Minimal residual disease measurement in blood by mass spectrometry identifies long-term responders in multiple myeloma Open Access

Multiple myeloma (MM) is the second-most common hematological neoplasm, affecting >185 000 patients annually around the world.¹ The disease is caused by the malignant transformation of plasma cells, which produce monoclonal protein (M protein). Quantification of bone marrow infiltration by malignant cells and the serum concentration of the M protein are the mainstays of myeloma disease assessment.² Historically, this assessment was performed by a combination of immunohistochemical staining of bone marrow biopsies; protein electrophoresis of serum and urine; and, more recently, immunoassays to quantify serum free light chains. However, with remarkable progress in treatment efficacy over the last 2 decades, these standard methods have proven to be insufficient markers of depth of response achieved with novel drugs combinations. This has led to widespread implementation of minimal residual disease (MRD) assessment in the bone marrow using next-generation sequencing (NGS), which can detect malignant plasma cells with a limit of detection (LoD) as low as 1 myeloma cell per million analyzed nucleated cells (ie, 10⁻⁶)³ by monitoring gene sequences specific to the M protein expressed by malignant plasma cells. Reaching MRD negativity in the bone marrow is associated with improved outcomes in both newly diagnosed (ND) transplant-eligible and ineligible patients, as well as in the relapsed/refractory setting.^4,5 Moreover, the depth of MRD evaluation (10⁻⁵ or 10⁻⁶) also correlates with survival outcomes, with a better prognosis for patients achieving negativity at the lower threshold.⁶ This remarkable prognostic performance of MRD in the bone marrow has led to the incorporation of MRD-guided treatment schema in clinical trials (eg, ATLAS, MASTER, MIDAS, and MASTER-2).^7,8 Nevertheless, there is still room for improvement, because bone marrow biopsy and aspiration is a costly and invasive procedure and is susceptible to false-negative results in cases of patchy disease infiltration or extramedullary disease.⁹ These limitations can be addressed by using techniques that enable sensitive assessment of M protein concentration in the peripheral blood through various mass spectrometry (MS) approaches.^10,11 Currently, intact immunoglobulin MS is the most advanced and validated platform for this type of evaluation. Matrix-assisted laser desorption ionization–time of flight (MALDI-TOF) allows for M protein detection with a sensitivity up to 100 times higher than standard serum protein electrophoresis (SPEP)/immunofixation.¹² Several studies have demonstrated that this technique produces results comparable with MRD evaluation in the bone marrow at a threshold of 10⁻⁵.^13-15 Sample preparation using liquid chromatography–MS (LC-MS) offers the potential for an additional log improvement in M protein detection sensitivity.¹⁶ Previous analyses suggest that combining MRD-testing modalities could discriminate outcomes in patients with MM after transplant.¹⁵ In this study, we pooled data from 3 prospective nonrandomized phase 2 studies coordinated by The University of Chicago, to compare the prognostic value of EXENT and LC-MS in the blood with MRD in the bone marrow at respective thresholds of 10⁻⁵ and 10⁻⁶ and to evaluate whether MS can improve identification of patients with prolonged duration of response.

Data were pooled from 3 prospective investigator-initiated phase 2 clinical trials coordinated by The University of Chicago. The KRd-ASCT (carfilzomib, lenalidomide, and dexamethasone–autologous stem cell transplant) study (NCT01816971) enrolled patients with NDMM who were eligible for ASCT, whereas the Dara-KRd (daratumumab-KRd; NCT03500445) and the Elo-KRd (elotuzumab-KRd; NCT02969837) trials recruited patients with NDMM regardless of ASCT eligibility. Only patients with available additional serum samples for exploratory analyses were included in this study. Details regarding patient enrollment and inclusion/exclusion criteria can be found elsewhere.^17-19 The studies were conducted in accordance with US Food and Drug Administration and International Conference on Harmonisation guidelines for good clinical practice, the Declaration of Helsinki, Health Canada, and any applicable health authority. The study protocols were approved by the institutional review boards or ethics committees of participating institutions. All patients provided written informed consent.

In the KRd-ASCT study, patients received 4 cycles (28 days per cycle) of KRd induction, followed by ASCT, 4 cycles of KRd consolidation, and 10 cycles of KRd maintenance, totaling 18 cycles of KRd. Single-agent lenalidomide maintenance therapy was recommended after completion of protocol therapy. In the Elo-KRd trial, patients received 12 to 24 cycles of Elo-KRd followed by Elo-Rd maintenance until disease progression, with the duration of Elo-KRd guided by MRD results in the bone marrow. The Dara-KRd study assigned patients to receive 24 cycles of Dara-KRd, after which patients were recommended to receive at least lenalidomide maintenance therapy. A detailed description of dosing in each protocol have been published previously^17-19; a summary of the treatment schema is presented in supplemental Figure 1. The studies, their protocols, and amendments were approved by the institutional review board at each participating institution. The studies are conducted in accordance with the Declaration of Helsinki and the International Conference on Harmonisation of good clinical practice guidelines.

