Predicting the posttransplant prognosis of patients with hepatocellular carcinoma: What lies beneath alpha-fetoprotein?

We read with interest the article by Berry and Ioannou1 describing the prognostic usefulness of pretransplant serum alpha-fetoprotein levels in a large series of patients who underwent liver transplantation for hepatocellular carcinoma in the United States between 2002 and 2011. They demonstrated that the serum alpha-fetoprotein level retained a significant prognostic capability after they controlled for several well-known recipient and graft characteristics and even when the tumoral burden, assessed with either the Milan criteria or the University of California San Francisco criteria, was within favorable limits. All in all, the results of this study1 add further evidence to support a prognostic role of serum alpha-fetoprotein in patients with hepatocellular carcinoma undergoing liver transplantation, and they thus confirm previous reports.2, 3 However, in our opinion, one of the most important results of this study is the fact that when the patients who underwent locoregional treatment before liver transplantation and whose alpha-fetoprotein levels were assessed multiple times were evaluated, the group whose serum alpha-fetoprotein levels decreased, even minimally and without normalization, had no excess posttransplant mortality, whereas the posttransplant mortality rate remained elevated for the patients with no alpha-fetoprotein response. Probably more importantly, the excess mortality was attributed to malignancies, although the explored database did not allow more in-depth discrimination between recurrent and de novo hepatocellular carcinoma.1 Nevertheless, what is noteworthy here is that alpha-fetoprotein seems to be a good indicator of the biological aggressiveness of the tumor, as the prognostic ability of this tumor marker was independent of tumoral burden. Indeed, it has been previously shown that high alpha-fetoprotein levels are associated with some histological and biological characteristics that denote poorer prognoses for patients with hepatocellular carcinoma, and we have recently shown that once a small hepatocellular carcinoma is successfully ablated, the prognosis is independent of the baseline serum alpha-fetoprotein levels.4, 5 Lastly, in a large series of patients with hepatocellular carcinoma, we have shown that increasing serum alpha-fetoprotein levels are consistently and significantly associated with poor tumor differentiation according to the Edmondson-Steiner classification (E. G. Giannini for the Italian Liver Cancer Group, unpublished data, 2013). In this respect, we feel that what is missing from Berry and Ioannou's study1 is a thorough analysis of hepatocellular histology. Because this figure is more easily and accurately obtained from explanted livers versus bioptical samples, an analysis of serum alpha-fetoprotein levels and hepatocellular carcinoma histology and differentiation might have provided valuable insights into the possible prognostic use of this tumoral marker as a proxy for poor differentiation and greater biological aggressiveness and thus supported its use on a sound pathophysiological basis also. Edoardo G. Giannini, M.D., Ph.D., F.A.C.G. Vincenzo Savarino, M.D. Gastroenterology Unit Department of Internal Medicine IRCCS-Azienda Ospedaliera San Martino-IST University of Genoa Genoa, Italy