Background: Role of antiangiogenic therapy is still undefined in ABC. We examined BV combined with G + C, a standard chemotherapy for ABC when the study began. Methods: Pts with ABC (inoperable stage III or IV), ECOG PS 0-1, normal organ and marrow function. Schema: BV 15 mg/ kg q 21 days; C 650 mg/m2 bid x 14 days, both starting day 1; G 1000 mg/m2 days 1 & 8; cycles repeated q 21 days. Primary objective was progression free survival (PFS), secondary objectives were overall survival (OS) toxicity, quality of life (QOL) using the FACT HEP tool and circulating tumor cell (CTC) number. Results: Fifty pts enrolled at 2 sites; 11(22%) gall bladder (GB), 29(58%) intrahepatic cholangiocarcinoma (IHC) and 10(20%) extrahepatic cholangiocarcinoma (EHC). Median age 63 (range 25- 84 years), 52% male, inoperable 16% and metastatic 84%. Cycles: median 8 cycles/patient (range 1- 33), median time on treatment was 5.8 months. Responses (RECIST): PR 12(24%), SD 24 (48%), PD 6(12%) and 8 (16%) 2 too early/6 clinical progression (nonevaluable). Clinical benefit rate (PR + SD) = 72 %. Median PFS: 8.1months (95% CI:5.3, 9.9). Median OS: 11.3 months (95% CI:8.1, 13.1). One year survival rate was 0.47 (95%CI: 0.32-0.60). Grade 3/4 toxicities ( > 5%): hematologic 28(58%), GI 18 (36%), infections 14 (28%) and thromboembolic events 6 (12%). Treatment discontinuation for toxicity (10%), progression (64%), death (4%), or other reasons (22%). Median OS in 21 (46%) pts with detectable CTCs was 9.4 months compared to 13.7 months in the 25(54%) pts without detectable CTCs at baseline.Patients with QOL scores after one and two cycles of therapy that were above median score at baseline for the whole group had higher OS of 13 months compared to 11.3 months. Conclusions: PFS and OS in ABC with the combination of gemcitabine, capecitabine and bevacizumab is relatively comparable to that reported with gemcitabine and cisplatin with acceptable toxicity. Lack of detectable CTCs and higher than median QOL score after one/two cycles of therapy may serve as biomarkers to predict improved outcome. Acknowledgement: Study is supported by a grant from Genentech. Clinical trial information: NCT01007552