Background: No standard treatment options are available for patients with advanced biliary cancer (BC) who fail gemcitabine/platinum based regimen. The most commonly used second line regimens are 5-fluorouracil (5-FU) with leucovorin (LV), or capecitabine based on limited data. Trametinib is an orally bioavailable, potent, and specific allosteric inhibitor of MEK1/2. In cell lines, increased sensitivity to trametinib was observed especially in cells with mutations in the MAPK pathway. Aberrant activation of the Ras/Raf/MAPK pathway occurs in more than 60% of BC indicating the importance of these pathways. Furthermore other MEK inhibitors have been studied in advanced BC demonstrating promising results including complete responses (Bekkai-Saab et al JCO 2012, Finn RS GI ASCO 2011) Methods: : An open label phase II trial was designed to randomly assign patients to receive trametinib ( 2.0mg daily) versus 5-FU ( 400 mg/m² bolus and 2400 mg/m² infusion over 48 hours) with LV( 400 mg/m²) every 2 weeks or capecitabine (1000 mg/m² BID 14 days on 7 days off) in refractory BC. Eligible patients must have histologically or cytologically documented BC (excluding ampullary cancer) following progression of first line chemotherapy. The primary endpoint is overall survival. Secondary endpoints include progression-free survival and overall response rate. Additional endpoints will assess a 16 gene expression signature as a biomarker of MEK efficacy and to measure inflammatory cytokines. MRI or CT scan will also be obtained for all patients at the time of study entry and at 6 weeks into treatment to estimate lean soft tissue and fat mass weight gain according to treatment with trametinib. Analysis will compare randomized groups, stratified by 5FU/LV versus capecitabine and cholangiocarcinoma versus gall bladder. Assuming a one-sided type 1 error of 10%, 80% power, and 2 years of accrual with an additional year of follow-up, approximately 80 eligible patients (40 per arm) are needed to detect an improvement in median OS from 5 months to 8.25 months (corresponding to a 1.65 hazard ratio). As of Jan 2015, 33 of the planned 89 patients have been randomized. Clinical trial information: NCT 02042443.