Introduction

Multiple Myeloma is a plasma cell malignancy characterized by anemia, bone lesions, hypercalcemia, renal failure and predisposition to infections. The outcomes of patients with newly diagnosed multiple myeloma (NDMM) continue to improve owing to the tremendous improvement in understanding of disease biology leading to progress in therapeutics and supportive care measures [1, 2]. The use of anti-CD38 monoclonal antibodies (anti-CD38 mAb), proteasome inhibitors (PI), immunomodulatory agents (IMiDs) and corticosteroids (dex) in quadruplet combinations, particularly when followed by upfront autologous stem cell transplantation (ASCT), resulted in unprecedented depth and duration of response in NDMM [3,4,5,6]. However, NDMM is quite heterogeneous in biology, presentation, and response to therapy [7, 8]. Current therapeutic standards therefore are insufficient for some patients and may be excessive for others.

The availability of efficacious therapies has catalyzed the need to develop more sophisticated methods to detect residual disease and understand treatment response. With our current triplet and quadruplet combinations, most patients will achieve deep responses using traditional paraprotein based testing, and discriminating outcomes is challenging. In this setting measurable residual disease (MRD) has been shown to be one of the most important dynamic prognostic markers for progression free (PFS) and overall survival (OS) [9]. In comparison to traditional measurement of paraprotein (surrogate) to characterize disease response, MRD provide a direct estimation of tumor burden (clonal plasma cells). Indeed, Oncologic Drugs Advisory Committee (ODAC) recommended acceptance of MRD as an endpoint for accelerated drug approval to the FDA due to robust individual-level surrogacy for PFS and OS, along with moderate to strong trial-level surrogacy for PFS [10].

ASCT has been a prime consolidative strategy for patients with transplant eligible NDMM. However, while it helps improve the depth, including MRD negativity, and duration of response in NDMM, it has short- and long-term toxicities. MRD provides a novel opportunity to accurately characterize the incremental benefit of ASCT on depth of response. By assessing the pre- and post-ASCT MRD, we have a direct estimation of clone size as opposed to paraprotein measurements which can be fraught with challenges such as heterogeneity in paraprotein clearance and inaccurate characterization of the disease burden. Utilizing MRD therefore helps isolate the effect of ASCT independent of induction regimen and can serve as a benchmark for the development of alternative consolidations strategies. The remarkable efficacy of immunotherapy in relapsed MM has led to their development in earlier lines of therapy. The premise is to improve the toxicity and efficacy balance in the setting of lower burden and less refractory disease as more common early in the disease course. T cell redirecting therapy (TCRT) including bispecific antibodies and chimeric antigen receptor T cell therapies are being investigated as an alternative and/or adjunctive strategy with ASCT in NDMM. However, to understand their relative efficacy, a clear benchmark for expected change in clone size and MRD burden for the ASCT consolidative strategy is vital for appropriate comparisons.

We sought to isolate the effect of ASCT, as legacy consolidative therapy in NDMM, on MRD burden and generate a benchmark for evaluation of TCRT in this setting.

Methods

We collected MRD status by next generation sequencing (NGS; clonoSEQ®) irrespective of IMWG response post triplet (PI+IMiD+dex) and quadruplet (triplet + anti-CD38 mAb) induction and 60–100 days after ASCT in consecutive patients at one large volume myeloma program along with participants of two single arm phase 2 trials with quadruplet induction followed by ASCT. We analyzed MRD at a threshold of 10−5 and 10−6 and as a continuous variable. We report the effect of ASCT on MRD negativity rates and quantitative disease burden according to type of induction therapy and cytogenetic subsets.

Patients and study design

We analyzed data from two clinical trials: Monoclonal Antibody-Based Sequential Therapy for Deep Remission in Multiple Myeloma (MASTER; NCT03224507) and Minimal Residual Disease Response-adapted Deferral of Transplant in Dysproteinemia (MILESTONE; NCT NCT04991103). MASTER is a single arm, phase 2 multicenter study of MRD guided quadruplet therapy for NDMM, while MILESTONE is a single center, phase 2 study evaluating the use of MRD for decision to proceed with ASCT following quadruplet induction [3, 11, 12]. We also included consecutive patients not participating in a clinical trial and receiving ASCT for NDMM following induction with triplet and quadruplet therapy as standard of care as a part of institutional guidance for management of NDMM who had pre and post ASCT MRD results. All patients collected autologous stem cells shortly after completion of triplet or quadruplet induction using granulocyte colony-stimulating factor, with or without plerixafor as mobilization per institutional policy. Patients received high dose melphalan (140–200 mg/m2 at investigator discretion) followed by ASCT.

