Utility of Prediction Scores for Hepatocellular Carcinoma in Patients with Chronic Hepatitis B Treated with Nucleos(t)ide Analogues

Hepatocellular Carcinoma Risk Scores from Modeling to Real Clinical Practice in Areas Highly Endemic for Hepatitis B Infection

Hepatocellular carcinoma (HCC) accounts for the majority of primary liver cancers and represents a global health challenge. Liver cancer ranks third in cancer-related mortality with 830,000 deaths and sixth in incidence with 906,000 new cases annually worldwide. HCC most commonly occurs in patients with underlying liver disease, especially chronic hepatitis B virus (HBV) infection in highly endemic areas. Predicting HCC risk based on scoring models for patients with chronic liver disease is a simple, effective strategy for identifying and stratifying patients to improve the early diagnosis rate and prognosis of HCC. We examined 23 HCC risk scores published worldwide in CHB patients with (n=10) or without (n=13) antiviral treatment. We also described the characteristics of the risk score's predictive performance and application status. In the future, higher predictive accuracy could be achieved by combining novel technologies and machine learning algorithms to develop and update HCC risk score models and integrated early warning and diagnosis systems for HCC in hospitals and communities.

Risk stratification and early detection biomarkers for precision HCC screening

Hepatocellular carcinoma (HCC) mortality remains high primarily due to late diagnosis as a consequence of failed early detection. Professional societies recommend semi-annual HCC screening in at-risk patients with chronic liver disease to increase the likelihood of curative treatment receipt and improve survival. However, recent dynamic shift of HCC etiologies from viral to metabolic liver diseases has significantly increased the potential target population for the screening, whereas annual incidence rate has become substantially lower. Thus, with the contemporary HCC etiologies, the traditional screening approach might not be practical and cost-effective. HCC screening consists of (i) definition of rational at-risk population, and subsequent (ii) repeated application of early detection tests to the population at regular intervals. The suboptimal performance of the currently available HCC screening tests highlights an urgent need for new modalities and strategies to improve early HCC detection. In this review, we overview recent developments of clinical, molecular, and imaging-based tools to address the current challenge, and discuss conceptual framework and approaches of their clinical translation and implementation. These encouraging progresses are expected to transform the current "one-size-fits-all" HCC screening into individualized precision approaches to early HCC detection and ultimately improve the poor HCC prognosis in the foreseeable future.

Comparative Performance of 14 HCC Prediction Models in CHB: A Dynamic Validation at Serial On-Treatment Timepoints

INTRODUCTION

To assess comparative performance of 14 hepatocellular carcinoma (HCC) prediction models in chronic hepatitis B (CHB) patients using on-treatment values at different timepoints.

METHODS

Based on a nationwide prospective cohort of 986 treatment-naive CHB patients undergoing entecavir therapy with every 26-week follow-up, 14 HCC risk scores were calculated using on-treatment values at week 26, 52, 78, and 104, respectively. Model performance predicting 3-year HCC was assessed using time-dependent area under the receiver operating characteristic curve (AUC) and calibration index. Model cutoffs were validated through common diagnostic accuracy measures.

RESULTS

During median 4.7-year follow-up, 56 (7.5%) developed HCC. Discrimination using on-treatment values within first 2 years was generally acceptable for most models (AUCs ranging from 0.68 to 0.81), except for REACH-B, NGM-HCC, and PAGE-B, although AUCs slightly decreased from week 26 to 104. Of these, REAL-B, CAMD, GAG-HCC, AASL-HCC, LSM-HCC, mPAGE-B, and mREACH-BII showed highest discrimination with AUCs ranging from 0.76 to 0.81, 0.72 to 0.76, 0.70 to 0.76, and 0.71 to 0.74 when reassessment at week 26, 52, 78, and 104, respectively. With reassessment within first 2 years, both REAL-B and CAMD calibrated well (Brier score ranging from 0.037 to 0.052). Of 9 models reporting cutoffs, REAL-B, AASL-HCC, and mPAGE-B using on-treatment values could identify 30%-40% of patients as low risk with minimal HCC incidence in the low-risk group (0.40% [REAL-B]-1.56% [mPAGE-B]).

