Real-world use of atezolizumab plus bevacizumab in patients with hepatocellular carcinoma and Child-Pugh A and B cirrhosis.

Abstract

393 Background: Atezolizumab plus bevacizumab (A+B) is the new standard of care for first-line treatment of advanced hepatocellular carcinoma (HCC). No evidence exists as to its use in routine clinical practice in patients (pts) with impaired liver function. Methods: This retrospective, multi-center observational study was conducted across 7 tertiary academic referral centres and collected 64 HCC pts consecutively treated with A+B. Efficacy outcome measures included overall (OS) and progression-free survival (PFS) calculated from time of A+B commencement and overall response rates (ORR) and disease control rates (DCR) defined by Response Evaluation Criteria in Solid Tumors (RECIST, v1.1). Safety outcomes included treatment-related adverse events (trAEs) graded (G) according to CTCAE v5.0. Results: Of 64 eligible pts, 54 had BCLC C stage HCC (84%), secondary to hepatitis C cirrhosis (n = 24; 37%), hepatitis B (n = 10; 16%), and non-viral etiologies (n = 40; 47%). Liver function was classified as Child-Pugh (CP) A in 46 patients (72%), B7 in 7 (11%), B8 in 8 (12%), and B9 in 3 (5%). Patients were of performance status (PS) ECOG 0 (n = 39; 61%) and 1 (n = 25; 39%). Pre-treatment upper-gastrointestinal endoscopy was performed in 44 patients (69%), with gastro-esophageal varices found in 18 pts (40%) and graded as 1 (n = 12, 27%), 2 (n = 4, 9%) and 3 (n = 2, 4%) respectively. After a median follow-up of 6.8 months (m) (95% confidence interval [CI], 5.5-8.0), median OS (mOS) was 11.7m (95% CI, 6.2-17.3) whereas median progression-free survival (mPFS) was 6.97m (95% CI, 2.9-11.0). ORR and DCR were 26% and 62% respectively. TrAEs of any grade were documented in 43 pts (67%): 12 pts (18%) had trAEs of G≥3: 7 (11%) atezolizumab-related and 5 (8%) bevacizumab-related. Toxicity led to treatment discontinuation in 3 pts (5%). Compared to CP-A, CP-B pts achieved shorter OS (11.7m [95% CI, 10.3-13.2] vs 6.5m (95% CI, 3.5-9.5), p = 0.029) and PFS (9.1m [95% CI, 5.4-12.8] vs 2.3m [95% CI, 1.7-2.9], p = 0.001) with no differences in ORR nor in DCR. The rate of trAEs did not significantly differ across CP classes. Median OS was significantly longer in patients achieving a radiologic response (12.7m [95% CI, not reached] vs 11.0m [95% CI, 5.5-16.5], p = 0.04). Presence and grade of varices was not associated to bevacizumab-related trAEs. Conclusions: This study confirms reproducible efficacy and safety of A+B in routine practice. Despite inferior OS and PFS compared to CP-A, A+B was associated with similar tolerability and radiologic response in CP-B patients, warranting prospective evaluation of A+B in this treatment deprived population.