Dataset Information

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

ABSTRACT:

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.

SUBMITTER: Papatheodoridis GV

PROVIDER: S-EPMC8144729 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:Prevalence of chronic hepatitis C (CH-C) infection in patients of Asian ancestry ranges between 1% and 20%. Interferon (IFN)- and ribavirin (RBV)-containing regimens for CH-C have a negative impact on patient-reported outcomes (PROs) during treatment. The aim of this study was to assess the impact of IFN-free RBV-free sofosbuvir (SOF)-based regimens on PROs in CH-C patients of Asian ancestry. In this observational retrospective study, the PRO data from 12 multicenter multinational phase 3 clinical trials (2012-2015, conducted in Europe, North America, Australia, and New Zealand) of SOF-based regimens with and without IFN, ledipasvir (LDV), and/or RBV were used. At baseline, during treatment, and post-treatment, patients completed 4 validated PRO questionnaires (SF-36, CLDQ-HCV, FACIT-F, and WPAI:SHP). The resulting PROs in Asian patients were compared across the treatment regimens. Of 4485 of the trials' participants, 106 patients were of Asian ancestry (55.7% male, 69.8% treatment-naïve, 17.0% cirrhotic). In comparison with other patients, the Asian CH-C cohort was younger, had lower BMI, and lower rates of pre-treatment psychiatric comorbidities (anxiety, depression, sleep disorders) (all P < .05). At baseline, Asian patients also had lower SF-36 physical functioning scores (on average, by -5.6% on a normalized 0-100% PRO scale, P = .001). During treatment, Asian CH-C patients experienced a decline in their PRO scores while receiving IFN and/or RBV-containing regimens (up to -19.6%, P < .001). In contrast, patients receiving LDV/SOF experienced no PRO decrement and improvement of some PRO scores during treatment (+9.0% in general health of SF-36, P = .03). After achieving SVR-12, some of the PRO scores in Asian patients improved regardless of the regimen (up to +9.3%, P < .001). In multivariate analysis of Asian patients, the use of LDV/SOF was independently associated with higher PRO scores during and soon after the end of treatment (betas +15.0% to +29.3%, all P < .05). Other predictors of PRO impairment included depression, type 2 diabetes mellitus, and cirrhosis. The use of IFN- and RBV-free LDV/SOF regimens leads to PRO improvement in Asian patients with CH-C during treatment. Achieving SVR-12 results in improvement of PRO scores.

Project description:Background & aimsAccurate hepatocellular carcinoma (HCC) risk prediction facilitates appropriate surveillance strategy and reduces cancer mortality. We aimed to derive and validate novel machine learning models to predict HCC in a territory-wide cohort of patients with chronic viral hepatitis (CVH) using data from the Hospital Authority Data Collaboration Lab (HADCL).MethodsThis was a territory-wide, retrospective, observational, cohort study of patients with CVH in Hong Kong in 2000-2018 identified from HADCL based on viral markers, diagnosis codes, and antiviral treatment for chronic hepatitis B and/or C. The cohort was randomly split into training and validation cohorts in a 7:3 ratio. Five popular machine learning methods, namely, logistic regression, ridge regression, AdaBoost, decision tree, and random forest, were performed and compared to find the best prediction model.ResultsA total of 124,006 patients with CVH with complete data were included to build the models. In the training cohort (n = 86,804; 6,821 HCC), ridge regression (area under the receiver operating characteristic curve [AUROC] 0.842), decision tree (0.952), and random forest (0.992) performed the best. In the validation cohort (n = 37,202; 2,875 HCC), ridge regression (AUROC 0.844) and random forest (0.837) maintained their accuracy, which was significantly higher than those of HCC risk scores: CU-HCC (0.672), GAG-HCC (0.745), REACH-B (0.671), PAGE-B (0.748), and REAL-B (0.712) scores. The low cut-off (0.07) of HCC ridge score (HCC-RS) achieved 90.0% sensitivity and 98.6% negative predictive value (NPV) in the validation cohort. The high cut-off (0.15) of HCC-RS achieved high specificity (90.0%) and NPV (95.6%); 31.1% of patients remained indeterminate.ConclusionsHCC-RS from the ridge regression machine learning model accurately predicted HCC in patients with CVH. These machine learning models may be developed as built-in functional keys or calculators in electronic health systems to reduce cancer mortality.Lay summaryNovel machine learning models generated accurate risk scores for hepatocellular carcinoma (HCC) in patients with chronic viral hepatitis. HCC ridge score was consistently more accurate than existing HCC risk scores. These models may be incorporated into electronic medical health systems to develop appropriate cancer surveillance strategies and reduce cancer death.