DISCUSSION

In this undergoing antiviral treatment CHB cohort, most HCC prediction models performed well even using on-treatment values during first 2 years, particularly REAL-B, AASL-HCC, CAMD, and mPAGE-B model.

Hepatocellular Carcinoma Prediction Models in Chronic Hepatitis B: A Systematic Review of 14 Models and External Validation

BACKGROUND & AIMS

The aim of our study was to characterize the performance of hepatocellular carcinoma (HCC) prediction models in chronic hepatitis B (CHB) patients through meta-analysis followed by external validation.

METHODS

We performed a systematic review and meta-analysis of current literature, followed by external validation in independent multi-center cohort with 986 patients with CHB undergoing entecavir treatment (median follow-up: 4.7 years). Model performance to predict HCC within 3, 5, 7, and 10 years was assessed using area under receiver operating characteristic curve (AUROC) and calibration index. Subgroup analysis were conducted by treatment status, cirrhotic, race and baseline alanine aminotransferase.

RESULTS

We identified 14 models with 123,885 patients (5,452 HCC cases), with REACH-B, CU-HCC, GAG-HCC, PAGE-B and mPAGE-B models being broadly externally validated. Discrimination was generally acceptable for all models, with pooled AUC ranging from 0.70 (95% CI, 0.63-0.76 for REACH-B) to 0.83 (95% CI, 0.78-0.87 for REAL-B) for 3-year, 0.68 (95% CI, 0.64-0.73 for REACH-B) to 0.81 (95% CI, 0.77-0.85 for REAL-B) for 5-year and 0.70 (95% CI, 0.58-0.80 for PAGE-B) to 0.81 (95% CI, 0.78-0.84 for REAL-B and 0.77-0.86 for AASL-HCC) for 10-year prediction. However, calibration performance was poorly reported in most studies. In external validation cohort, REAL-B showed highest discrimination with 0.76 (95% CI, 0.69-0.83) and 0.75 (95% CI, 0.70-0.81) for 3 and 5-year prediction. The REAL-B model was also well calibrated in the external validation cohort (3-year Brier score 0.066). Results were consistent in subgroup analyses.

CONCLUSIONS

In a systematic review of available HCC models, the REAL-B model exhibited best discrimination and calibration.

External validation of hepatocellular carcinoma risk scores in patients with chronic hepatitis B virus infection in China

Several scores have been proposed in untreated or treated patients with chronic hepatitis B (CHB) to predict risks of hepatocellular carcinoma (HCC) occurrence. However, it is still unclear which score suits all chronic hepatitis B virus (HBV)-infected patients well, regardless of whether they are chronic carriers or CHB patients. In this study, we validated and compared the predictability of CU-HCC, REACH-B, PAGE-B and mPAGE-B in patients with chronic HBV infection in China. 1,786 patients with no history of HCC were recruited, with 978 carriers and 808 CHB patients on antiviral therapy. Patients were classified into low- and high-risk groups according to the predefined cut-off values of 5, 8, 10 and 9 for CU-HCC, REACH-B, PAGE-B and mPAGE-B. The median follow-up period was 43.7months, during which 18 (1.0%) patients developed HCC. The areas under the receiver operating characteristic curves (AUROCs) of CU-HCC, REACH-B, PAGE-B and mPAGE-B scores to predict HCC risk at 36 months were 0.815, 0.703, 0.794 and 0.825, respectively (all p < 0.05). No significant difference among AUROCs of these scores was observed except those of mPAGE-B and REACH-B at 36 months. The cumulative incidence of HCC in low- and high- risk groups based on CU-HCC, REACH-B, PAGE-B and mPAGE-B were 0.4% vs. 3.2%, 0.7% vs. 1.5%, 0.2% vs. 1.3%, and 0.2% vs. 1.7% at 36 months, respectively (all p < 0.05, except PAGE-B, log-rant test). Both CU-HCC and mPAGE-B scores accurately predict HCC risk in Chinese chronic HBV-infected patients. Patients with CU-HCC <5 or mPAGE-B <9 could be exempt from HCC surveillance within 36 months.

