Prognosticating Survival in Hepatocellular Carcinoma with Elevated Baseline Alpha-fetoprotein Treated with...

To investigate laboratory parameters as predictors of overall survival (OS) for hepatocellular carcinoma (HCC) treated with radioembolization and develop/validate a scoring system.

With IRB approval, we included all patients with baseline alpha-fetoprotein (AFP) > 100 ng/dL from our prospectively acquired HCC radioembolization database. Neutrophil-lymphocyte ratio, albumin-bilirubin (ALBI), and AFP were measured at baseline and at 1-, 3-, and 6-month post-radioembolization Landmarks. OS was assessed from these Landmarks. Univariate/multivariate analyses were performed to evaluate OS predictability of these parameters. Baseline Imaging, Laboratory, and Combination scoring systems were developed. Developing/validating groups were created to investigate/validate the score's OS predictability. Time-dependent receiver operating characteristics (ROC) were evaluated. Patients were stratified into groups I, II, and III by using 25th and 75th percentile cutoffs according to change in Laboratory Score from baseline.

345/401 (86%), 238/401 (59%), and 167/401 (42%) patients had laboratory parameters available at the 1-, 3-, and 6-month Landmarks, respectively. ALBI and AFP were significant OS prognosticators at all Landmarks. The Laboratory Score [ALBI + (0.3 × LnAFP)] was developed/internally validated to predict OS from these Landmarks. Areas under the curve of time-dependent ROCs of the Baseline Imaging vs. Laboratory scores in predicting patient OS post 3 and 6 months Landmarks were 0.56 versus 0.82 and 0.57 versus 0.77, respectively. OS differences in groups I, II, and III according to change in Laboratory Score from baseline were significant (p < 0.001).

Post-radioembolization AFP and ALBI scores were significant OS prognosticators. A decrease in post-therapeutic Laboratory Score, which combines AFP and ALBI, correlates with an improved OS.