• Potential conflict of interest: Nothing to report.


We thank Dr. Pawlik and colleagues for their interest in our paper,1 but we respectfully disagree with their comments. First, our intention was not to propose a new treatment algorithm for hepatocellular carcinoma (HCC), but rather validate existing prognostic classifications in our patients with HCC.

Second, Fig. 2B was misinterpreted. The best overall survival was in patients who underwent surgical therapy followed by ablation as primary treatment, then systemic treatments and no treatment. In addition, 91% of those undergoing transplantation and 100% of those undergoing resection were Barcelona Clinic Liver Cancer (BCLC) Stage A; ablation as primary treatment was performed in 60% BCLC stage A, 36% stage B and 4% stage C. While some of the patients who underwent ablation or chemoembolization as primary treatment may have had tumor burdens in which resection may have been feasible, hepatic dysfunction or poor performance status hindered the ability to perform resection.

Third, the authors showed 2 studies as evidence for performing palliation with surgical resection for HCC >5 cm.2, 3 The study by Ng et al.2 found that small tumors had better survival than large tumors (<5 cm), but on multivariate analysis cirrhosis and multinodular tumors were among the factors that predicted poor prognosis. The study by Pawlik et al.3 showed a 3-year survival of 22% in patients with cirrhosis. Two randomized controlled trials showed that chemoembolization in patients with cirrhosis and HCC >5 cm had better 3-year survivals.4

Fourth, our algorithm and the BCLC classification do not limit resection according to tumor size. However, the best results of resection are in those with tumors <5 cm and normal hepatic function; and when limited to single compared to multinodular tumors as shown in the largest series to date of resection for HCC.5

Finally, the new American Joint Committee on Cancer (AJCC) is not considered the gold standard for staging or prognosis of HCC, and its usefulness has been challenged by others.6, 7 In fact, the panel of experts at the recent European Association Study Liver–American Association Study of Liver Disease–Japanese Society of Hepatology Monothematic HCC meeting endorsed the BCLC classification. The external validation of the AJCC classification excluded postoperative deaths,8 when an intent-to-treat analysis was performed, the AJCC classification was not an important prognostic variable,9 and other prognostic classifications were not compared to AJCC. However, the most important limitation of the new AJCC classification is that it is based on postoperative assessment and not at the time of HCC diagnosis when prognosis is of outmost importance. In addition, the new AJCC classification ignores important prognostic factors such as performance status, age, hepatic function and the presence of portal hypertension at the time of HCC diagnosis. Also, pathological information required to define microvascular invasion is available in less than one third of cases, and the new AJCC does not differentiate micro/macrovascular invasion (except for the main trunk), when both entities have different prognostic value as recognized by Poon et al.10

Jorge A. Marrero M.D., M.S.*, Robert J. Fontana M.D.*, Grace L. Su M.D.*, Hari S. Conjeevaram M.D.*, Anna S. Lok M.D.*, * Division of Gastroenterology, University of Michigan, Ann Arbor, MI.