Comparison between observed survival after resection of transplantable hepatocellular carcinoma and predicted survival after listing through a Markov model simulation.

There is still some debate on whether hepatic resection or liver transplantation should be the initial treatment for hepatocellular carcinoma (HCC) in compensated cirrhosis. Clinical data and observed survivals of 150 transplantable patients (within Milan criteria) resected for HCC were reviewed and their predicted survival after listing for liver transplantation was calculated using a Markov model simulation. Differences between observed and predicted survival estimates were explored by standardized differences (d). The mean observed survival within 5 years after surgery was 45.35 months, and the predicted survival after listing was 49.18 months (d = 0.265). The largest gain in life-expectancy with liver transplantation would be obtained in patients with Model for End-stage Liver Disease (MELD) score >9 (d = 0.403); conversely, observed and predicted survivals were similar in HCV+ patients (d = -0.002) and in patients with MELD ≤9 (d = -0.057). For T1 tumors, the observed mean estimate of survival after hepatic resection was higher than that predicted by the simulation (d = -0.606). In conclusion, in HCV patients and in those with very well compensated cirrhosis, hepatic resection could lead to results similar to those of transplantation strategy for HCC within Milan criteria; HCC T1 patients are probably best served by resection as first-line therapy rather than listing for transplantation.