Addition of stereotactic body radiotherapy (SBRT) to systemic chemotherapy in locally advanced cholangiocarcinoma (CC) (ABC-07): Results from a randomized phase II trial.

4006 Background: Locoregional treatments for inoperable, non-metastatic CC remain undefined. Single arm phase 2 data suggested that SBRT may improve clinical outcomes. We aimed to investigate the efficacy and safety of the addition SBRT with CisGem in locally advanced inoperable CC. Methods: ABC-07 (ISRCTN:10639376) is a phase II multicentre randomized trial for patients with inoperable, histologically confirmed locally advanced CC, WHO performance status 0-1. After registration patients showing no progression after 4 cycles of CisGem were randomized (2:1) for treatment after completing six cycles: either SBRT (50 Gy in 5 fractions or 67.5 Gy in 15 fractions, based on tumour size) or two additional cycles (cycles 7 and 8) of CisGem (CG-only). Primary endpoint was progression free survival (PFS). With 80% power and a 15% one-sided alpha, the study planned to randomize 65 patients to detect a median PFS increase from 10.4 to 17.4 months. Secondary endpoints included overall survival (OS), toxicity and patterns of failure. RT quality assurance was undertaken. Results: Between March 2016 and August 2022, 16 UK centres randomized 69 patients (SBRT, 45; CG only, 24). Median age was 66 (38 to 83) yrs, 58 (84.1%) perihilar location, 48(69.6%) stent in situ, median tumor size was 3.5cm (range 0.6-10.5). The median follow-up time was 20.7 months. In the SBRT arm, 43 completed 6 cycles CisGem (96%) and 41 (91%) received SBRT. In the CG-only arm, 18 (75%) completed 8 cycles. The median PFS from randomization was 8.6 and 9.0 months in the SBRT and CG-only arms, respectively (HR: 1.00; 95% CI: 0.58,1.70; p = 0.989). The first PFS events were 7 (16%) local and 24 (53%) metastatic relapses for SBRT, versus 7 (29%) and 7 (29%) in the CG-only arm, respectively. From registration, median OS time was 23.4 (95% CI: 14.6, 27.7) months for SBRT and 17.2 (95% CI: 10.2, NR) months for CG-only arm. Post-randomization, OS was 19.4 months (11.2, 24.6) for SBRT and 14.2 (7.0, NR) months for CG-only (HR: 0.79; 95% CI: 0.41,1.51; p = 0.47). Adverse events of grade ≥3 were observed in 33 (73%) and 21 (88%) patients in SBRT and CG-only arms, respectively. Of the infections, of which biliary sepsis was common, pre-cycle 6, SBRT had 19 (42%) cases vs. 7 (29%) in CG-only; post-cycle 6, SBRT had 17 (38%) sepsis cases vs. 6 (25%) in CG-only. There was 1 Grade 3 duodenal hemorrhage and 1 death due to sepsis in the SBRT arm. Primary causes of death included disease (20, 44% vs 11, 46%), sepsis (5, 11% vs 3, 13%), and hepatic failure (0 vs 3, 13%) for the SBRT and CG-only arms, respectively. Conclusions: SBRT did not show a PFS advantage over CG-only, yet a longer median OS time and better primary tumour control were observed without safety concerns. However, more mature survival data is still needed. Clinical trial information: 10639376.