Early peripheral blood and bone marrow MRD as prognostic markers in quadruplet-treated multiple myeloma without transplant

Abstract

Assessing measurable residual disease (MRD) in the peripheral blood (PB) of patients with multiple myeloma (MM) could mitigate issues of patchy and/or extramedullary disease leading to false negative MRD tests based on bone marrow (BM)-only assessments and reduce patient discomfort associated with serial BM sampling. Early disease response assessment using serum protein electrophoresis and immunofixation may not be able to account for interference with therapeutic monoclonal antibodies such as daratumumab (Dara) and elotuzumab (Elo), nor do they account for the extended half-lives of immunoglobulins due to recycling. Identification of a PB assay that better reflects a patient's response status compared to the conventional International Myeloma Working Group (IMWG) response criteria¹ could enable a more informative response-adapted treatment approach as part of decision-making regarding consolidative strategies following induction. PB MRD by flow cytometry has previously been found to be an early prognostic marker among patients receiving triplet induction and was shown to be prognostic when utilized much later in the maintenance setting.^{2, 3}

In this analysis, we investigated the potential role and prognostic significance of early MRD status after four cycles of carfilzomib-based quadruplet induction therapy by next-generation sequencing (NGS) in both the BM and PB as well as by conventional IMWG response criteria.

This was a post hoc analysis from two multicenter Phase II studies investigating the efficacy and safety of elotuzumab, carfilzomib, lenalidomide, and dexamethasone (Elo-KRd) (NCT02969837) and Dara-KRd (NCT 03500445) in patients with newly diagnosed MM without transplant intent.^{4, 5}

In both studies, a landmark analysis was performed after Cycle 4 (C4), which included MRD evaluation by clonoSEQ (Adaptive Biotechnologies, Seattle) NGS (limit of detection 6.8 × 10⁻⁷ with input of 20 μg DNA) in BM and PB samples. MRD assessment by NGS was performed on PB or BM mononuclear cells for patients with available suitable material and with an identified dominant clonotype. MRD < 10⁻⁶ in the BM was considered MRD negative, or MRD < 10⁻⁵ if insufficient input to determine 10⁻⁶ status. PB mononuclear cells were evaluated for MRD using the same assay; any quantifiable signal was considered positive in PB. Conventional response was assessed per IMWG response criteria.⁶

This landmark modified intent-to-treat (ITT) analysis focused on patients who remained on treatment at the end of C4 and had samples available for analysis (Supporting Information S1: Figure S1). Paired IMWG response and MRD status in the PB and BM were assessed for agreement and prognostic significance at the end of C4. Differences between groups were examined using the chi-square or Fisher's exact test. Agreement between BM and PB MRD status was assessed by Cohen's kappa coefficient. The Kaplan–Meier method was used to assess progression-free survival (PFS) and overall survival (OS) from the start of protocol therapy. Hazard ratios (HRs) and corresponding 95% confidence intervals (CIs) were estimated with the use of Cox proportional hazards regression models, and where applicable, the Bonferroni method was applied to adjust for multiplicity testing. The data cutoff date for this analysis was July 20, 2024.

Among 62 patients with IMWG response status and either a C4 BM or PB MRD result, 53 had a BM MRD result and 57 had a PB MRD result (Supporting Information S1: Figure S1 and Table S1). In the landmark analysis group with a median follow-up of 50 months, there were 17 PFS events, including 8 deaths (7 from progression). There were no differences in PFS (log-rank P = 0.90) or OS (P = 0.31) between those receiving Elo-KRd and Dara-KRd.

Of the 62 patients evaluable at C4, 19 (31%) achieved at least a complete response (CR). CR status at C4 was not associated with PFS (HR 0.47, 95% CI 0.13–1.63, P = 0.24) nor OS (HR 0.79, 95% 0.16–3.93, P = 0.78) using the landmark method (Supporting Information S1: Figure S2); the same was true for best IMWG response. Among the 20 patients with IgG kappa paraprotein at baseline and in a VGPR at C4, no patient had an M-protein > 0.1 g/L by mass spectrometry (the limit of detection for serum immunofixation), suggesting that interference from therapeutic monoclonal antibodies did not play a significant role in downgrading responses.

