Introduction

Spatial genomic heterogeneity is a well-recognized characteristic of multiple myeloma (MM), contributing to treatment resistance and disease progression. However, assessing this heterogeneity through multiple-site bone marrow biopsies is clinically challenging due to the invasive nature of the procedure. Liquid biopsy using circulating cell-free DNA (cfDNA) has emerged as a promising non-invasive alternative to assess the mutational landscape of MM. In this study, we investigated whether cfDNA reflects spatial heterogeneity by comparing mutational profiles between plasma and bone marrow samples in newly diagnosed MM (NDMM) patients with multiple focal lesions and diffuse bone marrow uptake on positron emission tomography/computed tomography (PET/CT).

Methods

Using a customized panel of 80 genes relevant to multiple myeloma biology, targeted sequencing was conducted on matched bone marrow and plasma samples. Variant allele frequency thresholds were set sample type-specifically: >1% for plasma to capture subclonal mutations and >5% for bone marrow to focus on clonal drivers. 18F-fluorodeoxyglucose PET/CT images were interpreted based on the number of focal lesions (FLs) and the degree of diffuse bone marrow uptake (DL). Metabolic uptake on PET/CT was evaluated using the Deauville score, with uptake greater than that of the liver (Deauville score ≥ 4) interpreted as positive.

Results

A total of 54 NDMM patients treated at Kyungpook National University Hospital between 2020 and 2023 were included in the study. The median age at diagnosis was 66.5 years, and 33 patients (61.1%) were male. Among them, 46 patients received a bortezomib-based triplet regimen, while 7 received a doublet therapy. On PET/CT imaging, 24 patients were classified as having high focal lesions (FL_high, defined as >3 focal lesions), and 20 patients showed high diffuse bone marrow uptake (DL_high, defined as Deauville score ≥ 4).

The average number of mutations per patient was significantly higher in plasma samples compared to bone marrow samples (3.76 vs. 2.52, p < 0.001). The overall concordance of mutations between bone marrow and plasma was 96.5%. The most frequently mutated gene in both samples was ATRX (BM: 61.1%, plasma: 59.3%), followed by NRAS (14.8%) in bone marrow and KMT2C (24.1%) in plasma. The genes with the lowest concordance between bone marrow and plasma were KMT2C (75.9%), RSP3A (81.5%), DNMT3A (87.0%), KRAS (90.7%), and NRAS (90.7%). Except for NRAS, these mutations were more frequently observed in plasma samples than in bone marrow.

In subgroup analysis by PET/CT findings, patients with FL_high showed a higher number of mutations than those with FL_low in both samples (BM: 2.71 vs. 2.40; plasma: 4.33 vs. 3.30, respectively), although statistical significance was not identified. Concordance of mutations between bone marrow and plasma was slightly lower in FL_high (96.46%) compared to FL_low (96.61%). In contrast, the average number of mutations was slightly lower in patients with DL_high than in those with DL_low (BM: 2.45 vs. 2.56; plasma. BM: 3.75 vs 3.76), without statistical significance. KMT2C mutations were more frequently detected in plasma samples of both FL_high and DL_high groups, with detection rates of 41.7% and 40.0%, respectively. Patients with the presence of KMT2C mutations in plasma showed inferior overall survival (median not reached vs. 26.3 months, p = 0.060).

Conclusions

While overall concordance between bone marrow and plasma samples was high, more mutations were detected in plasma. In patients with FL_high on PET/CT, plasma samples showed a higher number and broader diversity of mutations, suggesting that cfDNA analysis may provide a non-invasive alternative to biopsy for characterizing spatial genomic heterogeneity. KMT2C mutations were more frequently detected in plasma with FL_high and DL_high and were associated with poorer overall survival, highlighting the prognostic value of cfDNA-based genomic profiling.

This content is only available as a PDF.
2025
Sign in via your Institution