A system review of central nervous system tumors on children in China: epidemiology and clinical characteristics

Thus far, the Chinese Cancer Registry Annual Report has released cancer data for the period 2005 to 2017. In our study, the incidence and mortality data for childhood CNS tumors (0–19 age group) were collected from the same Report, which by 2017 had addressed approximately 31.39% of the Chinese population. ​The changes in the number of registries, and the coverage of the population from 2005 to 2017, are detailed in Table S1, and the distribution of cancer registries in China is shown in Figure S1. With the continuous improvement of economic conditions and tumor-monitoring capacity in China, the number of monitoring units is increasing, the quality of monitoring data is improving, and the coverage of the population is also rising. At present, the Chinese Cancer Registry Annual Report comprehensively reflects the development of cancer incidence and death within the country. The resource includes all individuals diagnosed with, or dying from, childhood CNS tumors (ICD-10 codes C70 ~ C72, D32 ~ D33 and D42 ~ D43) between 1 January and 31 December each year.

We extracted the population sizes and incidence/mortality cases for four age groups (0 years, 1–4 years, 5–9 years, 10–14 years, and 15–19 years). Because the incidence/mortality rate in the 0-year age group was too low for our purposes, we combined the related data in the 0-year age group with those for the 1–4-year age group. Incidence rates, mortality, and the annual percent change (APC) in each age group were calculated to capture the trend from 2005 to 2017.

This study had been approved by the Ethics Committee of Liaoning Cancer Hospital. Because the data for our study were derived from the Chinese Cancer Registry Annual Report (2005 to 2017), and the related published article, our study does not involve specific research objects, sensitive human information and data, or commercial interests, and an application for ethical approval was thus waived.

The Cancer Registry Annual Report did not provide the required information on childhood CNS-tumor positions and subtypes. Consequently, we performed a system review to analyze the locations and types of these tumors. Multiple literature databases (PubMed, Web of Science, China Science and Technology Journal Database (CQVIP), China National Knowledge Infrastructure (CNKI), and Wanfang Data) were searched via an established search strategy. This latter included (1) the key words “central nervous system tumors,” “brain tumor(s),” “brain neoplasm(s),” “spinal cord tumor(s),” “spinal cord tumor(s),” and “spinal cord neoplasm” for searches on PubMed and Web of Science. Meanwhile (2), the above terms were set as subject headings (including title, abstract, and keyword) for searches within the Chinese databases. The searches included all relevant studies published prior to June 30, 2023. All the retrieved references were restricted to “child” and “Chinese population.” We conducted this study, and we reported the related results, based on the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement, as issued in 2020 (Table S2). We also registered our review on the Open Science Framework (https://​doi.​org/​10.​17605/​OSF.​IO/​EX76Z.).

We created an electronic library with Endnote bibliographic software (version X6; Thomson Reuters, Inc., Philadelphia, PA), and all pertinent studies were retrieved from this electronic library. After duplicate studies were identified via the software, two authors (XM W, B Z) independently screened the titles/abstracts and full texts of the remaining studies. When disagreements occurred between the authors, a further author (CY W) was consulted, and agreement was reached by discussion.

The criteria for inclusion were as follows: (1) observational studies on primary tumor; (2) studies evincing a specific research time period and area; (3) a clear diagnostic criterion; (4) all cases being examined via imaging, and evincing a clear location, pathological type and grade; (5) when two or more studies came from the same population, the most recent article, or the article with the largest sample size, was selected. The criteria for exclusion were as follows: (1) studies failing to fulfil the above inclusion criteria; (2) the publication type was that of a review, conference abstract, or case report; (3) the data were not derived from the population, being derived (e.g.) from animal experiments. The International Classification of Childhood Cancer, third edition (ICCC-3), was used as a classification system to distinguish the pathological types of childhood CNS tumor.

We extracted basic information (First author, Year of publication, Study design, Time interval of case collection, Hospital of diagnosis and treatment, Source of cancer cases, Diagnostic criteria, and Number of subjects) from the included studies. To avoid the inclusion of duplicate cases, we also used three factors (Time interval of case collection, Hospital of diagnosis and treatment, Source of cancer cases) to identify the latter. The quality of the included studies was independently assessed by two authors (XM W, B Z) via a cross-sectional/prevalence study-quality assessment [6]. The detailed contents of the latter are shown in Tables S3 and S4.

