Comparative effectiveness of sorafenib, lenvatinib, and nivolumab as first-line systemic therapy for patients with advanced hepatocellular carcinoma and Child-Pugh class B cirrhosis treated at VA Medical Centers.

4102 Background: Up to 90% of cases of hepatocellular carcinoma (HCC) in the U.S. occur in patients with cirrhosis. As prognosis in patients with decompensated cirrhosis may be more related to liver function than tumor stage, patients with Child-Pugh (CP) class B and C cirrhosis are often excluded from randomized trials despite representing up to 30% of patients treated with systemic therapy in clinical practice. While prospective data such as the GIDEON registry and CHECKMATE-040 have demonstrated safety of sorafenib and nivolumab in CP B cirrhosis, studies of comparative effectiveness of sorafenib, lenvatinib, and nivolumab are lacking. Methods: We performed a retrospective cohort study of patients with HCC and CP B cirrhosis at U.S. Veterans Affairs Medical Centers who initiated first systemic therapy with sorafenib, lenvatinib, or nivolumab between 9/22/2017 and 2/13/2021. Overall survival (OS) by first systemic therapy was assessed with Kaplan Meier analysis and Cox proportional hazards modeling with pre-specified candidate covariates and a backward elimination approach for variable selection. Missing data were imputed by multiple imputation with chained equations. Results: Among 401 CP B patients undergoing first systemic therapy, 98% were male, 29% had macrovascular invasion (MVI), 35% had extrahepatic spread (EHS), 57%/29%/13% had CP score 7/8/9, 5% had hepatic encephalopathy, 37% had ascites, 63%/2%/36% had HCV/HBV/non-viral cirrhosis, 63% had prior embolization or ablation, and 35% had ECOG performance status of > = 2. Of these patients, 320 received sorafenib, 33 received lenvatinib, and 48 received nivolumab. In univariate analysis, median OS was 4.8 months (95% CI 3.8 – 6.0; 1 year OS 21.8%) in the sorafenib cohort, 6.9 months (95% CI 4.4 – 8.4 months; 1 year OS 27.8%) in the lenvatinib cohort, and 7.7 months (95% CI 3.9 – 9.1 months; 1 year OS 31.7%) in the nivolumab cohort. Covariates with at least one category significantly associated with survival that were retained in the multivariate model included CP score, MELD-Na, BMI, log(AFP), time from diagnosis to systemic treatment, ECOG performance status, Cirrhosis Comorbidity index, and VA complexity level (all p < 0.05); MVI, EHS, and age were retained in the model as pre-specified despite lack of statistical significance. Compared with sorafenib, nivolumab was associated with significantly lower hazard of death (adjusted HR 0.63; 95% CI 0.43 – 0.91; p = 0.015) while hazard of death for lenvatinib treatment was not significantly different (adjusted HR 0.83; 95% CI 0.54 – 1.28, p = 0.41). Conclusions: Among patients with HCC and CP class B cirrhosis undergoing first systemic therapy, nivolumab was associated with significantly lower hazard of death compared to sorafenib after adjusting for important covariates.