The following abnormalities when present by fluorescence in situ hybridization (FISH) constituted high risk cytogenetic abnormalities (HRCA): gain/amplification 1q, t(4;14), t(14;16), and del(17p) [13]. Patients were grouped according to the number of HRCA present in the clonal plasma cells (0, 1, or 2 + ).

MRD assessment

We collected the “first pull” bone marrow aspirate for MRD assessment using next-generation sequencing (NGS; ClonoSEQ® platform; Adaptive Biotechnologies, Seattle, WA) upon completion of induction therapy and 60–100 days after ASCT. We performed MRD assessment in all patients, simultaneously with and not influenced by IMWG response assessment, and reported MRD defined by thresholds of 10−5 and 10−6 at those timepoints. For patients with MRD ≥ 10−5 post induction, we describe quantitative relative changes in MRD burden, expressed by log10 of the relative reduction in MRD burden with AHCT, defined as log10(MRD pre-AHCT/MRD post-AHCT). MRD assessment and reporting aligns with the International Myeloma Working Group (IMWG) consensus criteria for response and MRD assessment in MM, and the international harmonization in performing and reporting MRD assessment in MM trials [14, 15].

Study oversight

The clinical trials were conducted under the University of Alabama at Birmingham O’Neal Comprehensive Cancer Center Data and Safety Monitoring Plan and in accordance with the International Council for Harmonization Good Clinical Practice Guidelines and the principles of the 2013 Declaration of Helsinki. All patients provided written informed consent for participation. Institutional review boards approved the protocol and related documents. The clinical trials are registered in clinicaltrials.gov (NCT03224507; NCT04991103). Amgen and Janssen provided material support for the MASTER clinical trial. The American Cancer Society Institutional Research Grant funds the MILESTONE trial internally. The non-clinical trial patients provided written informed consent to participate in an institutional database.

Statistics

We compared proportions using Fisher’s exact test and numerical variables across groups using the Mann–Whitney U test. A two-sided p < 0.05 was considered statistically significant. All analyses were conducted using IBM SPSS (v27.0).

Results

We identified 330 patients with pre- and post-ASCT MRD data. The median age of the entire cohort was 61 years (range 30-79), 185 (56%) males, 215 (65%) Non-Hispanic Whites (NHW). Three hundred and seventeen of the 330 (96%) patients had cytogenetic information at the time of diagnosis, 106 (33%) had 1 HRCA and 44 (14%) had 2 + HRCA. Of the 330 patients, 124 (38%) were clinical trial participants. Baseline characteristics are summarized in Table 1.

Table 1 Baseline Characteristics.
Full size table

Induction and ASCT effect on MRD

Of the 330 patients, 279 (85%) received quadruplets during induction therapy. The proteasome inhibitor in the quadruplet was carfilzomib in 113 (41%) and bortezomib on the remaining patients. The median number of quadruplets cycles prior to ASCT was 4 cycles (range 4–12). The best response to quadruplet induction therapy was stable disease (SD) in one patient, partial response (PR) in 50 (18%), very good partial response (VGPR) in 141 (51%), complete response (CR) in 17 (6%) and stringent complete response (sCR) in 70 patients (25%). Following ASCT, the IMWG best responses included 18 (6%) PR, 105 (38%) VGPR, 35 (13%) CR and 119 (43%) sCR (Supplementart Fig. 2). Two patients experienced progressive disease at or within 100 days following ASCT. Granular data on transition across MRD strata with ASCT is displayed in Fig. 1. Among patients receiving quadruplet induction, 82 (29%) achieved MRD < 10−5 following induction therapy, and 165 (59%) had MRD < 10−5 following ASCT. There were 42 patients (15%) who had achieved MRD < 10−6 post quadruplet induction, increasing to 125 (45%) post ASCT (Fig. 1, panel A). Among the 197 quadruplet patients with MRD ≥ 10−5 post-induction, ASCT lowered the MRD burden ≥ 1 log10 in 135 (68.5%) (Fig. 1, panel C). Patients with high disease burden (MRD ≥ 10−3) following induction with quadruplet were less likely to convert to MRD < 10−5 with ASCT (Fig. 1, panel A). A minority of patients with MRD > 10−3 converted to MRD < 10−5 (20%) and MRD < 10−6 (17%) with ASCT.