The incidence and mortality of childhood CNS tumors were classified by sex, age group (0–4, 5–9, 10–14 and 15–19), and area (urban or rural). Joinpoint software (Version 4.7.0.0) was used to assess the magnitude, and direction of trends over time, for incidence and mortality. The APC was calculated by means of generalized linear models, assuming a Poisson distribution. The Monte Carlo permutation method was deployed to assess the tests of significance, and the overall asymptotic significance level was maintained via a Bonferroni correction. The overall summary estimate for each clinicopathological characteristic was calculated via a random-effects model (DerSimonian-Laird) in the meta-analysis. Heterogeneity was assessed via I-square statistics (I²) and Cochran Q-statistics. The meta-analysis was performed using R software (Version 4.1.1). P < 0.05 was defined as statistically significant.

In China, from 2005 to 2017, a total of 8216 childhood CNS tumors were diagnosed among children aged 0 to 19. There was no significant increase in the incidence rate (APC: -0.1, 95% CI: -1.5 to 1.3). In terms of the area analysis, the incidence rate in rural areas increased significantly (APC: 6.2, 95% CI: 2.4 to 10.2), while no significant change was found in urban areas (APC: -1.6, 95% CI: -3.2 to 0.2). There was a significant decrease in the 15–19-year age group (APC: -2.2, 95% CI: -4.4 to -0.8). The incidence rate in males was higher than that in females, and the incidence in urban areas was also higher than that in rural areas (Fig. 1A, C, E and Table S5).

The number of childhood CNS tumor deaths was 4577. There was a significant increase in the mortality rate (APC:1.8, 95% CI: 0.3 to 3.4). There were also significant increases in CNS mortality from 2005 to 2017, in both the 5–9-year age group and the 10–14-year age group (APC in 5–9-year age group: 3.9, 95% CI: 2.2 to 5.7; APC in 10–14-year age group: 3.3, 95% CI: 0.5 to 6.1). As with the incidence rate, the mortality rate decreased with age. In terms of sex, the mortality rate in males increased significantly (APC: 1.9, 95% CI: 0.3 to 3.5), while no significant change was found in females (APC: 1.6, 95% CI: 0.1 to 3.3), and the mortality rate in males was also higher than that in females. As regards the area analysis, the mortality rate increased significantly in both urban and rural areas (APC in urban:1.6, 95% CI: 0.2 to 3.1; APC in rural: 4.4, 95% CI: 0.4 to 8.4). The mortality rate in urban areas was higher than that in rural areas. The extent of the increase in rural areas was, however, higher than that in urban areas, and the two gradually converged from 2005 to 2017 (Figs. 1B, D and F, and Table S5).

In terms of literature collection, 1654 potential studies were searched, across multiple databases. After a review of the titles and abstracts, 1594 studies were excluded. Of the remaining 60 studies, we found that 27 studies had duplicate cases, while studies with larger sample sizes and higher quality were included. Following a careful review of the full texts, 27 studies were excluded. In total, 16 studies satisfied the inclusion criteria for our meta-analysis [7‐22]. The study-selection process, and the results from the literature search, are shown in Fig. 2. Table S6 shows the characteristics of the included studies within the meta-analysis. All the included studies evinced a cross-sectional design.

A total of 12,932 children with CNS tumors were addressed by the 16 studies. Among the total population of included children, 56.4% were males (95% CI: 51.7% to 61.0%), and the male-to-female ratio was approximately 1.28:1. Regarding the age-group analysis, the proportion of low-grade tumors increased by age, but the ratio of high-grade and low-grade tumors decreased by age (Table 1). In terms of the analysis of childhood CNS-tumor location, intracranial tumors accounted for 92.6% (95% CI: 91.3% to 93.9%), while spinal-cord tumors accounted for only 7.4% (95% CI: 6.2% to 8.7%), with a ratio of 13.2:1. The ratio of supratentorial tumors to infratentorial tumors was 1.71:1. Among supratentorial tumors, the cerebral hemisphere (25.5%, 95% CI: 21.7% to 29.4%) and sellar region (18.4%, 95% CI: 14.4% to 24.4%) were the main tumor locations, accounting for 67.7% of supratentorial tumors. Among infratentorial tumors, meanwhile, the cerebellum (20.8%, 95% CI: 16.7% to 24.8%) and the vermis of cerebellum (12.4%, 95% CI: 6.8% to 18.1%) were the main tumor locations, accounting for 75.0% of infratentorial tumors (Fig. 3 and Table S7). In the pathological types of childhood CNS tumor, the top five pathological types were astrocytomas (26.8%, 95% CI: 23.9% to 29.6%), tumors of the sellar region (craniopharyngiomas) (14.0%, 95% CI: 12.0% to 16.0%), medulloblastomas (12.8%, 95% CI: 11.4% to 14.2%), intracranial and intraspinal germ-cell tumors (7.0%, 95% CI: 5.7% to 8.3%) and pituitary adenomas and carcinomas (6.7%, 95% CI: 2.8% to 10.6%) (Fig. 4 and Table S8).