H3K27M-mutant glial tumors arise predominantly in midline structures such as the thalamus, brainstem, and spinal cord (1, 2). The H3K27M-mutant diffuse midline glioma is a relatively newly described entity, debuting in the revised 2016 World Health Organization classification of tumors of the central nervous system. These tumors are more commonly seen in pediatric patients and are associated with a poor prognosis of typically less than one year because of their infiltrative nature and difficulty in achieving complete surgical resection (3–6). H3 mutations could occur either on the HIST1H3B/C (H3.1), HIST2H3C (H3.2), or H3F3A (H3.3) genes with a lysine to methionine amino acid substitution at codon 27 (K27M) and are mutually exclusive with isocitrate dehydrogenase 1/2 (IDH1/2) mutations (7).

Despite aggressive therapeutic approaches and advances in chemoradiotherapy and targeted therapy regimens, there has been no survival improvement for patients with H3K27M mutations over the recent years (3, 8). Published studies have shown several differences in clinicopathological parameters between pediatric and adult patients with H3K27M-mutant gliomas (9, 10), but no significant differences in overall survival (11, 12). In this study, individual patient data from published studies were integrated to investigate the clinical and prognostic differences between pediatric and adult H3K27M-mutant midline gliomas.

For the title and abstract screening, we identified 756 articles, and 74 of them were selected for full-text review. After reading full text, we included 43 studies comprised of 929 H3K27M-mutant midline gliomas for integrated analyses ( Figure 1 ) (2, 4, 7, 8, 10–12, 14–44). The characteristics of all included studies are shown in Table S1 . The median age of patients was 11.0 years (range, 1-82). The majority of tumors were found intracranially with brainstem or thalamus being the most frequent sites. The median OS was 11.3 months with 80% of patients expired at last follow-up. Stratified by age, 272 and 657 patients were separated into adult and pediatric groups, respectively.

Diffuse midline gliomas are mostly high-grade tumors and have a lethal outcome. These neoplasms can occur at any age from infants to elderly patients but are most commonly found in children (35, 45, 46). About 80-90% of these tumors harbor mutations in H3 genes, most of which are H3K27M genotypes (45). H3K27M-mutant midline gliomas are associated with a worse outcome compared to H3-wild-type tumors (4). Because midline gliomas with H3K27M mutations are clinically different from H3K27M-mutated cortical high-grade gliomas (47), we only focused on midline tumors to avoid the risk of bias. Given its rarity in adults, it is challenging to investigate the differences in clinical manifestations and prognosis between pediatric and adult groups (11). In this study, we conducted an integrated analysis of more than 900 H3K27M-mutated midline gliomas to examine the clinicopathological parameters and prognosis between adults versus pediatrics. Our results showed that pediatric tumors were not only clinically and prognostically different but also had distinct molecular profiles as compared to the adult group.

In this study of H3K27M-mutant midline tumors, young age portended an adverse prognosis with shorter PFS and OS as compared with adult patients. While earlier reports suggested that OS of pediatric and adult H3K27M-mutant gliomas are not different (8.9 vs 9.3 months, respectively), only a small number of patients were compared (11). Schulte et al. (48) reported a higher survival rate of adult H3K27M midline gliomas compared to children in published case series and hypothesized that adults might have a better survival. Notably, our results revealed that pediatric patients had a higher rate of tumor resection and radiation administration than adult patients, substantiating that pediatric H3K27M-mutant midline gliomas carry worse outcome despite more aggressive treatment attempts than in adult patients. Our findings showed that tumors in pediatric patients had more aggressive pathologic features including a significantly advanced histological grade and Ki67 proliferation index, which may explain the poor survival of the pediatric cohort. Hoffmann et al. reported that pediatric DIPG with age of less than 3 and more than 10 years were associated with longer-term survival compared to those aged 3-10 (49). In this study, although the p-value of the log-rank test demonstrated significant differences between these three age subgroups, pairwise comparison showed that the differences between infants vs young children and infants vs adolescents did not reach statistical significance.

