Comparison of next-generation sequencing (NGS) and next-generation flow (NGF) for minimal residual disease (MRD) assessment in multiple myeloma

Multiple myeloma (MM) is a clonal plasma-cell malignancy in which residual disease can persist below the detection limit of conventional methods, leading to relapse even among patients achieving deep responses (CR/sCR). The International Myeloma Working Group (IMWG) has defined MRD negativity as the absence of clonal malignant plasma cells at a sensitivity of at least 10^-5, necessitating highly sensitive and standardized assays. Two principal standardized approaches exist for bone marrow MRD in MM: immunophenotyping by next-generation flow (NGF, EuroFlow) and molecular techniques such as next-generation sequencing (NGS). NGF achieves sensitivities down to approximately 2 × 10^-6 but requires high-quality samples and expertise; NGS targets immunoglobulin gene rearrangements and has rapidly entered clinical use, though FDA-cleared clonoSEQ is not broadly accessible. This study aimed to validate the applicability and prognostic utility of a commercial IGH-targeted NGS panel (LymphoTrack®) for MRD detection in MM, comparing its performance head-to-head with EuroFlow NGF in a cohort of transplant-eligible patients assessed 3 months post-ASCT.

The paper contextualizes MRD as a robust prognostic marker in MM, with IMWG criteria requiring sensitivities of at least 10^-5. Prior MRD approaches include: (1) NGF immunophenotyping, with EuroFlow’s two-tube, eight-color protocol capable of sensitivities near 2 × 10^-6 but dependent on sample quality and expert analysis; and (2) molecular assays detecting clonal immunoglobulin rearrangements. Historical allele-specific qPCR (ASO-PCR) is labor-intensive with limited applicability (approximately 40–75% of MM patients). FDA-cleared NGS (clonoSEQ) is validated but not widely available or cost-feasible for many centers. Imaging (PET-CT, MRI) provides complementary assessment but lacks standardized MRD procedures comparable to NGF and NGS. The study addresses the need for alternative, commercial, standardized NGS options and direct comparison with NGF.

Design and cohort: Retrospective analysis of 106 newly diagnosed, transplant-eligible MM patients enrolled in the Spanish GEM2012 trial. Inclusion required: (i) previously identified clonotypic IGH rearrangements at diagnosis, (ii) NGF at diagnosis and per-protocol follow-up, and (iii) sufficient genomic DNA for MRD. MRD assessment was performed at day ~100 post-ASCT using two independent bone marrow pulls for parallel flow and molecular studies. High-risk cytogenetics (t(4;14), t(14;16), or del17p) were defined per IMWG; R-ISS was applied. Ethics approvals and informed consent were obtained. NGS (LymphoTrack® IGH) workflow: Baseline clonotypes were identified using BIOMED-2 primers, PCR, and Sanger sequencing. For MRD, LymphoTrack® IGH primers targeting IGH framework regions amplified V(D)J sequences via a one-step, indexed PCR enabling multiplexing (up to 24 samples/run). Target input was ≥650 ng DNA to reach ~10^-5 sensitivity (~100,000 cell equivalents; 6.5 pg DNA/cell). A clonal B-cell line spike-in (100 cells) served for absolute quantification. Amplicons underwent purification (AMPure XP), quality/quantity checks (TapeStation 4200; KAPA kit or Qubit), library prep at 12–20 pM, and sequencing on Illumina MiSeq (v3 kits, 2 × 251 bp), aiming for ~1 million reads/sample. NGS analysis and MRD calling: Invivoscribe MRD Data Analysis Tool v1.1.0 tracked sample-specific IGH CDR3 clonotypes. Absolute quantification used spike-in cell counts and read ratios. Invalid runs had <20,000 total reads. MRD-positive required ≥2 identical clonotypic reads; otherwise MRD-negative (undetectable). FastQ files were also analyzed with Vidjil and Arrest/Interrogate. Patients were categorized by MRD status (positive/negative) and levels: negative; positive at ≥10^-5, between 10^-5 and 10^-4, or ≥10^-4; and by conventional response (VGPR/PR vs sCR) and cytogenetic risk. NGF workflow: Bone marrow was processed within 24 hours using EuroFlow NGF (two eight-color tubes). Data were merged in INFINICYT v2.0. MRD positivity required ≥20 aberrant plasma cells. Hemodilution was assessed via reductions in non-plasma populations (mast cells, erythroblasts, B-cell precursors). The GEM2012 NGF performance has been detailed elsewhere. Statistical analysis: SPSS v23.0 was used. Fisher’s exact and Mann–Whitney tests assessed categorical and continuous variables. Agreement between methods was tested with Bland–Altman plots; correlation by linear regression. Survival analyses used Kaplan–Meier and log-rank tests. Cox regression provided univariate and multivariate hazard ratios (HRs). PFS was from MRD assessment to progression/death/last follow-up; OS from MRD assessment to death/last follow-up. Two-sided p < 0.05 was significant.

