Models can help predict liver cancer in people cured of hepatitis C

17 Sep 2020 Liz Highleyman
Originally published on www.infohep.org

Three new methods using readily available clinical parameters can help predict who will develop hepatocellular carcinoma (HCC) after achieving sustained virological response to hepatitis C treatment, according to studies presented at the recent 2020 Digital International Liver Congress.

Over years or decades, chronic hepatitis C virus (HCV) infection can lead to the development of cirrhosis and liver cancer. People who are cured of hepatitis C are less likely to develop HCC than those who are not successfully treated, but some risk remains, especially for people who have already progressed to advanced fibrosis or cirrhosis. Liver cancer is often diagnosed late, when it is more difficult to treat, so better methods to predict who is most likely to develop HCC are urgently needed.

Jessica Azzi of the ANRS-AFEF HEPATHER Study Group in Paris presented findings from an analysis of HCC predictive factors after sustained virological response (SVR) to direct-acting antiviral (DAA) therapy. The researchers developed a prognostic score and compared its performance against existing scoring systems.

The analysis included 3929 people with chronic HCV infection in the ANRS C022 HEPATHER cohort who were treated with DAAs and achieved SVR. They were negative for hepatitis B, did not yet have HCC and had not undergone liver transplantation.

Of the 2829 patients with cirrhosis (stage F4), 191 (6.8%) developed HCC. In contrast, among the 1097 people with moderate to advanced fibrosis (stage F2-F3), just 15 (1.4%) developed liver cancer. HCC occured approximately two years after SVR, on average.

The researchers identified eleven variables associated with development of HCC, including male sex, age greater than 64 years, HCV genotype 3, prolonged prothrombin time (a measure of blood clotting ability), alpha-fetoprotein (AFP) level, FIB-4 score (a fibrosis index based on laboratory tests), elevated cholesterol, oesophageal varices, prior treatment with interferon-based therapy, duration of HCV infection and being underweight. However, elevated cholesterol, body mass index and HCV duration were significant only for those with stage F2-F3 fibrosis.

They then assigned points for each of these variables and calculated a composite HCC risk score. Scores below 6 were considered low risk, scores of 6 to 11 were medium risk and scores of 11 or above were high risk.

After three years of follow-up, just 2% of people with low risk scores developed HCC, rising to 7% for those with medium risk scores and 23% for those with high risk scores. Those with high scores began to see a steep increase in liver cancer incidence even during the first year of follow-up, while those with low or medium scores saw a more gradual rise.

The score offers a "practical and easy tool to use in clinical practice to estimate HCC risk" and can help identify a high risk subgroup of patients for whom liver cancer screening would be cost-effective, Azzi said, adding that real-world studies are warranted to validate the scoring system.

In another French study, Prof. Pierre Nahon of Hôpital Jean Verdier in Paris and colleagues aimed to identify routinely measured biomarkers and changes in alpha feto-protein (AFP) levels before and after SVR that are associated with a higher risk of HCC in people with cirrhosis.

The researchers used data from ANRS CirVir, a prospective cohort of people with compensated cirrhosis who underwent regular HCC surveillance. The analysis included 717 participants who were followed for a median of 5.6 years; 413 of them (58%) achieved SVR. Levels of liver enzymes (ALT, AST and GGT), bilirubin, AFP, albumin, platelets and prothrombin time were assessed every six months.

Looking at participants prior to SVR (including those who were never cured), the investigators identified three clusters: people with inflammation and high AFP levels (26%), those who experienced liver failure (28%) and those with the least impaired lab values (46%). People with inflammation and high AFP had the highest risk of developing HCC, with the liver failure cluster not far behind. Those with the most favourable lab values had a much lower risk.

Among those who achieved SVR, 26% had persistent liver impairment, 23% continued to have elevated liver biomarkers and 22% had the least impaired values. In this group, people with persistent liver impairment (16%) or elevated biomarkers (14%) had a higher HCC risk, while those with the most favourable values had a lower rate (8%). Prof. Nahon suggested that people with persistent liver impairment after being cured of HCV may have been treated too late.

The investigators concluded that the pre-SVR inflammation and liver failure clusters represent two different risk profiles that together accounted for more than half of people who developed liver cancer. These profiles can persist even after SVR, identifying subgroups of patients who remain at risk for HCC.

"These analyses, based on novel statistical methods, suggest HCC surveillance can be refined and improved in order to tailor patient management to achieve optimum outcomes and increase cost-effectiveness," Prof. Nahon said in an EASL press release.

Finally, Prof. Gamal Shiha of the Egyptian Liver Research Institute and Hospital and colleagues aimed to develop a non-invasive scoring model for individualised HCC risk prediction for people with advanced fibrosis or cirrhosis. He noted that a "one size fits all" screening strategy for a growing population of people treated for hepatitis C may not be feasible, particularly in low- and middle-income countries.

The researchers screened more than 200,000 people for hepatitis C in 73 villages in Egypt. Of those screened, 7.8% were found to be HCV RNA positive and offered DAA treatment. The investigators identified 2372 people with no evidence of existing liver cancer who achieved SVR and completed at least a year of follow-up; nearly three-quarters had cirrhosis while the rest had stage F3 fibrosis.

Over a median two years of observation, 109 people (4.6%) developed HCC after the end of treatment. All but eight already had cirrhosis before HCV treatment initiation. Age over 54 years, male sex, pre-treatment fibrosis stage (F4 vs F3) and AFP and albumin levels were identified as risk factors for the development of HCC.

The researchers used these factors to develop a GES score by assigning points for each relevant variable. Participants were stratified into three groups with low risk (a score of 6.0 or lower; 58%), intermediate risk (6.0-7.5; 25%) or high risk (over 7.5; 18%). The score demonstrated high predictive accuracy, with HCC incidence rates of 1.2% for the low risk group, 3.3% for the intermediate risk group and 7.1%for the high risk group.

The GES score was then validated in two other groups. In an internal cohort of 422 patients with cirrhosis and 265 with stage F3 fibrosis, 70%, 14% and 17%, respectively, were classified low, intermediate and high risk. Over a median nine months of observation after the end of treatment, 14 people developed HCC: two (0.2%) in the low risk group, two (2.1%) in the intermediate group and 10 (8.7%) in the high risk group.

In an external cohort of 947 patients with cirrhosis and 394 with stage F3 fibrosis, 34%, 18% and 48% were deemed low, intermediate and high risk. Over about two years of post-treatment follow-up, only one person (0%) in the low risk group, six people (2.1%) in the intermediate group and 39 (6.1%) in the high risk group developed HCC.

This simple GES score using readily available parameters "can accurately stratify patients according to HCC risk," the researchers concluded. Shiha suggested that identifying people who will not benefit from continued HCC surveillance based on their estimated risk could enable a personalised surveillance strategy targeting those who are at high risk.

"These three studies reflect the complexity of understanding hepatocarcinogenesis and refute the idea that cure of HCV is equal to eliminating the risk of liver cancer," said Prof. Jordi Bruix of the University of Barcelona. "The proposed scores potentially represent a useful clinical tool to help inform patients about the risk of developing HCC after HCV is cured. These data also reinforce the importance of implementing HCC screening programmes in DAA-treated patients and the need to reinforce research efforts to identify the causes of liver cancer development despite cure of HCV."