DATA SOURCES AND SEARCHES

We conducted a search for literature published in Medline, PsycInfo, the Cochrane Central Register of Controlled Trials, and the Cochrane Database of Systematic Reviews from database inception to March 2013. The search strategy included terms for HCC, screening/screening, treatment modalities, and adverse effects including psychological harms of screening/screening. The detailed search strategy is provided in Appendix A. We obtained additional articles from systematic reviews, reference lists of pertinent studies, reviews, editorials, and by consulting experts. We also searched for ongoing and recently completed studies on ClinicalTrials.gov.

STUDY SELECTION

This review was commissioned by the VHA Oncology Program Office and the VHA HIV, Hepatitis and Public Health Pathogen Program. A protocol describing the review plan was posted to a publicly accessible website before the study was initiated.¹⁶ The analytic framework and key questions which guided this review were developed in conjunction with a panel of VA and non-VA technical experts and are provided in Appendix B, Figure 1. Detailed inclusion and exclusion criteria are provided in Appendix C. We used a “best-evidence” approach to guide study design criteria depending on the question under consideration and the literature available.¹⁷ We prioritized controlled clinical trials, then comparative observational studies, then large cohort studies. To assess the effects of screening on HCC-specific and all-cause mortality, we included clinical trials and observational studies providing primary data in adult populations. We use the term “screening” to refer to any program in which tests – including ultrasonography, computed tomography, magnetic resonance imaging, and/or alpha-fetoprotein levels – were done explicitly to look for HCC in asymptomatic patients. Studies had to include a contemporary comparison group of patients who did not undergo screening and had testing done only to evaluate symptoms. We excluded observational studies that did not account for basic confounding factors such as age, sex, and liver disease severity. Because we anticipated few clinical trials comparing screening to no screening, we also included trials comparing different frequencies of screening. We included studies of any population with chronic liver disease, with or without cirrhosis, but excluded studies of patients with prior HCC. To assess the harms of screening, we abstracted any reported adverse effects data from studies included from the above search. We also additionally searched for trials or observational studies focused on potential harms of HCC screening.

To assess the benefits and harms of treating HCC found as a result of screening, we included trials or large prospective cohort studies examining the effects of liver resection, transplant, radiofrequency ablation, transarterial chemoembolization, or sorafenib, compared to no treatment in patients with early stage HCC (defined as the equivalent of Barcelona Clinic Liver Cancer (BCLC) Stage A, or early-stage HCC by the Milan criteria).^18,19 We included studies with mixed populations of patients with early and advanced disease, but not studies including only patients with advanced disease. Because comparative effectiveness studies would not directly address the incremental effects of screening or treating screen-detected disease, we excluded studies comparing 2 or more active treatments without an untreated control group. We found no trials and only a small number of comparative observational studies of liver resection, transplantation, and radiofrequency ablation, so we included non-comparative cohort studies of these interventions if they included consecutive patients with adequate long-term follow-up and, in the case of OLT for which several large cohorts were available, large sample size (n > 500). We prioritized systematic reviews of such studies if available.

In order to better understand the quality and content of existing recommendations guiding the practice of HCC screening, we systematically searched for HCC screening guidelines. Among published guidelines, we identified the 3 most widely disseminated guidelines representing distinct geographic areas including North America,⁸ Europe,⁹ and Asia.¹⁰

Seven investigators reviewed the titles and abstracts of citations identified from literature searches, and 2 reviewers independently assessed the selected full-text articles for inclusion based on the eligibility criteria shown in Appendix C. Disagreements were resolved through consensus.

DATA EXTRACTION AND QUALITY ASSESSMENT

From each study, we abstracted study design, objectives, setting, population characteristics (including sex, age, race/ethnicity, liver disease etiology and severity), subject eligibility and exclusion criteria, number of subjects, years of enrollment, mode and frequency of screening, adjusted and unadjusted mortality, and adverse events. A second author checked each entry for accuracy.

Two reviewers independently assessed the quality of each trial using a tool developed by the Cochrane Collaboration.²⁰ Disagreements were resolved through discussion. Each study was given an overall summary assessment of low, high, or unclear risk of bias. We graded the strength of evidence for outcomes using published criteria which consider the consistency, coherence, and applicability of a body of evidence, as well as the internal validity of individual studies.²¹

Though there is no widely accepted standard for quality assessment of observational studies, we adapted existing assessment tools.^22,23 For the observational screening studies, we additionally adapted causal inference criteria²⁴ relevant to this review and specifically assessed: 1) methods for ascertaining screening status, and 2) use of an inception cohort. We do not report an overall summary assessment for observational studies because there are no validated criteria for doing so.

We assessed the quality of published guidelines using the AGREE framework.²⁵

DATA SYNTHESIS AND ANALYSIS

We did not perform meta-analyses of screening or treatment interventions because of the dearth of trial data and the clinical heterogeneity among the small number of trials. Rather, we qualitatively synthesized the results of trials and observational studies.