Abstract
BACKGROUND: There are a lack of studies about whether radiation dose escalation synergizes with immunotherapy and targeted therapy in pancreatic cancer. In this study, we performed a secondary analysis to investigate whether a high radiation dose rather than a low dose plus pembrolizumab and trametinib provided improved survival compared with gemcitabine in post-operative locally recurrent pancreatic cancer. METHODS: In this open-label, randomised, controlled, phase 2 trial, eligible patients with pancreatic ductal adenocarcinoma characterized by mutant KRAS and positive immunohistochemical staining of PD-L1 and documented post-operative local recurrence were randomly assigned using an interactive voice or web response system, without stratification, to receive stereotactic body radiation therapy (SBRT) with doses ranging from 35 to 40Gy in five fractions, pembrolizumab 200 mg every three weeks and oral trametinib 2 mg once daily (SBRT + K + M) or SBRT and gemcitabine (1000 mg/m(2)) on day 1 and 8 of each 21-day cycle (SBRT + G) until disease progression in our hospital in China. Those had radiotherapy, immunotherapy or targeted therapy were excluded. Patients and investigators were not masked to the assignment. In each arm, patients were stratified based on biologically effective dose (BED(10); α/β = 10) of 60-65Gy and BED(10) ≥65Gy. The primary endpoint was overall survival (OS) and the secondary endpoint was progression-free survival (PFS). All patients received their assigned treatment and were included in the efficacy and safety analyses. This study is registered with ClinicalTrials.gov, NCT02704156. FINDINGS: Between Oct 10, 2016, and Oct 28, 2017, 147 of 170 randomly assigned participants were eligible for inclusion in this analysis. In BED(10) of 60-65Gy group, 34 and 29 patients had SBRT + G and SBRT + K + M, respectively. While there were 42 and 42 patients with SBRT + G and SBRT + K + M in BED(10) ≥65Gy group. Patients in the SBRT + K + M group had longer OS compared with the SBRT + G group, but this did not reach statistical significance (median: 15.1 vs. 12.4 months, HR 0.67 [95%CI 0.43-1.04]; p = 0.071). For BED(10) of 60-65Gy, OS was similar between patients in the SBRT + K + M and SBRT + G groups (median, 13.6 vs. 12.4 months; HR 0.69 [95% CI 0.41-1.16]; p = 0.16). For BED(10) of ≥65Gy, PFS was prolonged with SBRT + K + M versus SBRT + G (median: 8.6 vs. 5.0 months, HR 0.48 [95% CI 0.31-0.77]; p = 0.0021). For BED(10) of 60-65Gy, there was no significant difference in PFS between the two groups (PFS: median, 7.9 vs. 4.3 months; HR 0.69 [95% CI 0.42-1.15]; p = 0.16). In BED(10) of 60-65Gy group, 7 (20.6%) and 8 patients (27.6%) with SBRT + G and SBRT + K + M had grade 3 or 4 adverse events (p = 0.52). In BED(10) ≥65Gy group, 8 (19.0%) and 12 patients (28.6%) with SBRT + G and SBRT + K + M had grade 3 or 4 adverse events (p = 0.31). No treatment-related death occurred. INTERPRETATION: Dose escalation of SBRT may improve PFS with pembrolizumab and trametnib versus gemcitabine for patients with post-operative locally recurrent pancreatic cancer. However, benefits of PFS did not translate into longer OS. This may be ascribed to small sample size and post-hoc analysis that was not powered to determine the significance. Therefore, synergy of high dose of SBRT with immunotherapy and targeted therapy required further investigations in phase 3 trials. FUNDING: Shanghai Shenkang Centre and Changhai Hospital.