Abstract
727 Background: Pancreatic cancer is the third leading cause of cancer mortality in the U.S. and remains among the deadliest solid tumors, with an estimated 67,440 new cases and 51,980 deaths projected in 2025. Pancreatic cancer carries a poor prognosis despite advances in multimodality care. Standard treatment typically centers on surgery when feasible and modern systemic therapy (e.g., FOLFIRINOX or gemcitabine-based regimens), with radiation used in borderline-resectable and unresectable disease as external beam radiation therapy (EBRT; conventional or SBRT) for local control and symptom relief. In contrast, pancreatic brachytherapy is rarely utilized in the U.S. and remains understudied at scale, with limited comparative-effectiveness data versus EBRT. Methods: The National Cancer Database (NCDB) was queried for adults (≥18 years) with pancreatic cancer diagnosed from 2004–2020 who received either EBRT or brachytherapy. Overall survival (OS) was estimated using Kaplan–Meier and compared by log-rank. Multivariable Cox proportional hazards models adjusted for NCDB analytic stage group, age, sex, race, Charlson–Deyo comorbidity score, year of diagnosis, and insurance status. Results: Among 83,808 patients treated with radiation, 385 (0.5%) received brachytherapy and 83,423 received EBRT. Median OS was 18.0 months (95% CI, 14.7–21.4) with brachytherapy versus 15.5 months (95% CI, 15.4–15.7) with EBRT (log-rank χ²=18.7; p<0.001). In multivariable analysis, brachytherapy was associated with reduced mortality compared with EBRT (aHR 0.75; 95% CI, 0.67–0.84; p<0.001). Independent predictors of worse OS included higher stage (HR 1.10 per category), older age (HR 1.02 per year), and greater comorbidity (HR 1.08 per point). Survival also improved with more recent diagnosis year (HR 0.97 per later year). Conclusions: In this large NCDB cohort, brachytherapy—though infrequently utilized—was associated with improved OS compared with EBRT, even after adjustment for clinical and sociodemographic factors. Given the observational design and NCDB limitations (e.g., lack of data on performance status, radiation dose/technique, sequencing with systemic therapy, or local control), selection bias cannot be excluded. These findings support further prospective trials and propensity-matched/center-level analyses to validate a potential survival advantage and identify patients most likely to benefit. Survival outcomes for patients with pancreatic cancer treated with brachytherapy vs EBRT (NCDB analysis 2004-2020). Treatment N Median OS, months (95% CI) Unadjusted log-rank p Adjusted HR (95% CI)* p-value Brachytherapy 385 18.0 (14.7–21.4) <0.001 0.75 (0.67–0.84) <0.001 EBRT 83,423 15.5 (15.4–15.7) <0.001 1.0 (Reference) — *Adjusted for stage, age, sex, race, Charlson-Deyo comorbidity score, year of diagnosis, and insurance status.