OP034 LIVER DECOMPENSATION AS A NEW OUTCOME AND PROGNOSTIC INDICATOR IN CIRRHOTIC PATIENTS WITH HEPATOCELLULAR CARCINOMA TREATED WITH ATEZOLIZUMAB PLUS BEVACIZUMAB: A REAL-LIFE MULTICENTER STUDY

Introduction: The combination of Atezolizumab plus Bevacizumab (atezobeva) has changed the treatment landscape of advanced hepatocellular carcinoma (HCC), but its efficacy and safety in patients with impaired liver function or portal hypertension is still debated. Aims & Methods: In this multicenter, real-life, study we compared 247 patients with unresectable HCC treated with atezobeva, matched to 494 patients treated with tyrosine kinase inhibitors (TKIs). We evaluated the impact of liver function and decompensation on treatment outcomes. Results: Patients in the atezobeva group showed longer median overall survival (OS; 18·30 vs 10·20 months, p<0·0001), longer time to progression (TTP; 13·60 vs 10·10 months, p=0·0002) and increased progression-free survival (PFS; 10·50 VS 6·10 months, p<0·0001) compared with the TKIs group. Although survival was better in Child Pugh A patients than in Child Pugh B patients (20·20 vs 9·83 months, p=0·0008), no clear differences in TTP and overall treatment safety were observed· Liver decompensation occurred in 63 (25·51%) patients, specifically 41 (27·89%) Child Pugh A and 22 (51·16%) Child Pugh B and was associated with worse survival (10·50 vs 20·20 months; p=0·006). Among decompensated patients, 26 patients (41·26%) were able to resume atezobeva after decompensation, achieving an OS comparable to those who never decompensated (20·87 vs 20·2 months; p=0·77), and significantly better than those who permanently stopped treatment (8·07 months; p=0·02). Treatment interruption following decompensation was an independent predictor of survival (HR 2·04, 95% CI 1·25-3·33, p=0·004), while ALBI score 2-3 (HR 2·21, 95%CI 1·07-4·55; p=0·03), Child Pugh class B (HR 1·81, 95% CI 1·03-3·18; p=0·04), portal hypertension (HR 2·54, 95%CI 1·37-4·72; p=0·003) and sAEs (HR 1·71, 95%CI 1·04-2·82; p=0·04) emerged as predictors of decompensation. Time to decompensation was similar for patients with ALBI 2, regardless of Child Pugh class, and the probability of recovery from decompensation was similar for Child A and B patients. Conclusion: In this real-life multicentric study atezobeva was effective in both Child Pugh A and B patients. Child Pugh B patients despite their higher risk of decompensation have an equal chance of recovery. The possibility to resume treatment after an episode of decompensation highlights the necessity of integrated hepato-oncological management, optimizing the therapeutic efficacy of atezobeva.