2191 - Outcomes Following Total Neoadjuvant Therapy with Radiation in Borderline Resectable Pancreatic Cancer
Presenter(s)
H. J. Roberts1, P. S. Pathak2, Z. Guan3, B. Y. Yeap3, C. R. Blaszkowsky1, J. L. Koenig1, A. E. Silberstein4, C. Weekes2, H. Singh2, J. N. Allen2, D. P. Ryan2, L. S. Blaszkowsky2, A. R. Parikh2, M. L. Peters2, C. Fernandez-del Casti5, M. Qadan6, P. J. Fagenholz5, J. Harrison5, K. Lillemoe5, T. S. Hong7, and J. Y. Wo1; 1Department of Radiation Oncology, Massachusetts General Hospital, Harvard Medical School, Boston, MA, 2Division of Medical Oncology, Department of Medicine, Massachusetts General Hospital, Harvard Medical School, Boston, MA, 3Department of Biostatistics, Massachusetts General Hospital and Harvard Medical School, Boston, MA, 4Harvard Radiation Oncology Program, Boston, MA, 5Department of Surgical Oncology, Mass General Brigham, Boston, MA, 6Department of Surgery, Massachusetts General Hospital, Harvard Medical School, Boston, MA, 7Dana-Farber Cancer Institute, Boston, MA
Purpose/Objective(s): The role of radiation (RT) in pancreatic cancer remains controversial, and there is no standard approach in the borderline setting.
Materials/Methods:
This retrospective study included patients with borderline resectable pancreatic cancer (BRPC) by NCCN criteria treated with total neoadjuvant therapy (TNT) with chemotherapy followed by RT at a single academic center (2016–2025). Restaging scans occurred at 2 and 4 months after chemotherapy initiation and before RT. OS and PFS were estimated from RT start using Kaplan–Meier. Multivariable (MV) logistic regression identified factors associated with resection, and MV Cox regression identified factors associated with overall survival (OS) and progression-free survival (PFS) (no MV analysis for local control (LC) due to few events). Resection covariates included age (continuous), post-NT CA19-9 response (normalization or =50% reduction with absolute value <100 mg/dL vs not), and ECOG PS (0 vs >1). OS/PFS models included age, post-NT CA19-9 response, and total RT dose (BED10 >/ =90 Gy of external + intraoperative RT (IORT)), analyzed separately for resected and unresected patients with pN+ included for resected patients.Results: Among 162 patients with BRPC, median follow-up for 61 survivors was 16.3 mo (1.2–104.6). Chemotherapy was primarily FOLFIRINOX (90%) and/or gemcitabine/nab-paclitaxel (18%) delivered over a median of 4 months. Most received 50.4–58.8 Gy in 28 fractions (69%); others received 40 Gy in 5 fractions (14%), 30 Gy in 10 fractions (10%), or alternative regimens (8%). Concurrent chemotherapy was given in 90% (capecitabine 75%, n=121). Surgical exploration occurred in 88% (n=143), with 65% resected (n=106) (88% R0 (n=93)); 30% (N=32) were pN+. IORT was delivered to 72 (50%) patients (median 10 Gy, 8–15), including 23 (16%) unresected. 19 patients were not explored (progression = 12, comorbidity = 2, other/unknown = 5). Median OS for the full, resected, and unresected cohorts was 21, 32.9, and 11 mo; median PFS was 9.5, 15.6, and 3.3 mo. 2-year OS, PFS, and LC rates were 47%, 22%, and 69%. On MV analysis, CA19-9 response was associated with higher resection rates (OR 3.16, p=0.003) and improved OS in resected (HR 0.52, p=0.057) and unresected patients (HR 0.41, p=0.014). BED10 > 90 Gy was associated with improved PFS in unresected patients (HR 0.34, p=0.003).
Conclusion:
TNT including RT for BRPC was associated with high rates of surgical conversion, R0 resection, and favorable survival in resected patients. Post-treatment CA19-9 strongly predicted resection and OS, supporting its prognostic value. Among unresected patients, higher RT dose achieved with IORT (BED10 >90 Gy) was associated with improved PFS, supporting the potential benefit of dose-escalated RT among unresectable patients. These results support ongoing prospective studies on dose escalation (NRG LAP100).