PATH-02. Molecular characterization of brain tumors from stereotactic needle biopsy surgical waste

Background Stereotactic needle biopsies of brain tumors are often performed in midline or deep-seeded brain regions and are risky procedures, with additional samples for molecular diagnostics conferring additional morbidity risk. To address this issue, we investigated whether tumor DNA recovered from surgical waste (specifically material washed from biopsy needles and absorbed into non-adherent gauze) can be used for standard molecular diagnostics and more novel envisioned rapid diagnostics. Methods Surgeons were instructed to save needle aspiration and/or rinse fluid used during stereotactic needle biopsies, and/or any non-adherent gauze used to examine and stage biopsy tissue for clinical analysis. Cell-free DNA was then extracted and sequenced using a variety of methods (NGS oncopanel, Nanopore targeted and methylation-aware sequencing). Results Almost all samples (96.9%; n = 31/32) yielded enough DNA for targeted sequencing and 100% of analyzable samples yielded at least one clinically confirmed somatic driver mutation. Tumor purity correlated with that of clinically indicated biopsy samples (Pearson R2 = 0.75). Waste sequencing improved sensitivity of immunohistochemical staining for IDH mutation testing (100% vs. 62.5%; n = 5/8) due to the prevalence of untested mutant IDH alleles. Nanopore methylation sequencing resulted in highly-confident and correct machine-­learning-based classifications (n = 2, K27M altered diffuse glioma and IDH mutant astrocytoma). “Ultra-rapid” H3K27M determination returned a confident H3F3A K27M variant call within 65 minutes of sample receipt, supporting the idea of future intra-operative sequencing complementing frozen section. Finally, a tumor organoid was generated from biopsy waste from a recurrent IDH-mutant tumor, allowing experimental in vitro drug screen. Conclusions High-purity tumor material from stereotactic biopsy surgical waste can be used to molecularly characterize brain tumors, reduce required number of biopsy cores for molecular diagnostics, and perform “ultra-rapid” and organoid workflows. Further, we present an optimized protocol for collection/banking of surgical waste material that can be implemented at academic and community hospitals.