Improved Criteria for Identifying Transplant Candidates with Hepatocellular Carcinoma

A new model that includes alpha-fetoprotein level predicts 5-year recurrence of posttransplant HCC better than the standard Milan criteria.

Liver transplantation is the best treatment option for early stages of hepatocellular carcinoma (HCC). Currently, the Milan criteria are used to identify transplantation-eligible patients with a low risk (10%–15%) for posttransplantation tumor recurrence. However, recent evidence has raised concern that these criteria might be too restrictive.

Using data from a cohort of 537 patients in France who received a liver transplant for HCC, researchers developed a new predictive model that included alpha-fetoprotein (AFP) level. This AFP model was then validated in a separate cohort of 435 liver transplant recipients with HCC who were part of a national transplantation program in France and were followed prospectively.

Regression analysis showed that the number of tumors, tumor size, and AFP level were independent predictors of 5-year tumor recurrence (the primary endpoint). Three combinations of these variables defined patients as low risk:

• One to three nodules with maximum tumor diameter <3 cm and AFP 1000 ng/mL
• One to three nodules with maximum tumor diameter 3 to 6 cm and AFP 100 ng/mL
• More than four nodules with maximum tumor diameter <3 cm and AFP 100 ng/mL

In a simplified model, researchers transformed Beta coefficients of variables from the regression model into points, adding these to obtain a score. A score of 2 was the cutoff for low versus high risk. In the validation cohort, a score >2 (vs. 2) was associated with a higher 5-year recurrence rate (50.6% vs. 8.8%, P<0.001) and a lower 5-year survival rate (47.5% vs. 67.8%, P=0.002). The AFP model was superior to the Milan criteria in predictive performance based on a comparison of predicted and observed recurrence events with each. Some patients identified by the AFP model as low risk (e.g., patients with AFP 100 ng/mL) did not meet Milan criteria, and some identified as high risk (e.g., patients with AFP >1000 ng/mL) did meet the Milan criteria.

Comment: These new selection criteria identified not only transplantation candidates with low risk for HCC recurrence who would have been excluded using the standard Milan criteria, but also those at high risk for HCC recurrence who would have been eligible under those criteria. With the rise in the number of patients with HCC listed for transplantation, continual refinement of the selection criteria is critical. This AFP model could be the next iteration of those criteria.

— Atif Zaman, MD, MPH

Published in Journal Watch Gastroenterology September 28, 2012

Citation(s):

Duvoux C et al. Liver transplantation for hepatocellular carcinoma: A model including alpha-fetoprotein improves the performance of Milan criteria. Gastroenterology 2012 Oct; 143:986.

Medline abstract (Free)

Source