Prediction of progression-free survival in patients presenting with hepatocellular carcinoma within the Milan criteria

Transplantation is the treatment of choice for hepatocellular carcinoma (HCC) meeting the Milan criteria. HCC and chronic liver diseases have distinct natural histories for which an equitable transplant policy must account. We enrolled and prospectively followed at a single center 206 consecutive HCC patients that presented within the Milan criteria. Patients were treated per the Barcelona Clinic Liver Cancer (BCLC) algorithm; 95% received resection, ablation, or transarterial chemoembolization. The median follow-up was 16 months. Progression occurred in 84 patients, and 8 patients died. Risk factors for the time to disease progression (death or progression beyond T2) were analyzed in 170 patients with a complete data set. Risk factors with the strongest relationship to progression included tumor diameter and tumor persistence/recurrence after local therapy (hazard ratios of 1.51 and 2.75, respectively, when transplanted patients were censored at the time of transplantation and hazard ratios of 1.53 and 3.66, respectively, when transplantation was counted as an event; P < or = 0.0001). To evaluate the current Model for End-Stage Liver Disease (MELD) exception, we compared the expected progression rate (PR) with our observed PR in 133 stage T2 patients. The current policy resulted in a large overestimation of the PR for T2 HCC and an unsatisfactory performance [Harrell's concordance index (C index) = 0.60, transplant censored; C index = 0.55, transplant as progression]. Risk factors for progression that were identified by univariate analysis were considered for multivariate analysis. With these risk factors and the patients' natural MELD scores, an adjusted model applicable to organ allocation was generated, and this decreased the discrepancy between the expected and observed PRs (C index = 0.66, transplant censored; C index = 0.69, transplant as progression). In conclusion, the current MELD exception largely overestimates progression in T2 patients treated according to the BCLC guidelines. The tumor response to resective or ablative treatment can predict tumor progression beyond the Milan criteria, and it should be taken into account in models designed to prioritize organ allocation.