Assessing risk scores for predicting hepatocellular carcinoma in Thai patients with chronic hepatitis B

Chronic hepatitis B (CHB) infection-associated hepatocellular carcinoma (HCC) is a major health problem in Asian countries. Several HCC risk prediction models have been developed using either treated or untreated CHB patients. However, there is limited validation of these risk scores in a treated and untreated mixed CHB patient cohort. This study analysed and validated HCC risk scores among 2208 CHB patients who enrolled in the HCC surveillance programme in Thailand during July 2010. The baseline clinical and radiologic data of these CHB patients were applied to calculate various HCC risk scores. There were 20 patients (0.9%) with HCC development at the 5.9-year follow-up. The areas under the receiver operating characteristic curves (AUROCs) predicting HCC risk at 5 years were 0.80 (0.68-0.91), 0.73 (0.60-0.85), 0.79 (0.67-0.91), 0.70 (0.58-0.82), 0.72 (0.59-0.85), 0.76 (0.63-0.87) and 0.77 (0.64-0.89) for the GAG-HCC, CU-HCC, REACH-B, PAGE-B, mPAGE-B, CAMD and AASL scores, respectively. The overall HCC risk scores were accurate and comparable. However, the subgroup analysis revealed better HCC-risk-predictive performance in the treated patients, while performance was less helpful in those not fulfilling criteria for antiviral therapy. Clinicians should be aware of these data when using the HCC risk score in untreated CHB patients.

Predictive performance of newer Asian hepatocellular carcinoma risk scores in treated Caucasians with chronic hepatitis B

Background & Aims

Recently, several risk scores for prediction of hepatocellular carcinoma (HCC) were developed in cohorts of treated Asian patients with chronic hepatitis B (CHB), but they have not been assessed in non-Asian patients. We evaluated the predictability and comparative utility of our PAGE-B and recent Asian HCC risk scores in nucleos(t)ide analogue (NA)-treated adult Caucasian patients with CHB, with or without well-documented compensated cirrhosis but not previous diagnosis of HCC.

Methods

We included 1,951 patients treated with entecavir/tenofovir and followed up for a median of 7.6 years. The c-statistic was used to estimate the predictability of PAGE-B, HCC-Rescue, CAMD, mPAGE-B, and AASL score for HCC development within 5 or 10 years. The low- and high-risk group cut-offs were used for estimation of negative (NPV) and positive predictive values (PPV), respectively.

Results

HCC developed in 103/1,951 (5.3%) patients during the first 5 years and in another 39/1,428 (2.7%) patients between years 5 and 10. The 3-, 5-, and 10-year cumulative HCC rates were 3.3%, 5.9%, and 9.6%, respectively. All scores offered good 5- and 10-year HCC prediction (c-statistic: 0.78–0.82). NPVs were always >99% (99.3–100%), whereas PPV ranged between 13% and 24%.

Conclusions

In NA-treated Caucasian patients with CHB including compensated cirrhosis, HCC risk scores developed in NA-treated Asian patients offer good 5- and 10-year HCC predictability, similar to that of PAGE-B. PAGE-B and mPAGE-B scores are simpler in clinical practice, as they do not require an accurate diagnosis of cirrhosis, but the addition of albumin in mPAGE-B score does not seem to offer an advantage in patients with well compensated liver disease.

Lay summary

Several risk scores for prediction of hepatocellular carcinoma (HCC) were recently developed in cohorts of treated Asian patients with chronic hepatitis B (CHB). In Caucasian patients with CHB treated with oral antivirals, newer Asian HCC risk scores offer good 5- and 10-year HCC predictability, similar to that of PAGE-B. For clinical practice, PAGE-B and mPAGE-B scores are simpler, as they do not require an accurate diagnosis of cirrhosis.