BM MRD samples by NGS at C4 were available in 53 patients, of whom 18 (34%) were MRD negative using the best available threshold (11 at 10⁻⁶ and 7 at 10⁻⁵). Using the landmark method, BM MRD negativity at C4 was associated with superior PFS (HR 0.11, 95% CI 0.015–0.85; P = 0.034; Figure 1A) and OS (HR not estimable, log-rank P = 0.033; Figure 1B).

PB MRD samples by NGS at C4 were available in 57 patients, of whom 42 (74%) were MRD negative. The median input of cellular equivalents of PB mononuclear cells was 1.9 million cells; 49/57 samples reached 10⁻⁶ sensitivity, and 8 reached 10⁻⁵ sensitivity.

Using the landmark method, PB MRD positivity at C4 was associated with inferior PFS (HR 5.27, 95% CI 1.94–14.33, P = 0.001; Figure 1C) and OS (HR 7.74, 95% CI 1.49–40.14, P = 0.015; Figure 1D); this association remained significant in bivariate analyses that accounted for ISS and the presence of ≥1 HRCA (Supporting Information S1: Table S2). The 4-year PFS for patients with C4 PB MRD negativity was 83% (95% CI: 66%–92%) versus 40% (95% CI: 14%–65%) for C4 PB MRD positivity (P = 0.0053).

There were 57 patients with paired IMWG response and PB MRD status (Supporting Information S1: Table S3). Of the 15 patients with PB MRD positivity at C4, 4 (27%) patients were in ≥CR; 28/42 (67%) patients with PB MRD negativity at C4 were in a PR or VGPR. When assessing the combined prognostic impact of IMWG response and PB MRD status, we found a significant association with PFS (log-rank P = 0.0054) but not with OS (log-rank P = 0.13), mainly driven by worse outcomes among patients who were PB positive and <CR (Figure 2A,B).

In comparing paired BM and PB MRD status (n = 48), there was only 52% agreement (Cohen's kappa 0.085, 95% CI 0.021–0.356). In total, 10/11 (91%) patients with PB MRD positivity at C4 also had BM MRD positivity; conversely, 22/37 (59%) patients with BM MRD positivity at C4 had PB MRD negativity (Supporting Information S1: Table S4 and Figures S3 and S4).

All 10 matched BM MRD-positive/PB MRD-positive samples at the C4 timepoint had higher concentrations of aberrant cells in the BM; BM MRD concentrations were a median log difference of 1.68 (range 0.07–3.52) higher than PB MRD concentrations.

When assessing for the combined prognostic impact of PB and BM MRD status, we found a significant association with PFS (log-rank P = 0.0012) and OS (log-rank P < 0.001), mainly driven by worse outcomes in the PB/BM MRD-positive group (n = 10) compared to PB/BM MRD negative (n = 15) or only one positive (n = 23) (Figure 2C,D).

This is the first study to assess the prognostic impact of PB and BM MRD assays early during extended quadruplet regimens for MM in the absence of ASCT. Together, PB and BM MRD assessments at C4 offer two prognostic insights. C4 PB MRD positivity by clonoSEQ identified patients at high risk for early relapse and was more predictive of outcomes than conventional IMWG response criteria, while early attainment of BM MRD < 10⁻⁶ marked a trajectory of sustained MRD negativity and prolonged PFS and OS. These complementary insights underscore the value of incorporating both compartments in early MRD assessment.

Importantly, the results did not change in an intention-to-treat analysis that imputed missing C4 MRD data as positive. Reported MRD negativity rates should be interpreted in the context of a landmark analysis, which excluded patients who experienced early progression or death.

PB paraprotein-based assays, including serum protein electrophoresis, immunofixation, and even mass spectrometry, are limited in assessing early response due to delayed paraprotein clearance. In contrast, post-induction MRD testing in PB and BM using NGS provides a more immediate and highly prognostic assessment of disease burden, outperforming IMWG criteria. Moreover, PB MRD may also be valuable in contexts such as oligosecretory disease, extramedullary involvement, or post-chimeric antigen receptor (CAR) T-cell therapy as a less invasive alternative to serial BM biopsies or extensive imaging.

