165P - Prognostic factors in sorafenib-treated hepatocellular carcinoma: Multicentre analysis of a European population sample

Background

Sorafenib has been the standard-of-care in advanced hepatocellular carcinoma (HCC). Prognostic factors such as BCLC stage and macroscopic vascular invasion (MVI), together with predictive factors of response such as neutrophil-lymphocyte ratio (NLR), cirrhosis etiology and extrahepatic spread (EHS) were found to be associated to survival outcomes. Nevertheless, these factors need to be fully validated. We intend to explore and identify prognostic factors for overall survival (OS) in a European population sample treated with sorafenib.

Methods

Multicentre retrospective evaluation of patients’ (pts) charts of HCC pts with ECOG performance status (PS) ≤2 treated with sorafenib between 2008 and 2019 in two Portuguese centres. Alfa- fetoprotein (AFP) < 200 ng/mL, EHS, Model for End-Stage Liver Disease-Na (MELD-Na) < 17, cutaneous toxicity, NLR < 2.3, MVI and previous locoregional treatment (LRT) were considered for stratification. Each factor was analyzed by Kaplan-Meier method for OS estimation and a multivariate Cox regression was performed.

Results

Eighty-six pts were analyzed, 15% female and with a median age (med) of 65.5 years [36-83]. Considering underlying cause, 45% pts had alcohol-related liver disease, 23% hepatitis C virus, 14% hepatitis B virus and 5% non-alcoholic steatosis hepatitis. Median OS was 8.5 months (m). Kaplan-Meier analysis of OS proved to be statistical significant in the following factors: AFP < 200 ng/mL med OS 12.5 m and ≥ 200 ng/mL 5.4 m (p < 0.0001); MELD-Na < 17 med OS 10.4 m and ≥ 17 4.8 m (p=0.015); absence of EHS med OS 10.0 m and presence of EHS med OS 7.3 m (p=0.028); previous LRT med OS 11.2 m and no LRT med OS 6.6 m (p=0.005). Multivariate Cox analysis identified extra-hepatic involvement (HR 1.002, p=0.041) and no previous LRT (HR 1.871, p=0.009) as risk factors.

Conclusions

HCC pts represent a heterogeneous population with many factors that potentially affect prognosis, such as geographic etiology difference in hepatic cirrhosis. In our sample, analysis of single-variable subgroups estimated greater OS in pts with AFP < 200 ng/mL, MELD-Na < 17, absence of EHS and previous LRT. The multivariate regression showed EHS and no previous LRT are independent factors for worse prognosis.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

Has not received any funding.

Disclosure

All authors have declared no conflicts of interest.