Impact of ROS1, ALK, and/or MET expression level on the therapeutic efficacy of GEMOX with and without cetuximab in ABTC: A post hoc analysis of a randomized phase II trial.

Abstract

267 Background: Our randomized phase II trial showed adding cetuximab to GEMOX marginally improved the ORR and median PFS in ABTC patients regardless their KRAS mutation status. Recently, MET overexpression was noted in 11.7% to 16.2% of BTC, and associated with poor OS. ROS1 kinase fusion has also been detected in 8.6% of IHCC.The aim of the study is to evaluate the impact of ROS1, ALK and/or MET (RAM) overexpression on the clinical outcomes of ABTC patients. Methods: FFPE tissue sections that were prospectively collected for biomarker study from all 122 patients were subjected for determining the expression of ROS1, ALK and MET by IHC in a central laboratory. Results: Of 110 patients with enough tissue section for IHC of all three markers, 18 tumors were found to over-express ROS1 (in 9), ALK (in 5) and/or MET (in 6). One tumor over-expressed all three biomarkers. As compared with RAMlow tumors, the 18 RAMhigh tumors were significantly more of IHCC and without prior surgery, and with more frequent concurrent KRAS mutation. RAMhigh tumors was associated with significant inferior median OS than RAMlow tumors, 5.7 vs 11.8 months (p=0.04). Of the latter 92 patients, adding cetuximab improved the therapeutic efficacy of GEMOX in ABTC, with ORR of 31.8% vs 10.4% (p=0.02), DCR of 61.4% vs 37.5% (p=0.04) and median PFS of 7.0 vs 4.5 months (p=0.02). OS was only marginally affected, with median OS of 12.5 vs 10.4 months (p=0.36). The ORR in KRASwt and KRASmut patients receiving C-GEMOX and GEMOX was 34.5% vs 11.8% (p=0.04) and 26.7% vs 7.1% (p=0.33), respectively; whiles the DCR was 65.5% vs 50.0% (p=0.31) and 53.3% vs 7.1% (p=0.01), respectively, and the median PFS was 7.1 vs 5.8 months (p=0.07) and 7.0 vs 1.9 months (p=0.05), respectively. The results suggest that, the presence of KRAS mutation did not preclude the benefit of adding cetuximab to GEMOX in RAMlowsubpopulation. Conclusions: ROS1/ALK/METhigh ABTC had poor survival after C-GEMOX/GEMOX; while C-GEMOX significantly improved the ORR, DCR and PFS as compared to GEMOX alone in patients with ROS1/ALK/METlow tumors regardless their KRAS mutation status. Biomarker-driven design is warranted for future ABTC trials.