Objective: The Barcelona Clinic Liver Cancer algorithm is the most widely used staging system for hepatocellular carcinoma, but the intermediate stage of this classification comprises a very heterogeneous group of patients with different survival probabilities. The aim of our study was to construct a simple prognostic index for identifying subgroups of patients with different prognoses within the intermediate stage.
Patients and methods: Three-hundred and seven patients were retrospectively analyzed and randomly divided into a training sample (n=205), from which the model was developed, and a test sample (n=102), to independently assess the model's performance.
Results: Four variables were retained in the final multivariate model: hepatic failure, number of nodules, α-fetoprotein, and albumin, with hazard ratios equal to 2.22 (95% confidence interval: 1.52-3.24), 1.47 (1.00-2.18), 2.34 (1.56-3.52), and 1.75 (1.26-2.44), respectively. The score system was derived by summing up the linear weights assigned to the four covariates according to the observed regression coefficients. The score ranged between 4 and 13; to avoid sparse-data bias arising from small numbers within strata, only four categories (4-5, 6-7, 8-9, 10-13) were identified. The prognosis worsened significantly with increasing score and the C-index for discriminatory accuracy was equal to 0.66 (95% confidence interval: 0.60-0.72). The score was validated in the test sample (log-rank test P=0.02). Similar results were found when evaluating the score as a continuous variable.
Conclusion: The simple prognostic index predicts survival in patients with intermediate-stage hepatocellular carcinoma. This score might help guide treatment selection and patient stratification in clinical studies.