Validation of Modified Albumin‐Bilirubin‐Tumor Stage as a Prognostic Model in Early‐Stage Hepatocellular Carcinoma Treated With Radiofrequency Ablation

ABSTRACT

Survival after radiofrequency ablation (RFA) for very early‐ or early‐stage hepatocellular carcinoma (HCC) remains heterogeneous because prognosis depends on both tumor burden and hepatic reserve. We retrospectively evaluated treatment‐naive patients with Barcelona Clinic Liver Cancer (BCLC) Stages 0 or A HCC treated with RFA as initial therapy from January 2002 to September 2016. Patients with Child‐Pugh class A or B and complete ablation were included. Four prespecified total scores, Child‐Pugh‐tumor (CP‐T), albumin‐bilirubin‐tumor (ALBI‐T), modified ALBI‐tumor (mALBI‐T), and platelet‐ALBI‐tumor (pALBI‐T) were compared using the Akaike information criterion (AIC), Harrell concordance index (C‐index), time‐dependent area under the receiver operating characteristic curve, calibration, and subgroup analyses. Of 1238 patients screened, 950 were included; 831 (87.5%) were Child‐Pugh class A. During a median follow‐up of 5.44 years, 417 deaths occurred; 5‐ and 10‐year overall survival rates were 58.0% and 35.8%, respectively. The American Joint Committee on Cancer eighth‐edition T stage illustrated the best tumor‐stage performance and was used in the total scores, while mALBI‐T had the lowest AIC (5052.2) and highest Harrell C‐index (0.628; 95% confidence interval, 0.601–0.655). Paired bootstrap differences versus CP‐T, ALBI‐T, and pALBI‐T were statistically significant, but modest, while mALBI‐T performed best at 5 years and in Child‐Pugh class A patients, whereas all models showed limited discrimination in Child‐Pugh class B patients. Apparent calibration was generally close, and mALBI‐T also had the best fit for recurrence‐free survival. mALBI‐T provides a simple prognostic model for long‐term survival stratification after complete ablation in early‐stage hepatocellular carcinoma, especially in Child‐Pugh class A patients.