Abstract
Abstract
Background
Detection of circulating tumor DNA (ctDNA) in the blood and cerebrospinal fluid (CSF) of pediatric patients with central nervous system (CNS) tumors may improve diagnostic and disease monitoring capabilities. However, factors such as the blood brain barrier complicate the ability to reliably detect ctDNA in the blood. Accurately distinguishing tumor-derived DNA from germline DNA carries unique challenges.
Methods
From 2021–2024, longitudinal blood and/or CSF samples were collected from pediatric CNS tumor patients throughout the course of their treatment. Demographic information, histologic diagnoses, molecular features, and treatment responses were tracked. DNA was extracted from plasma, CSF, tumor, and whole blood samples. Using a novel classification algorithm combining whole genome and targeted genomic sequencing, DNA was analyzed for tissue of origin.
Results
Forty-five patients were enrolled with 181 total plasma samples and 41 CSF samples. Sixteen patients were selected to undergo proof-of-principle analysis consisting of six (37%) patients with diffuse midline glioma, five (31%) with high-grade glioma, four (25%) with medulloblastoma, and one (6%) with ependymoma. Five (31%) patients had metastatic disease at diagnosis. All patients in the cohort underwent radiation and 12 (75%) received chemotherapy or immunotherapy. The average number of plasma samples per patient was 4.37 (range 1-8) and CSF samples was 0.9 (range 0-3). Cell-free DNA (cfDNA) was successfully detected in 70/70 (100%) plasma samples and 10/15 (66%) CSF samples, including some patients without radiographic evidence of disease. Further analysis to confirm whether this isolated cfDNA is tumor derived is ongoing and will be presented.
Conclusion
Early data demonstrate our classification algorithm is highly successful at detecting cfDNA from the blood and CSF of pediatric CNS tumor patients. Further analysis is needed to validate the tissue of origin of this circulating DNA and to determine the clinical applicability of our novel classification algorithm.