The albumin-bilirubin (ALBI) model in hepatocellular carcinoma (HCC) may better predict hepatic failure in patients with traditionally low-risk cirrhosis following definitive stereotactic body radiotherapy (SBRT)

Abstract

Stereotactic body radiotherapy (SBRT) in hepatocellular carcinoma (HCC) is safe and effective in Child Pugh (CP) A cirrhosis, though a subset of patients may be at higher risk for progressive liver failure. The albumin-bilirubin (ALBI) model is a simple, novel, and validated prognostic assessment of hepatic cirrhosis in patients with HCC, but its utilization during liver SBRT is not well studied. Compared to conventional stratifications of cirrhosis, we hypothesize that the ALBI model is a sensitive predictor of hepatic function, survival, and liver failure following definitive SBRT for HCC in patients with CP-A cirrhosis. From 2009 to 2016, 40 patients with HCC and CP-A cirrhosis completed SBRT. The mean dose was 45 Gy (40 to 50 Gy) in 4–5 fractions. Local control defined as the absence of tumor progression was assessed with CT or MRI every 3–6 months after SBRT. ALBI, MELD (model of end stage liver disease), and CP scores were recorded before treatment and every 3–4 months after. ALBI scores were calculated from the formula (log(bilirubin[mcmol/L]) × 0.66) − (albumin[g/L] × 0.085) and classified as ALBI-1, ALBI-2, or ALBI-3. Initial scores were measured against progression to CP-B cirrhosis and survival. Kaplan-Meier statistical analysis was conducted via MedCalc software. Forty patients with 49 HCC lesions were analyzed in this IRB approved retrospective study. Median tumor diameter was 3.5 cm (1.4–8.9 cm). Eight patients underwent orthotropic liver transplant (median time to transplant was 12 months). Prior to SBRT, 17 patients had ALBI-1 and 23 had ALBI-2 scores.The actuarial local control at 2 years was 98%. With the median follow-up of 23.4 months, overall survival (OS) was 92 and 60% at 1 and 2 years. Freedom from progression to CP-B was 74 and 62% at 1 and 2 years. Compared to pre-treatment ALBI-2 patients, ALBI-1 patients correlated with higher freedom from CP progression (HR = 3.66, P = 0.03), but not OS (P = 0.09). At the last follow-up, 22 of 40 patients (55%) progressed to a higher ALBI grade (12 from ALBI-1 to 2 and 10 from ALBI-2 to ALBI-3). The 2-year OS for patients with an ALBI-1 (n = 5), ALBI-2 (n = 25), and ALBI-3 (n = 10) at last follow-up was 80, 74, and 30%, respectively (P = 0.03). Our study indicates that the ALBI index more precisely predicted liver failure and survival in HCC following SBRT compared to CP scoring. Its application may help select the most ideal candidates for SBRT among the often heterogeneous CP-A population. Further study regarding precautionary measures such as functional treatment planning for ALBI-2 patients treated with liver SBRT may be warranted.