Global trends in incidence rates of childhood liver cancers: A systematic review and meta-analysis

Background Childhood liver cancers are relatively rare, hence inferences on incidence trends over time are limited by lack of precision in most studies. Objective To conduct a systematic review and meta-analysis of published contemporary trends on childhood liver cancer incidence rates worldwide. Data sources PubMed, EMBASE, CINAHL, Web of Science. Study selection and data extraction English-language peer-reviewed articles published from 1 January 2008 to 1 December 2019 that presented quantitative estimates of incidence trends for childhood liver cancer and diagnostic subgroups. Review was conducted per PRISMA guidelines. Two authors independently extracted data and critically assessed studies. Synthesis Random effects meta-analysis models were used to estimate pooled incidence trends by diagnostic subgroups. Heterogeneity was measured using the Q and I-2 statistics and publication bias evaluated using Egger's test. Results Eighteen studies were included, all based on population-based cancer registries. Trends were reported on average for 18 years. Overall pooled estimates of the annual percentage change (APC) were 1.4 (95% confidence interval [CI] 0.5, 2.3) for childhood liver cancers, 2.8 (95% CI 1.8, 3.8) for hepatoblastoma and -3.0 (95% CI -11.0, 4.9) for hepatocellular carcinoma. Sub-group analysis by region indicated increasing trends for childhood liver cancers in North America/Europe/Australia (APC 1.7, 95% CI 0.7, 2.8) whereas corresponding trends were stable in Asia (APC 1.4, 95%CI -0.3, 2.7). Publication bias was not detected for any of these analyses. The I-2 statistic indicated that the heterogeneity among included studies was low for combined liver cancers, moderate for hepatoblastoma and high for hepatocellular carcinoma. Conclusions Incidence is increasing for childhood liver cancers and the most commonly diagnosed subgroup hepatoblastoma. Lack of knowledge of the etiology of childhood liver cancers limited the ability to understand the reasons for observed incidence trends. This review highlighted the need for ongoing monitoring of incidence trends and etiological studies.