ID #581 Confounding Factors in CSF-derived cfDNA: The Seattle Experience

Background

Analysis of CSF-derived cell-free DNA (cfDNA) is a promising approach in evaluating pediatric brain tumors, with implications in diagnosis, disease-monitoring, and treatment. Its utilization in patient care requires correlation with clinical findings, as test limitations and confounding factors can cause false-negative or false-positive results. Herein we highlight cfDNA alterations with the potential to impact test interpretation.

Methods

Extracted cfDNA from CSF samples collected from brain tumor patients who consented to research was evaluated using a targeted next-generation sequencing (NGS) panel and/or low-pass whole genome sequencing (LP-WGS), 88 and 285 samples respectively. Results were discussed at multidisciplinary tumor boards.

Results

Patient 1: A 9-year-old diagnosed radiologically with a diffuse infiltrative pontine glioma lacked H3K27M-positive cells upon biopsy. NGS of cfDNA, following CAR-T cell therapy, revealed copy gain of EGFR exons 9-17 (partial) and deletion of introns 9-16 consistent with the EGFR “kill-switch” engineered into CAR-T cells rather than a tumor-derived alteration.

Patient 2: An 8-year-old with minute biopsies of spinal neoplasms histologically consistent with pilocytic astrocytomas. LP-WGS of cfDNA identified gain of part of chromosome 3p and loss of part of chromosome 7p, concerning for a higher-grade neoplasm. Multidisciplinary review revealed these findings were not tumor-derived but stemmed from the patient’s known germline translocation t(3;7).

Patient 3: A 14-month-old with a pineal region mass had a differential diagnosis that included a germ cell neoplasm. LP-WGS of cfDNA demonstrated gain of 12p, a finding characteristic of germ cell tumors. In the absence of additionally identified copy alterations, the patient’s germline trisomy 12p could have produced a false-positive result.

Conclusions

As we continue to push the frontiers in cfDNA testing, it is important to consider confounding factors when interpreting results. Indeed, clinical correlation within our multidisciplinary tumor board was key to a correct understanding of our patients’ biology.