Key points are not available for this paper at this time.
Abstract BACKGROUND Aggressive brain cancers represent the most lethal forms of tumors in pediatric population. Their molecular drivers and co-alterations are currently guiding therapeutic strategies. Due to heterogeneous nature of these tumors, they are quickly developing resistance to treatment. Recognizing intrinsic and extrinsic mechanisms underlying therapeutic resistance, international scientific community has recognized an urgent need to create precise models of brain tumors capable of predicting therapeutic responses. So far, patient-derived cell lines (PDCL) remain the most reliable model to estimate and mimic evolving scenarios. The primary challenges lie in derivation methods and in cellular and spatial validation. METHODS In the context of the PEDIAMODECAN program, we collected 105 fresh brain tumor samples, including 34 LGG, 24 HGG, 19 ependymomas, 19 medulloblastomas, and 9 other rare entities, to concurrently generate PDCLs and patient-derived xenografts (PDX). RESULTS Regarding the derivation process itself, we successfully maintained 80% of samples in 2D and 54% in PDX models. As anticipated, the LGG group showed the higher percentage of PDX failure (90%). Nevertheless, for high-grade subtypes, PDX generation rate increased to 85,7%. No correlation was found between derivation success and a shorter interval between surgical removal and dissociation processes. Histological analysis of PDXs models confirmed the similar microscopic features to those of primary tumors. Next-Generation Sequencing (NGS), methylome assays and spatial transcriptomic analyses of derived models matched those of the original patient tumors. Epigenetic approaches were proved to be the most effective methods for validating models, in addition to spatial transcriptomic analyses that highlighted similar inter-tumor heterogeneity between PDX models and primary tumor counterparts. CONCLUSIONS Establishing a systematic PDCL/PDX biobank for all pediatric brain tumors represents one of the most effective approaches to develop models capable of accurately reproducing tumoral heterogeneity, histological and molecular diversity, and facilitating the design predictive screenings for targeted therapies combined with irradiation.
Deschuyter et al. (Tue,) studied this question.