- Applicability and input/sensitivity: NGF had higher cell input and sensitivity than NGS. Median cell input: NGS 210,672 (IQR 136,382–264,110), NGF 9,200,276 (IQR 5,702,369–10,305,207). Median sensitivity: NGS ~10^-5, NGF ~2 × 10^-6. NGS input constraints and sample concentration requirements limited applicability compared with NGF. - MRD rates: At ~3 months post-ASCT, MRD negativity was 50.0% by NGS (53/106) and 54.7% by NGF (58/106). Distribution across MRD levels differed modestly between methods (Table 2). - Concordance and correlation: Strong correlation between NGS and NGF MRD quantitation (r = 0.951; R^2 = 0.905). Concordant results in 91/106 (86.8%); 15/106 (14.2%) discordant (5 NGF+/NGS−; 10 NGF−/NGS+). Most NGS-only positives (7/10) were below 10^-5. - Survival by MRD status: MRD negativity by either method associated with superior outcomes. • PFS: Negative vs positive had significantly better 3-year PFS (p < 0.001 by both methods). Median PFS for MRD-positive: 46.7 months (NGS) and 34.2 months (NGF); not reached for MRD-negative. • OS: 3-year OS for MRD-negative vs MRD-positive: NGS 96.2% vs 77.3% (p < 0.01); NGF 96.6% vs 74.9% (p < 0.01). - Hazard ratios (Cox): Univariate—PFS HR for MRD-negative vs positive: NGF 0.20 (95% CI 0.09–0.44; p < 0.001); NGS 0.29 (95% CI 0.14–0.63; p = 0.002). OS HR: NGF 0.18 (95% CI 0.05–0.62; p = 0.007); NGS 0.21 (95% CI 0.06–0.75; p = 0.02). Multivariate—MRD negativity by NGF predicted lower risk of progression/death (PFS HR 0.20, 95% CI 0.09–0.45; p < 0.001) and OS benefit (HR 0.21, 95% CI 0.06–0.74; p = 0.02). MRD negativity by NGS was significantly associated with decreased risk of death (HR 0.21, 95% CI 0.06–0.75; p = 0.02); R-ISS I/II also independently favorable. - Subgroup insights: MRD negativity correlated with achieving CR/sCR. Among MRD-positive patients, high-risk cytogenetics conferred worse PFS/OS compared with standard-risk; MRD-negative outcomes were favorable regardless of cytogenetic risk. Five patients with double-negative MRD relapsed; all were in CR at MRD assessment, underscoring potential false negatives (patchy disease) or extramedullary relapse.

The study demonstrates that MRD negativity assessed by either a commercial IGH-targeted NGS assay (LymphoTrack®) or EuroFlow NGF is strongly prognostic for improved PFS and OS in transplant-eligible MM. Despite NGS applicability being affected by pre-analytical variables (sample quality, concentration needs, different marrow pulls) leading to lower cell inputs and a higher effective detection limit than NGF, the quantitative correlation and Bland–Altman agreement between NGS and NGF were excellent. These findings support the interchangeability of the two standardized approaches for prognostication when comparable sensitivity and appropriate sampling are ensured. The results highlight that depth of detection (cell input/sensitivity) and sampling consistency (same marrow pull) are crucial to minimize false negatives, given the patchy nature of MM bone marrow involvement and potential extramedullary disease. Importantly, MRD status outperformed conventional response categories and mitigated the adverse impact of high-risk cytogenetics among MRD-negative patients, reinforcing MRD as a key endpoint in clinical trials and clinical practice.

In a cohort of 106 MM patients assessed 3 months post-ASCT, a commercial IGH-targeted NGS assay (LymphoTrack®) showed high concordance with EuroFlow NGF for MRD detection and provided comparable, clinically meaningful prognostic information. MRD negativity by either method correlated with significantly improved PFS and OS. NGF achieved greater effective sensitivity due to higher cell input and fewer pre-analytical constraints, while NGS performance was impacted by sample handling and input limitations. The study supports MRD assessment as a robust endpoint in MM trials and underscores the need for standardized pre-analytical procedures, harmonized sensitivity (ideally down to 10^-6 when feasible), and consistent sampling (same marrow pull) to optimize MRD detection. Future work should include prospective, harmonized comparisons using identical sampling and matched sensitivity thresholds, and integration with imaging for extramedullary disease.

- Pre-analytical variability: Different bone marrow pulls were used for NGF and NGS; NGS often required sample concentration, biasing applicability and favoring NGF. - Sensitivity and input disparities: NGS analyses frequently had lower cell inputs (median ~2.1 × 10^5 cells) vs NGF (~9.2 × 10^6), resulting in higher detection limits and potential false negatives. - Incomplete adjustment for pre-analytical bias: These factors could not be fully accounted for in multivariate analyses. - Disease biology: Patchy marrow involvement and possible extramedullary disease can yield false-negative marrow MRD results; five double-negative patients relapsed. - Single timepoint post-ASCT assessment; limited sample size (n=106) from a single trial cohort may affect generalizability. - Only IGH-targeted NGS was evaluated; broader IGK/IGL coverage or alternate NGS platforms were not assessed.