Integrated cell-free DNA and RNA sequencing recapitulates bone marrow genomic and transcriptomic alterations in multiple myeloma

Abstract

Abstract Background: Molecular profiling of multiple myeloma (MM) is typically performed on CD138⁺ plasma cells enriched from bone marrow aspirates. While informative, this approach has significant limitations, including sampling a single spatial site, missing subclonal populations due to tumor heterogeneity, and requiring an invasive, often painful procedure that limits longitudinal monitoring. Circulating tumor-derived nucleic acids in peripheral blood offer a promising minimally invasive alternative. We hypothesized that integrated sequencing of both cell-free DNA (cfDNA) and cell-free RNA (cfRNA) from plasma could recapitulate key genomic and transcriptomic features observed in bone marrow-based profiling, while also detecting alterations originating from spatially distinct or inaccessible disease sites. Methods: Bone marrow aspirates and peripheral blood plasma were collected from 19 patients with newly diagnosed MM prior to induction therapy. Comprehensive genomic and transcriptomic profiling was performed by Genomic Testing Cooperative using CLIA-certified next-generation sequencing on both cellular and cell-free nucleic acids. From bone marrow, DNA and RNA were extracted from CD138⁺ enriched cells. From blood, cfDNA and cfRNA were isolated from plasma collected in EDTA tubes. Both specimen types were assayed using targeted DNA and RNA sequencing panels covering 302 DNA and >1600 RNA genes relevant to hematologic malignancies, including MM. DNA sequencing captured all coding exons and flanking intronic regions with a sensitivity of 1% variant allele frequency for single-nucleotide variants (SNVs) and small insertions and deletions in bone marrow, and as low as 0.1–0.5% variant allele frequency for hotspot mutations in cfDNA and cfRNA. RNA sequencing employed hybrid capture to detect gene fusions, transcript variants, expression outliers, and B and T cell immune repertoire metrics. DRAGEN (Illumina) was used for variant calling, complemented by BCFtools and VarDict. Unique molecular identifiers were used for error correction, and duplicates were removed prior to analysis. All called variants were compared to databases generated from a pool of normal samples established at the validation of the assay to remove noise. Variants were compared to those from the same run, which typically included 80 to 100 samples, and checked against ClinVar, COSMIC, dbSNP, and our local database for pathogenic entries. All identified variants were inspected and confirmed within the BAM file, with a minimum of four reads in both directions required to confirm a call. Copy number alterations, translocations, and somatic mutations were identified and annotated. Plasma cell burden was inferred from mRNA expression of CD138 and BCMA. Bone marrow plasma cells were enumerated by CD138 immunohistochemistry and multi-parameter flow cytometry. Circulating plasma cells were enumerated by the Menarini CellTracks Analyzer II and multi-parameter flow cytometry. Results: Matched bone marrow and peripheral blood samples from 19 newly diagnosed MM patients were profiled using integrated DNA and RNA sequencing. The median number of coding SNVs detected in the bone marrow was 7 per sample (range, 2-12). Circulating cfDNA or cfRNA assays captured a high proportion of these marrow events, with a median concordance to bone marrow DNA and RNA of 85.7% (range, 50-100%). Within each tissue type, concordance between matched DNA and RNA assays was also high. In the bone marrow, a median of 80% (range, 33.3-100%) of DNA-detected events were also observed in RNA. In peripheral blood, a median of 33.3% (range, 0-100%) of cfDNA-detected events were supported by cfRNA. Circulating myeloma cells were detected in all patients, with counts ranging from 414 to >50,000 cells per 4 mL of blood, and their proportion was highly concordant with bone marrow plasma cells by immunohistochemistry (Pearson correlation = 0.70). Conclusion: Our findings demonstrate that integrated sequencing of plasma-derived cfDNA and cfRNA effectively recapitulates the key genomic and transcriptomic alterations identified in matched bone marrow samples. This high degree of concordance validates the use of liquid biopsy as a reliable, minimally invasive method for molecularly characterizing MM at diagnosis. Ultimately, this multiomic approach holds significant promise for improving longitudinal monitoring of disease evolution, therapeutic response, and detecting mechanisms of treatment resistance.