Real-world multicentre analysis of patients with hepatocellular carcinoma (HCC) treated with atezolizumab and bevacizumab (A+B): Efficacy, esophagogastroduodenoscopy (EGD) uptake and bleeding complications.

4105 Background: The IMbrave150 trial established A+B as a standard of care for HCC and an EGD within 6 months (mos) of starting treatment to detect esophagogastric varices and prevent bleeding complications was advised. However, the value of performing EGD in all patients (pts) is unknown. Methods: We conducted a retrospective analysis of all HCC pts treated with A+B between July 2020 to August 2022 at 5 cancer centres from the Canadian provinces of Ontario, Alberta and Manitoba (members of the HCC-CHORD Consortium). Pts characteristics and treatment history including efficacy, EGD details and bleeding events were collected. Overall survival (OS), progression-free survival (PFS) and response rate (RR) were calculated. A comparative analysis was conducted to compare the correlation between the EGD uptake and bleeding events. Results: A total of 112 pts were identified (median age 66 years, 87% male, 24% East Asian, 67% liver cirrhosis, 33% HCV, 25% HBV, 15% NASH, 23% BCLC B, 71% BCLC C, 91% Child-Pugh A, 29% main portal vein invasion, 45% ALBI grade 1 and 54% ALBI grade 2). Prior to systemic, 61% had locoregional therapies. 90% received A+B as first-line therapy, while 9% received it as second-line and 1% as third-line. Outcomes of pts treated with A+B are shown. Before starting A+B, 78 pts (70%) had completed an EGD within 6 mos. Of these, 32 pts (41%) had evidence of varices on EGD, and 15 (20%) required treatment with either banding or beta-blockers. All bleeding events in this population was 15% (n=17). Bleeding rates in the EGD and non-EGD groups were 18% (n=14) and 9% (n=3), respectively. Bleeding adverse events were 5%(n=6) gastrointestinal vs 10%(n=11) non-gastrointestinal (6 epistaxis, 1 ecchymosis, 1 periodontal and 3 unspecified). The GI bleeding rates in the EGD and non-EGD groups were 6% (n=5) and 3% (n=1). Conclusions: Outcomes of HCC pts treated with A+B in Canada are comparable to those observed in the IMbrave150 trial. Our study detected and treated varices at twice the rate of the IMbrave150 trial among the EGD group, reflecting real-world treatment of a high-risk population. Yet, pts who had no EGD before starting A+B, presumably due to a low risk of portal hypertension based on physicians’ judgement, did not experience more bleeding/GI bleeding than those who had an EGD. This supports the approach of selective EGD prior to A+B, and more guidelines are needed in this area. Outcomes and response rate of patients(pts) treated with Atezolizumab with Bevacizumab in this study. [Table: see text]