The effects of adjuvant radiotherapy on survival outcomes in polymorphous adenocarcinoma: a retrospective cohort study.

Background: Polymorphous adenocarcinoma (PAC) is the second most common malignant salivary tumor with a 10-year disease-specific survival (DSS) of 94%. While PAC is generally non-aggressive and surgery is often curative, some show aggressive features warranting adjuvant radiation therapy (RT). However, the impact of RT on survival outcomes remains unclear. The aim of this study is to investigate survival outcomes between patients treated with primary surgery Methods: The cohort consisted of patients listed in the National Cancer Database (NCDB) who were positive for PAC of all head and neck subsites and underwent surgical resection, with or without RT, between 2004 and 2020. Patients were then divided into subsets based on margin status. Patients lacking survival, follow-up, or staging data and those with Charlson-Deyo scores >0 were excluded to limit the confounding effects of co-morbid conditions and overall health status. Patient follow-up was performed in accordance with the treating institution. Overall survival (OS) was used as the primary endpoint. Cohort-level OS was adjusted for the following variables: age, sex, facility type, race, insurance payor, tumor stage, nodal stage, margin status, grade, salivary gland type, RT, and chemotherapy. Descriptive statistics and Cox regressions were performed. Covariates adjusted for included the following: sex, facility type, race, insurance payor, tumor stage, nodal stage, margin status, grade, salivary gland type, RT, and chemotherapy. All statistical analyses were performed using R Studio. Results: A total of 1,292 patients were included [32% male; 72% White; mean, 60 years; 92% minor gland primary and 8% major gland (parotid, submandibular, and overlap); 55% T1, 25% T2, 6% T3, 11% T4a, and 3% T4b; 94% N0 and 6% N+]. Multivariable Cox analysis showed a significant difference in OS for patients undergoing RT [hazard ratio (HR) =0.61, 95% confidence interval (CI): 0.39-0.96]. Positive margins were associated with worse OS (HR =1.69, 95% CI: 1.22-2.35) compared to those with negative margins. On subgroup analysis, patients with positive margins who received RT had significantly longer survival (HR =0.24, 95% CI: 0.10-0.56). Conclusions: While OS for PAC is high, 21.7% of patients had positive margins and thus a higher risk for local, regional, and distant recurrence. These patients benefit from RT and demonstrate significantly longer OS; thus, RT should be considered for cases with positive margins. Notable limitations in these conclusions include the retrospective design, use of NCDB data, which does not report specific follow-up data, and lack of recurrence data.