Fig. 1
figure 1

Sankey plots showing change in MRD strata with ASCT in (A) Patients treated with quadruplet induction, (B) patients treated with triplet induction and. C Quantitative changes in MRD burden after ASCT in patients with MRD ≥ 10−5 post induction, based on exposure to CD38 monoclonal antibody in induction.

Full size image

Of the 330 patients, 51 (15%) received triplet therapy during induction. For patients receiving triplet induction, the best IMWG response following induction was PR in 18 (35%), VGPR in 22 (43%), CR in 5 (10%) and sCR in 6 (12%). Following ASCT, the IMWG best responses included 5 (10%) PR, 25 (49%) VGPR, 13 (25%) CR and 8 (16%) sCR (supplemental figure 2). Among triplet patients, the rate of MRD < 10−5 post-induction was 16% increasing to 41% post-ASCT. The rate of MRD < 10−6 post induction was 4% increasing to 24% post ASCT (Fig. 1, panel B). Among patients with MRD ≥ 10−5 post-induction, ASCT lowered the MRD burden ≥ 1 log10 in 56% patients (Fig. 1, panel C). The incremental decrease in disease burden with ASCT was similar among patients who received triplets and quadruplets induction (Fig. 1, panel C).

Overall, 69 (21%) patients received reduced dose melphalan (140 mg/m2) as ASCT conditioning regimen per investigator discretion. Melphalan dose intensity did not influence the incremental reduction in disease burden with ASCT (Supplemental figure 1).

Cytogenetic risk and ASCT effect on MRD

Among patients who received quadruplet induction (N = 279), 270 had baseline cytogenetic information available at the time of diagnosis, 97 (36%) had 1 HRCA, 40 (15%) had 2 + HRCA. Following quadruplet induction therapy, MRD < 10−5 was 19%, 31% and 28% improved to 53%, 66% and 65% post-ASCT for patients with 0, 1 and 2 + HRCA, respectively (Fig. 2). The MRD < 10−6 post induction was 17%, 14%, 10% following induction, improving to 41%, 50%, 48% post-ASCT for patients with 0, 1 and 2 + HRCA, respectively. Patients with ultra-high-risk MM (2 + HRCA) derived greatest reduction in disease burden with ASCT (Figs. 2 and 3).

Fig. 2
figure 2

Sankey plots showing change in MRD strata with ASCT among patients with (A) 0 HRCA, (B) 1 HRCA, and (C) 2 + HRCA.

Full size image
Fig. 3
figure 3

Quantitative changes in MRD burden after ASCT in patients with MRD ≥ 10−5 post induction across different cytogenetic subsets for (A) all patients, (B) patients who received quadruplet induction.

Full size image

Among all patients with MRD ≥ 10−5 post induction, ASCT lowered the MRD burden ≥ 1 log10 for 60%, 72% and 81% of patients with 0, 1 and 2 + HRCA, respectively. Similar results were seen in the subset of patients who received quadruplet induction (Fig. 3).

Discussion

Our study isolates the impact of ASCT on disease burden using direct quantification of clonal plasma cells rather than a surrogate biomarker (paraprotein). We provide granular information on reduction in disease burden in different cytogenetic subsets and show that patients continue to experience a similar magnitude of reduction in disease regardless of incorporation of anti-CD38 mAb in induction. In aggregate, our study establishes a benchmark for the design of future clinical trials by defining expected MRD trajectories with quadruplet induction followed by ASCT.

In NDMM, MRD negativity has been associated with longer PFS and OS, with deeper MRD (<10−6) responses being more discriminatory in delineating patient outcomes (compared to 10−5 or greater) [9, 10, 16]. Given its prognostic significance, MRD was reviewed favorably as an early clinical endpoint to support the accelerated approval of novel therapies alongside confirmatory demonstration of longer-term traditional endpoints [10]. It is being routinely evaluated in the clinic among patients with NDMM and those achieving MRD negativity experience improved long term outcomes. For decades, ASCT has been a cornerstone of anti-plasma cell directed therapy with impressive efficacy with well-known short-term toxicities [17, 18]. However, with deeper understanding of disease evolution and biology, we are beginning to unpack the long-term toxicities of this therapy in a population that is likely to live longer with improved therapeutics [19]. Several datasets have provided insight into its efficacy particularly in patients with ultra-high-risk MM (2 + HRCA) and our data again supports a large effect size for this subset [11, 13, 20,21,22]. Greater benefit among patients with ultra-high risk MM is likely due to the use of alkylator therapy (melphalan) as an additional non-cross-resistant approach to reduce the potentially more proliferative disease not eliminated by novel agents in a complementary manner. In fact, log kill hypothesis, which is expected for melphalan, hypothesizes that fractional killing is independent of initial disease burden demonstrated by similar magnitude of disease burden reduction irrespective of monoclonal antibody use in induction. However, the role, timing and value of ASCT for all eligible patients, particularly those with standard risk MM and single HRCA MM where available induction strategies have provided deep responses and improved outcomes is being questioned. Establishing an MRD trajectory benchmark for the current standard of quadruplet induction and ASCT provides a tool for sample size estimation and optimal design of studies deploying novel therapies in addition to or in replacement of ASCT.

