Pediatric ependymoma in the molecular era: real-world experience with diagnostic correlation and long-term clinical outcomes.
Summary
This retrospective study evaluated the diagnostic and prognostic impact of DNA methylation array (MA) profiling in pediatric ependymoma. Out of 63 patients, MA profiling reclassified nearly a quarter of initial diagnoses, with a higher rate before 2019, underscoring the importance of integrated molecular diagnosis. Posterior fossa group A (PFA) ependymoma was the predominant subgroup. Gross total resection was the only factor significantly associated with improved survival, while chromosome 1q gain, exclusively found in PFA patients, was linked to a high relapse rate.
Analysis
CLINIQUE: This study highlights the significant clinical impact of DNA methylation array profiling for pediatric ependymoma diagnosis, leading to a reclassification of 23.8% of cases, with a particularly high rate (36.1%) in the pre-2019 cohort. This justifies the systematic integration of molecular diagnosis for all pediatric ependymoma cases, as it enables more precise classification and potentially better risk stratification. Gross total resection is confirmed as a positive prognostic factor (OS 75% vs 40%, p = 0.017), while chromosome 1q gain is identified as a poor prognostic biomarker, associated with a high relapse rate despite standard therapy. These findings could influence future treatment guidelines and clinical trial design for PFA patients with 1q gain, suggesting a clinical impact within 3-5 years. BIOMOL: The main discovery is the effectiveness of DNA methylation array profiling for molecular classification of pediatric ependymomas, enabling significant reclassification and identification of subgroups like PFA. The technology utilized includes DNA methylation profiling, RNA sequencing, and copy number variation analysis, providing a robust multi-omic approach. The identification of chromosome 1q gain as a high-risk marker, exclusively in PFA patients, represents an important translational discovery. While specific analytical validity is not detailed, the impact on diagnostic reclassification demonstrates high specificity and sensitivity in distinguishing tumor entities. These findings validate the use of these techniques for advanced clinical testing and patient stratification. BIOINFO: The study utilized t-SNE (t-distributed Stochastic Neighbor Embedding) analysis for data visualization and clustering, which is a standard but effective application for exploring high-dimensional datasets such as methylation profiles. The integration of these bioinformatic analyses with clinical and histological data led to a significant improvement in diagnostic accuracy. While the study does not describe a novel algorithm, it demonstrates the successful clinical application of existing bioinformatic tools to refine pediatric brain tumor classification. Clinical deployment of such bioinformatic workflows could standardize and automate molecular classification, replacing or complementing manual review and prioritizing patients for multidisciplinary tumor board discussions or specific clinical trials.