New Model Estimates HCC Risk in Patients With Chronic HBV

A new prognostic model could potentially predict and stratify the risk for hepatocellular carcinoma (HCC) among patients with chronic hepatitis B (CHB) who are noncirrhotic and not indicated for antiviral treatment.

The model, called Revised REACH-B or reREACH-B, stems from cohort studies in Hong Kong, South Korea, and Taiwan, and looks at the nonlinear parabolic association between serum hepatitis B virus (HBV) DNA levels and HCC risk.

“Current clinical practice guidelines don’t advocate antiviral treatment for patients with CHB who don’t show elevated alanine aminotransferase (ALT) levels, even in those with high HBV viral loads,” said co-author Young-Suk Lim, MD, PhD, professor of gastroenterology at the University of Ulsan College of Medicine and Asan Medical Center in Seoul, Korea.

“This stance is rooted in the notion that patients in the immune-tolerant phase are at very low risk for developing HCC,” Lim said. “However, the immune-tolerant phase includes patients with HBV DNA levels who face the highest risk for HCC, and many patients with moderate HBV viremia fall into an undefined gray zone.”

The study was published online in Annals of Internal Medicine.

Validating reREACH-B

During a course of CHB, HBV viral loads and HCC risks evolve over time due to viral replication and host immune responses, Lim explained. Most patients typically move to seroclearance and an “inactive hepatitis” phase, but about 10%-20% can progress to a “reactivation” phase, where HBV DNA levels and ALT levels increase, which can increase HCC risk as well.

In a previous cohort study in Taiwan, a prognostic model called Risk Estimation for HCC in CHB — or REACH-B — found the risk for HCC increases tenfold with increasing levels of HBV DNA up to 5 log₁₀IU/mL in noncirrhotic patients with CHB, regardless of ALT levels. Another cohort study in South Korea found a nonlinear parabolic association between HCC risk and HBV DNA levels up to 9 log₁₀ IU/mL, with the highest risks found for moderate HBV DNA levels around 6 log₁₀ IU/mL.

In this study, Lim and colleagues developed a prognostic model to integrate the nonlinear relationship and validated it externally, as well as compared it with the previous REACH-B model. The Revised REACH-B model incorporates six variables: Age, sex, platelet count, HBV DNA level, ALT, and hepatitis B e-antigen (HBeAg).

The study included 14,378 treatment-naive, noncirrhotic adults with CHB and serum ALT levels

Among the Asan cohort, the mean age was 45 years, 29.9% were HBeAg positive, median HBV DNA levels were 3.1 log₁₀ IU/mL, and the median ALT level was 25 U/L. In the external cohort, the mean age was 46 years, 21% were HBeAg positive, median HBV DNA levels were 3.4 log₁₀ IU/mL, and the median ALT level was 20 U/L.

In the Asan cohort, 435 patients (6.3%) developed HCC during a median follow-up of 10 years. The annual HCC incidence rate was 0.63 per 100 person-years, and the estimated cumulative probability of developing HCC at 10 years was 6.4%.

In the external cohort, 467 patients (6.3%) developed HCC during a median follow-up of 12 years. The annual HCC incidence rate was 0.42 per 100 person-years, and the estimated cumulative probability of developing HCC at 10 years was 3.1%.

Overall, the association between HBV viral load and HCC risk was linear in the HBeAg-negative groups and inverse in the HBeAg-positive groups, with the association between HBV viral load and HCC risk showing a nonlinear parabolic pattern.

Across both cohorts, patients with HBV DNA levels between 5 and 6 log₁₀ IU/mL had the highest risk for HCC in both the HBeAg-negative and HBeAg-positive groups, which was more than eight times higher than those HBV DNA levels ≤ 3 log₁₀ IU/mL.

For internal validation, the Revised REACH-B model had a c-statistic of 0.844 and 5-year area under the curve of 0.864. For external validation across the three external cohorts, the reREACH-B had c-statistics of 0.804, 0.808, and 0.813, and 5-year area under the curve of 0.839, 0.860, and 0.865.

In addition, the revised model yielded a greater positive net benefit than the REACH-B model in the threshold probability range between 0%-18%.

“These analyses indicate the reREACH-B model can be a valuable tool in clinical practice, aiding in timely management decisions,” Lim said.

Considering Prognostic Models

This study highlights the importance of recognizing that the association between HBV DNA viral load and HCC risk isn’t linear, said Norah Terrault, MD, chief of Gastroenterology and Hepatology at the Keck School of Medicine at the University of Southern California, Los Angeles.

“In contrast to most chronic liver diseases where liver cancer develops only among those with advanced fibrosis/cirrhosis, people with chronic hepatitis B are at risk prior to the development of cirrhosis,” she said. “Risk prediction scores for HCC can be a useful means of identifying those without cirrhosis who should be enrolled in HCC surveillance programs.”

For instance, patients with HBV DNA levels 10 IU/mL or > 8 log₁₀ IU/mL don’t have an increased risk, Terrault noted. However, the highest risk group appears to be around 5-6 log₁₀ IU/mL.

“Future risk prediction models should acknowledge that relationship in modeling HCC risk,” she said. “The re-REACH-B provides modest improvement over the REACH-B, but further validation of this score in more diverse cohorts is essential.”

The study received financial support from the Korean government and grants from the Patient-Centered Clinical Research Coordinating Center of the National Evidence-based Healthcare Collaborating Agency and the National R&D Program for Cancer Control through the National Cancer Center, which is funded by Korea’s Ministry of Health and Welfare. Lim and Terrault reported no relevant disclosures.

Carolyn Crist is a health and medical journalist who reports on the latest studies for Medscape Medical News, MDedge, and WebMD.