What question did this study set out to answer?

To assess the effectiveness of cerebrospinal fluid liquid biopsy in detecting ctDNA in pediatric CNS tumor patients.

June 26, 2026Open Access

ID #648 Cerebrospinal fluid liquid biopsy: results following implementation into clinical practice for patients with pediatric central nervous system tumors

Key Points

To assess the effectiveness of cerebrospinal fluid liquid biopsy in detecting ctDNA in pediatric CNS tumor patients.
Implemented LBSeq4Kids with 123 cerebrospinal fluid samples from 48 patients, collection via lumbar puncture, Ommaya, and other methods.
Performed low-pass whole genome sequencing to analyze ctDNA levels at various clinical timepoints.
Conducted serial sampling in half the cohort for longitudinal molecular monitoring of tumor evolution.
61% of CSF samples had detectable ctDNA, with variations by clinical timepoint: 64% at diagnosis and 44% at end-of-therapy.
65% of liquid biopsy results differed from primary tumor sequencing, detecting additional variants in 35% of cases.
In patients with positive cytology, 87% had detectable ctDNA, indicating potential for improved monitoring and management of disease.

Abstract

Abstract Liquid biopsy (LB) is emerging in the care of pediatric central nervous system (CNS) tumors. Low-pass whole genome sequencing of cerebrospinal fluid (CSF) was validated as a clinical test (LBSeq4Kids) at our institution in 2022. We report 123 CSF LBs from 48 patients (ages 7 months to 20 years) with CNS tumors at timepoints ranging from diagnosis through therapy to radiographic recurrence. Histologic diagnoses included medulloblastoma (n = 24), other embryonal tumor (n = 8), glioma (n = 9), others (n = 7). Collection utilized lumbar puncture (n = 73), Ommaya (n = 30), ventricular shunt (n = 10), ventriculostomy (n = 4), and other (n = 6). Twenty-four patients (50%) underwent serial sampling (mean: 4.2; range 2–10+ samples), enabling longitudinal molecular monitoring. Circulating tumor DNA (ctDNA) was detected in 72/123 samples (59%), variable by timepoint: 14/22 (64%) at diagnosis, 8/18 (44%) at end-of-therapy (EOT), and 9/13 (69%) at radiographic recurrence. Compared to primary tumor sequencing, 31/48 (65%) LB results were different from primary tumors, with 17/48 (35%) detecting additional variants. In cases with serial sampling 35/63 (56%) of LBs demonstrated clonal evolution as evidenced by accrual of additional alterations. In patients with positive CSF cytology, 13/15 (87%) had detectable ctDNA and in patients with negative cytology 54/98 (55%) had positive ctDNA. In patients with radiographic disease at the time of LB, 44/60 (73%) had detectable ctDNA and in patients without radiographic disease, 25/56 (45%) had ctDNA present. Outcomes for 7 patients with positive ctDNA at EOT varied: 2 has disease recurrence and died of disease, 3 received continuation therapy without recurrence, and 2 remain recurrence-free off therapy (8 and 14 months off therapy). LBSeq4Kids is feasible across various collection methods, demonstrates higher sensitivity than cytology, and provides insights into tumor evolution and sub-radiographic disease.

Read Full Paperexternally

Ask AI

Helpful

Bookmark

View Full Paper