While earlier studies reported a high frequency of H3 K27M mutation in both adult and pediatric high-grade spinal gliomas (50), we found a significantly higher frequency of adult H3 K27M spinal cord tumors as compared to the pediatric population. These results were consistent with published population-level data on high-grade gliomas and glioblastomas of the spinal cord that these neoplasms are most commonly seen in adults (51, 52), as opposed to DIPGs or brainstem gliomas which are more likely to occur in the pediatric population (3, 8). Several reports indicated the aggressiveness of H3K27M-mutant diffuse midline gliomas is independent of their anatomical locations (1, 53). However, other published data reported that patient survival may be influenced by different anatomical locations (42, 54). When we stratified cases into subgroups of intracranial and spinal tumors, we still observed the same result for pediatric intracranial tumors and a similar trend for pediatric spinal gliomas, possibly due to the relatively small sample size of spinal cord gliomas. These results confirmed the uniformly short survival of H3K27M midline gliomas in the pediatric population, regardless of anatomical location.

Compared to pediatric tumors, adult H3K27M-mutant midline gliomas were also genetically different. We found that ACVR1 mutations were exclusively seen in pediatric patients, particularly in children less than six years of age. On the other hand, the incidence of FGFR1, NF1 mutations, and MGMT methylation were more significantly prevalent in adult patients. The lower frequency of MGMT hypermethylation in pediatric patients may contribute to the poor response to temozolomide and a worse PFS/OS as compared to adult cohort. FGFR1 mutations are associated with prolonged survival in H3K27M-mutant gliomas (55), which may influence the more favorable outcome in adult patients. Other genetic alterations such as EGFR alterations, TERT promoter mutations, and mutations in the tumor suppressor PTEN are less likely to exist in high-grade gliomas in children (56, 57). However, we observed an insignificant difference in the prevalence of these genetic markers between pediatric and adult H3K27M-mutated midline gliomas. Hopefully, these observations may help understand the molecular profile of midline gliomas as treatment.

This study is the first study to demonstrate the differences in demographics, clinical manifestations, prognoses, as well as molecular backgrounds of pediatric versus adult H3K27M-mutant midline gliomas. Because of the rarity of H3K27M-mutant midline gliomas in adult patients, it is challenging to observe the significant differences between pediatric and adult cohorts in institutional studies (11). Our results are of clinical interest as they may aid in refining natural history expectations and in tailoring therapeutic approaches. However, this study has certain limitations. First, we could not avoid selection bias originating from the included datasets because most of them were retrospective studies. In the multivariate Cox regression model adjusted for multiple clinical covariates, the prognostic implication of age remained significant, validating the independent role of this parameter in the risk stratification of H3K27M-mutant midline gliomas. Additionally, there are important clinical factors that we could not incorporate into the multivariate analysis due to missing data, such as the Karnofsky Performance Scale, tumor size, and disease stage. Pediatric patients may have limited self-care ability resulting in lower KPS and advanced disease stages compared with adults. Future prospective studies are needed to confirm the results of this study. Another bias could occur due to differences in treatment protocols across included institutions. Finally, although we strictly screened the included patients using patient identification code, demographic information, and clinical data to avoid duplicating patients, there are still possibilities of overlap among the dataset which may affect the analyses.

In conclusion, the present study demonstrated that pediatric H3K27M-mutant midline gliomas are demographically, clinically, prognostically distinct from adult tumors. H3K27- mutant midline gliomas in pediatric patients were associated with a significantly shorter PFS and OS. Additionally, the underlying molecular backgrounds of these two cohorts were also different. Our results may improve our biological understanding of H3K27M-mutant midline gliomas.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding author.

HV, conceptualization, data curation, formal analysis, investigation, methodology, project administration, software, validation, writing original, review, and editing. TN, HL, AJ, MH, JB, and RM-K, data curation, formal analysis, investigation, methodology, review, and editing. IFD, data curation, formal analysis, investigation, methodology, project administration, validation, review, editing, and supervision. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.