•

In treated Caucasian patients with chronic hepatitis B, newer Asian hepatocellular carcinoma risk scores offer good 5- and 10-year predictability, similar to that of PAGE-B.

•

PAGE-B and mPAGE-B scores are simpler in clinical practice, as they do not require an accurate diagnosis of cirrhosis.

•

The addition of albumin in mPAGE-B does not seem to offer an advantage in patients with well-compensated liver disease.

Predictors for the progression of hepatic cirrhosis to hepatocellular carcinoma under long-term antiviral therapy

OBJECTIVE

Patients diagnosed with hepatitis B virus (HBV)-related hepatic cirrhosis have the potential for progression to hepatocellular carcinoma (HCC) even while undergoing long-term nucleos(t)ide analog (NA) therapy. This study investigated the predictors for the progression of hepatic cirrhosis to HCC under long-term NA therapy.

METHODS

This retrospective study enrolled 898 patients diagnosed with HBV-related hepatic cirrhosis. They received NA therapy between January 2012 and January 2015. The values for the liver stiffness measurement (LSM), laboratory tests, and disease history were collected. The diagnostic specificity of the serum, was assessed with a receiver operating characteristic curve.

RESULTS

The overall 2- and 3-year cumulative incidence of HCC was 6.8% and 15.15%, respectively. The LSM values were higher in the patients who had progressed to HCC. The serum PIVKA-II levels were more efficient than the serum AFP levels for the diagnosis of early HCC as the larger area under curve (0.866 vs. 0.687). The multivariate logistic regression analysis showed that HCC occurrence was significantly associated with the baseline LSM value (odds ratio = 1.035). At the end of the study, the death rate for the patients with larger LSM values was higher than that for those with lower LSM values (67.88% vs. 39.90%).

CONCLUSION

Patients with HBV-related cirrhosis have the potential for progression to HCC even under long-term NA therapy. The LSM value and the serum PIVKA-II level are significant predictors of HCC occurrence.

Clinical utility of hepatocellular carcinoma risk scores in chronic hepatitis B

BACKGROUND

Several risk scores have been recently developed to predict hepatocellular carcinoma (HCC) in chronic hepatitis B (CHB) patients. We systematically assessed the performance of the available HCC risk scores.

METHODS

Literature search was performed to identify all published studies reporting development or external validation of HCC risk scores in CHB patients.

RESULTS

Until March 2019, 12 scores were developed in untreated Asian and 7 scores in treated Asian (n = 6) or Caucasian (n = 1) patients. All scores provided significant predictions for HCC development in the derivation and validation cohorts of their original studies (c-statistic: 0.76-0.95) and usually classified patients into low, medium and high HCC risk groups. Eleven independent studies and three studies developing their own scores have validated externally some scores in Asian (GAG-HCC:5, CU-HCC:6, REACH-B:6, REACH-Bm:4, LSM-HCC:3, PAGE-B:5) or Caucasian/mixed origin patients (GAG-HCC:4, CU-HCC:4, REACH-B:4, PAGE-B:2). All scores offered acceptable predictability in almost all independent Asian cohorts (c-statistic: 0.70-0.86), but only PAGE-B and recently modified PAGE-B (mPAGE-B) offered good predictability in all independent Caucasian and/or Asian cohorts. Negative predictive values for 5-year HCC prediction were ≤99% (95%-99%) in most independent cohorts assessing Asian risk scores and 99%-100% in all independent cohorts (Caucasian/mixed origin:2; Asian:3) assessing PAGE-B and/or recently mPAGE-B.

CONCLUSIONS

Direct comparison of the newest HCC risk scores in independent patient cohorts of different origin remains intriguing, although statistical associations may not be directly transferable to clinical practice. PAGE-B and recently mPAGE-B score seem to offer persistently high predictability for Caucasian and/or Asian treated patients with low HCC risk who require no surveillance.