New risk model may help select patients with HCC for liver transplantation

Last Updated: 2017-06-08

By Joan Stephenson

NEW YORK (Reuters Health) - A new risk score to predict five-year survival after liver transplant of patients with hepatocellular carcinoma (HCC) displayed better prognostic ability than other models when tested in a database of nearly 14,000 patients.

In a retrospective cohort analysis, the Hazard Associated with Liver Transplantation for Hepatocellular Carcinoma (HALT-HCC) model was “superior to the currently adopted model (Milan criteria) and others previously described” in predicting overall mortality in patients with HCC who underwent transplantation, corresponding author Federico N. Aucejo, of the Cleveland Clinic in Ohio, told Reuters Health by email.

“Possible implications for clinical practice by adopting the HALT-HCC model as the new standard are superior patient selection criteria, resulting in improved survival and more effective organ utilization,” he said.

The basis for the Milan criteria, considered the gold standard for nearly two decades for selecting patients with HCC who are most likely to benefit from liver transplantation, is tumor morphology (size and number of tumors).

However, the Milan criteria “are far from perfect,” the researchers noted, because they assess only tumor morphology, not aspects of tumor biology (such as genetic factors) that influence a tumor’s metastatic potential. In addition, the Milan model is “binary” (meaning a patient either does or does not meet the model’s criteria), which excludes variance that might be associated with outcomes, the authors noted.

“We decided to create a continuous risk score incorporating multiple and diverse factors inherent to tumor morphology, tumor biology, and patient disease severity,” Dr. Aucejo said.

To develop the HALT-HCC model, the researchers using data from 420 patients with HCC who underwent liver transplantation at the Cleveland Clinic Foundation (CCF) between 2002 and 2014.

Three variables - the tumor burden score (TBS), alpha-fetoprotein (AFP), and MELD-sodium (MELD-Na), indicating tumor morphology, tumor biology, and disease severity, respectively - were “significantly associated with overall survival in the multivariate analysis,” the researchers report in The Lancet Gastroenterology and Hepatology, online May 22.

“These characteristics were used to generate an easily computed continuous risk equation” for HALT-HCC, they said. When they applied the model to the CCF cohort, HALT-HCC scores ranged from 2.40 to 46.42, with increasing scores indicating increasingly worse five-year overall survival.

To validate and calibrate the model’s assessment of risk, the researchers used nationwide data from 13,717 patients with HCC from the Scientific Registry of Transplant Recipients (SRTR) who underwent transplantation during the same period as the CCF cohort.

In the SRTR cohort as well, prognosis worsened with increasing HALT-HCC score, with five-year overall survival rates of 78.7%, 74.5%, 71.8%, and 61.5% for quartiles 1 to 4, respectively (p<0.0001).

Multivariate Cox modeling showed that HALT-HCC was significantly associated with overall survival (hazard ratio, 1.06 per point), even after adjustment for risk factors not related to HCC.

In addition, the HALT-HCC predictive model identified patients who were within Milan criteria and had a poor outcome after transplantation, and patients who were beyond Milan criteria and exhibited a favorable outcome, Dr. Aucejo said.

The model should be further validated in large international cohorts to address its value across heterogeneous ethnicities, he said.

“While waiting for a personalized medicine approach based on genomic data to provide useful information for predicting which patients will have a poor oncological outcome, integration of tumor features with patients’ characteristics will probably make the extension of liver transplantation criteria for hepatocellular carcinoma safer,” Drs. Patrizia Burra, Alberto Zanetto, and Martina Gambato, of Padua University Hospital in Italy, note in an accompanying commentary.

The HALT-HCC is “a new tool that could help us to move towards an integration-oriented selection of patients in the setting of liver transplantation for hepatocellular carcinoma,” they write.

The inclusion of such additional factors as severity of liver disease and a tumor activity blood test to inform predictions of how well patients who undergo liver transplantation for HCC will do “is one major advancement of HALT-HCC over existing models,” Dr. Scott Biggins, of the University of Washington in Seattle, told Reuters Health by email.

“The most important finding of the study is that the HALT-HCC model can identify groups that under the current system would be excluded from liver transplant but actually could be suitable patients for the procedure,” said Dr. Biggins, who studies ways to improve recipient selection for liver transplantation but who was not involved in the new work.

“Conversely, the HALT-HCC can be used to identify patients the current system predicts should do well with liver transplant but actually might not benefit as much as other patients,” he said.

“If adopted by policy makers, the use of the HALT-HCC model could improve the use scarce lifesaving donor livers,” he said.

SOURCE: http://bit.ly/2qMbZT2 and http://bit.ly/2rhxG1a

Lancet Gastroenterol Hepatol 2017