In summary, we demonstrate that both PB MRD and BM MRD status at C4, assessed by clonoSEQ, were independently prognostic for PFS and OS in frontline therapy for MM without transplant intent (Supporting Information S2: Visual Summary). In contrast, conventional IMWG response at the same timepoint lacked prognostic significance. These findings highlight the potential utility of early PB and BM MRD status to inform risk-adapted treatment strategies. BM MRD negativity after 4–6 cycles of induction—achieved in 23%–27% after four cycles of Dara-KRd or Isa-KRd^{7, 8}; and in 50% after 6 cycles of Isa-KRd⁹—may justify deferring frontline ASCT in selected patients. This approach is supported by the MIDAS trial, where patients achieving MRD < 10⁻⁵ after induction had comparable rates of MRD < 10⁻⁶ following either ASCT or additional Isa-KRd.⁹ Conversely, PB MRD positivity at C4 flags a high-risk group that may benefit from treatment intensification, including ASCT or even CAR T-cell therapy. While later MRD assessments by BM and/or emerging PB modalities such as mass spectrometry may guide therapy de-escalation in patients with deep responses, early MRD assessment in both PB and BM provides a valuable opportunity to identify patients in need of a more aggressive approach. Future trials should integrate PB MRD into response-adapted strategies to personalize treatment intensity and improve outcomes in MM.

Benjamin A. Derman: Conceptualization; investigation; writing—original draft; methodology; writing—review and editing; formal analysis; data curation; supervision; project administration. Jennifer Cooperrider: Writing—review and editing; investigation; project administration. Anna Pula: Writing—review and editing; project administration; formal analysis; data curation; investigation. Heidi Simmons: Writing—review and editing; resources; methodology; investigation; conceptualization. Tadeusz Kubicki: Writing—review and editing; formal analysis; project administration; data curation; resources; investigation. Jeffrey Zonder: Writing—review and editing; investigation; resources; project administration. Aimaz Afrough: Writing—review and editing; investigation; project administration; resources. David Grinblatt: Writing—review and editing; investigation; project administration; resources. Jacalyn Rosenblatt: Resources; project administration; writing—review and editing; investigation. Andrew Kin: Resources; project administration; writing—review and editing; investigation. David Avigan: Investigation; writing—review and editing; project administration; resources. Sunil Narula: Resources; project administration; writing—review and editing; investigation. Shayan Rayani: Investigation; writing—review and editing; project administration; resources. Ken Jiang: Resources; data curation; investigation; methodology; writing—review and editing. Ajay Major: Writing—review and editing; investigation; resources; project administration. Theodore Karrison: Resources; project administration; writing—review and editing; investigation; conceptualization; methodology. Allison Jacob: Conceptualization; methodology; writing—review and editing; investigation; project administration; resources. Andrzej J. Jakubowiak: Conceptualization; investigation; writing—review and editing; resources; project administration; supervision.

B.A.D. declares consultancy for Janssen, Sanofi, COTA, and Canopy; independent reviewer of a clinical trial for BMS; research funding from Amgen and GSK. J.C., Ta.Ku., A.K., S.N., S.R., K.J., A.M., and Th.Ka declare no competing financial interests. A.P. declares honoraria from Janssen, Roche, and Amgen; research funding from Sanofi and GSK. H.S. and A.J. are employees of Adaptive Biotechnologies. J.Z. declares employment by BMS; advisory role for Takeda, Sanofi, and Regeneron, and research funding from BMS and Janssen. A.A. declares advisory roles for Karyopharm, BMS, Sanofi, Janssen, and Pfizer; research funding from AbbVie, Adaptive Biotechnologies, K36 Therapeutics, Janssen, and Regeneron. D.G. declares consultancy for BMS. J.R. declares advisory role for Janssen; data safety monitoring committee for Karyopharm; research funding from BMS and Sanofi. L.D.A. declares honoraria from GSK, Amgen, Janssen, Bristol Myers Squibb/Celgene, Prothena, AbbVie, BeiGene, Cellectar, and Sanofi; advisory roles for Janssen, Amgen, GSK, Bristol Myers Squibb/Celgene, AbbVie, Prothena, BeiGene, Cellectar, and Sanofi; research funding from Celgene, Janssen, and BMS. D.A. declares advisory roles for BMS, Takeda, Chugai, Kite Pharma, and Janssen. A.J.J. declares advisory roles for AbbVie, Amgen, BMS, Celgene, GSK, Gracell, Janssen, Karyopharm, and Sanofi.

The studies contained herein were approved by the Institutional Review Boards of each participating institution. All procedures involving human participants were conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants.

This research received no funding.