.58 (0.39-0.85), p<0.01). Demographics, extracapsular extension, lymph-vascular invasion, HPV status, subsite, and number of positive nodes were not associated with positive margin rates.
Conclusions: TORS is typically performed at academic centers for tonsil and BOT-based tumors. The overall PMR is 16.8%. 10% of patients undergoing TORS have T3 or T4 disease, and, when controlling for demographic and clinical factors, higher T stage was strongly associated with increased risk of positive margins, with T3 and T4 tumors having the highest odds. Furthermore, non-academic centers carry twice the odds of positive margins as compared to academic. Further research is needed to determine the efficacy of TORS in tumors with high T stages and to understand reasons for the disparity between academic and non-academic centers utilizing TORS.