Individualized DNA junctions for glioma response assessment via longitudinal liquid biopsies

Grade 4 astrocytomas, including glioblastoma, are aggressive primary brain tumors that invariably recur despite maximal surgical resection and chemoradiation. Current monitoring strategies rely on magnetic resonance imaging (MRI) using RANO 2.0 criteria, which lacks sensitivity and specificity for detecting early or subtle changes in disease burden. Repeated tissue biopsies are not feasible, leaving clinicians without reliable real- time biomarkers to guide patient care or assess novel therapies. DNA junctions are unique genomic breakpoints formed by tumor-specific rearrangements that represent a promising new class of individualized liquid biopsy biomarkers. Individualized assays can be designed for each patient’s junctions, allowing precise detection and monitoring of tumor burden. We have demonstrated that high copy amplified junctions are detectable in plasma from patients with high-grade gliomas, with abundance correlating to disease burden and treatment response. However, plasma monitoring is limited to patients with amplifications (~50%). In contrast, low-copy junctions are essentially ubiquitous across gliomas and are reliably detectable in cerebrospinal fluid (CSF). The increasing use of intraventricular CSF access devices in neuro-oncology provides a unique opportunity for longitudinal junction-based monitoring. We hypothesize that longitudinal CSF sampling can enable sensitive, individualized monitoring of glioma burden via DNA junctions. In Aim 1, we will determine the clinical utility of patient-specific DNA junctions in CSF for monitoring glioma during standard-of-care treatment. We will assess their accuracy in detecting decreased disease burden with cytoreductive therapy, their stability during surveillance of stable disease, and their ability to discriminate true progression from pseudoprogression. In Aim 2, we will compare plasma and CSF as complementary sources for junction-based monitoring. We will systematically evaluate baseline detection of low- versus high-copy junctions, analyze paired plasma-CSF samples for their ability to track disease burden, and define the contexts in which each fluid provides unique or overlapping value. In Aim 3, we will evaluate the accuracy of DNA junctions for disease monitoring in patients receiving anti-B7H3 CAR-T therapy. We will test the performance of plasma-amplified junctions and directly compare serial CSF and plasma junction dynamics to clinical outcomes on trial. Paired CSF and plasma will be collected longitudinally at every MRI under existing access protocols and clinical trials. Junction dynamics will be compared to imaging and clinical outcomes, with rigorous assessment of cytoreduction, stability, progression, and pseudoprogression. By integrating both plasma and CSF, we will define the contexts in which each fluid provides optimal monitoring value. This work directly addresses a critical unmet need for accurate and minimally invasive biomarkers in gliomas by leveraging patient-specific genomics, ongoing CSF collection protocols, and our upcoming CAR-T clinical trials. Successful completion will establish CSF DNA junctions as a robust liquid biopsy tool to inform standard-of-care response assessment during each patient’s disease course.