Circulating Tumor DNA As a Minimally Invasive Predictor of Early Relapse in Multiple Myeloma

Background: Achieving minimal residual disease (MRD) negativity in multiple myeloma (MM) is associated with improved progression-free (PFS) and overall survival (OS). However, current MRD assessment relies on invasive bone marrow biopsies and cannot detect extramedullary disease. This study aimed to evaluate the sensitivity of detecting circulating tumor DNA (ctDNA) in cell-free DNA (ccfDNA) using a combination of ultra-deep targeted sequencing and low-pass whole-genome sequencing (WGS) to identify patients with MM who are at high-risk of early relapse.

Methods: Serial plasma samples were collected at four-time points from 24 patients with newly diagnosed MM during enrollment in a clinical trial investigating lenalidomide maintenance therapy after induction therapy (NCT02538198). This cohort included 13 patients who achieved sustained MRD negativity and 11 patients who lost or never achieved MRD negativity during the first year of maintenance (non-sustained MRD). Cell-free DNA was extracted from 4 ml of plasma. Low-pass (10X) whole-genome sequencing (WGS) was used to assess tumor burden and detect copy number alterations (CNAs). Ultra-deep (5000X) targeted sequencing with error-correcting unique molecular identifiers (UMI) was performed on 76 recurrently mutated genes in MM to detect single nucleotide variants (SNVs) and insertions and deletions (indels).

Results: CNAs by low pass WGS were detected in ccfDNA in all patients, including those achieving sustained MRD negativity. Patients achieving sustained MRD negativity showed a significant decrease in cancer cell fractions after induction therapy, while patients with non-sustained MRD negativity showed an increase (mean change in fraction of tumor DNA: -0.006 vs. 0.05, respectively, Wilcoxon rank sum test P value: 0.0006). Detectable ctDNA (fraction of tumor DNA > 0.01) preceded clinical progression by a median of 252 days in patients who relapsed. Additionally, patients with detectable ctDNA had a lower PFS and were associated with a higher risk of disease progression after maintenance therapy compared to those with a tumor DNA fraction ≤ 0.01 (log-rank P value = 0.009). Non-synonymous SNVs in FAM46C, KRAS, NRAS, and STAT3 were exclusively detected in patients with non-sustained MRD negativity.

Conclusion: This study demonstrates that ccfDNA analysis using a combination of low-pass WGS and ultra-deep targeted sequencing can be employed to monitor tumor dynamics and predict disease progression in MM. This approach may overcome limitations associated with current MRD assessments which require bone marrow sampling and are associated with high cost, morbidity and a reduced opportunity for serial monitoring. Further, ccfDNA analysis facilitates increased sensitivity to detect both extramedullary and bone marrow relapse. The detection of CNAs and SNVs in ccfDNA may provide valuable information to inform treatment decision-making and improve patient outcomes. Validation in larger prospective studies to confirm these findings could establish the clinical utility of this approach as a new standard in MM management.