Quadruplet therapies have significantly improved the PFS and OS of patients with newly diagnosis multiple myeloma for both transplant eligible and ineligible patients [4,5,6, 23,24,25]. Several large, randomized studies in transplant eligible NDMM have evaluated the quadruplet induction, ASCT and post-ASCT quadruplet consolidation followed by maintenance (such as CASSIOPEIA, GRIFFIN, PERSEUS). However, the magnitude of incremental benefit and isolated effect of ASCT in disease control remains not well described since MRD is assessed following post-ASCT consolidation and not immediately following ASCT and exact effect of the ASCT procedure on disease burden is unclear [4]. Additionally, there is a lot of heterogeneity among the reporting of MRD across different trials, differences in testing methodology (flow cytometry vs. NGS), numbers of cycles of induction/consolidation, and response category of patients in whom MRD was assessed. The recently reported randomized phase 3 ISKIA study provided more granular data in this setting. Transplant eligible patients with NDMM patients were enrolled and randomized to receive a carfilzomib containing quadruplet with or without CD38 mAb, followed by ASCT. In ISKIA, pre-ASCT MRD < 10−5 was 45% increasing to 64% post ASCT (MRD < 10−6 increasing from 27% to 52% post ASCT) compared to our report of 29% pre-ASCT increasing to 59% post-ASCT (MRD < 10−6 15% increasing to 45% post ASCT). Compared to ISKIA, our study is larger with more granular data in quadruplet treated patients in a population which is more representative of practice patterns in the US (older patients, bortezomib as the proteasome inhibitor).

We also tested MRD on all consecutive patients regardless of IMWG response to capture. This leads to important observations. For instance, patients with high disease burden (MRD > 10−3) following quadruplet therapy, a third of the population, 55% of which were in ≥ VGPR and 7% in ≥ CR, are unlikely to reach MRD < 10−5 and MRD < 10−6 with ASCT alone and are in greatest need for novel approaches.

Our analysis has several limitations. Our population is more heterogenous than typical clinical trial populations both in demographics and in regimen used for induction therapy. We mitigated that by emphasizing the analysis in patients receiving quadruplet induction. We only included patients who were MRD evaluable using NGS, those patients were baseline samples could not be retrieved or clonogenic sequences could not be identified were not included. Yet another inherent limitation is that the data is conditional upon patients proceeding to ASCT after induction therapy and did not include patients who could not proceed due to toxicity, or early disease progression during induction precluding upfront ASCT. We did not attempt to correlate MRD endpoints before or after ASCT with long term follow up. While this correlation has been unambiguously demonstrated, the trials and clinical practice included in this analysis adapted the post ASCT to the very MRD endpoint, either by intensification or deintensification of therapy, invalidating or at least blurring such analysis. A small number of patients in this study were treated on a clinical trial protocol where decision to proceed ASCT was deferred among patients achieving MRD < 10−5 post induction following quadruplet induction. While the absolute number of patients who achieved this milestone from very small, there is a small downward bias in estimation of true pre-ASCT MRD rate.

TCRT and other drug classes such as CELMoDs™ have demonstrated outstanding depth and duration of response in relapsed MM [26,27,28,29,30]. Their unparalleled efficacy in the multiply relapsed setting has catalyzed their rapid movement into earlier lines with hope of improving durability of response and the potential for curative approach to NDMM [31]. Additionally, patients such as those who have functional high-risk disease without significant response or persistent MRD burden following treatment with quadruplets with and without ASCT may be candidates for treatment intensification [7]. Several studies are ongoing and being planned to directly compare these head-to-head against ASCT (NCT05257083) and others as post-ASCT consolidation therapies (NCT05243797, NCT05317416, NCT05827016, NCT06045806, NCT04133636, NCT05434689). In a newly framed landscape where MRD will be employed as endpoint for accelerated drug approval, our study provides granular, strata-specific benchmark to design trials testing TCRT in NDMM.