Serum samples for MS assessment were collected after 18 cycles of KRd (end of protocol therapy) and after 1 year of follow-up in the KRd-ASCT study; after cycles 4, 8, 12, 18, and 24 in the Elo-KRd trial; and after cycles 4, 8, 12, and 24 in the Dara-KRd trial. In all 3 trials, screening serum samples were used to identify the mass-to-charge ratio of the +2-charge state from the M protein’s light chain for subsequent tracking. Stored serum samples were analyzed using the EXENT immunoglobulin isotypes (GAM) assay (EXENT Solution, Binding Site, part of Thermo Fisher Scientific), which uses MALDI-TOF MS. EXENT is an integrated clinical analyzer under development, consisting of an automated liquid handler (EXENT-iP500), a MALDI-TOF mass spectrometer (EXENT-iX500), and instrument control/data analysis software (EXENT-iQ) for detecting and quantifying monoclonal immunoglobulins in serum. This method allows for the identification of M protein with a lower LoD of 0.0015 g/dL. Only samples found to be negative for M protein by EXENT were reflexed to LC-MS testing, as previously described.²⁰ 

At the same time points as MS evaluation, MRD was assessed by NGS using clonoSEQ (Adaptive Biotechnologies, Seattle, WA; LoD 6.8 × 10⁻⁷ with input of 20 μg DNA). Results were reported at thresholds of 10⁻⁵ or 10⁻⁶, depending on the number of nucleated cells analyzed and the uniqueness of the tracked clonal immunoglobulin gene sequence.

Time-to-event analyses such as progression-free survival (PFS) and overall survival (OS) were performed using the Kaplan-Meier method, indexed from the start of treatment, and groups were compared using the log-rank test. Hazard ratios (HRs) and 95% confidence intervals (CIs) for survival data were derived from Cox proportional hazards regression models. Statistical analysis was performed using R software (version 4.3.0). Sensitivity, specificity, positive predictive value, and negative predictive value were calculated, with MRD by NGS as a reference. The degree of agreement was calculated with Cohen κ.

The pooled analysis included 97 patients: 36 from the KRd-ASCT study, 38 from the Dara-KRd study, and 23 from the Elo-KRd trial. The median follow-up for the entire cohort was 55.9 months (range, 7.1-128.9), with differences observed across the pooled trials: 102.0 months (range, 22.2-128.9) for KRd-ASCT, 40.1 months (range, 7.1-64.1) for Dara-KRd, and 49.6 months (range, 11.1-76.2) for Elo-KRd. The median age was 59 years. Of 97 patients, 39 (41%) were female; 17 of 97 (18%) were Black; and 10 of 97 (10%) were Hispanic or Latino. In terms of disease characteristics, 13 patients (13%) presented with International Staging System stage III, and 43 patients (44%) had at least 1 high-risk cytogenetic abnormality, defined as 17p deletion, t(4;14), t(14;16), or 1q gain/amplification. These characteristics were generally comparable across the 3 pooled trials, with the exception of Eastern Cooperative Oncology Group performance status, sex, and race (Table 1). Of note, 68 (70%) patients had an immunoglobulin G myeloma isotype.

MS analysis was performed on 266 serum samples collected at cycle 4 (n = 52), cycle 8 (n = 56), cycle 12 (n = 48), cycle 18 (n = 64), and cycle 24 or 1-year follow-up (n = 46). Paired MRD results from the bone marrow were available for 86 patients, encompassing 212 evaluations, of which 171 allowed for MRD assessment at a threshold of 10⁻⁶. A detailed breakdown of paired samples at each time point is provided in supplemental Table 1.

According to standard International Myeloma Working Group criteria, 75 patients (77%) achieved a stringent complete response (CR) as their best overall response, 5 patients (5%) achieved a CR, resulting in a rate of at least CR of 82%. Additionally, 17 patients (18%) achieved a very good partial response. MS negativity by EXENT, as a best response, was reported in 67 (69%) patients and when the negative samples were reflexed for LC-MS testing, the number of EXENT-negative and LC-MS–negative patients was 35 (36%; Figure 1). The outcomes for each of the 3 trials are presented supplemental Figure 2. MRD negativity at 10⁻⁵ was achieved as a best response in 66 of 86 evaluable patients (77%), and at a threshold of 10⁻⁶ in 36 of 69 patients (52%).

When analyzing response kinetics, negativity rates began to plateau after 8 cycles for MRD assessments in the bone marrow and after 12 cycles for MS-based evaluation. EXENT and LC-MS negativity was observed as early as after 4 cycles of treatment (Figure 2). The overall agreement between EXENT and MRD at 10⁻⁵ was 69% (Cohen κ = 0.36), with the highest agreement at the cycle 24 or 1-year follow-up time point (76%). For the discrepant cases, there were 19% EXENT positive/NGS 10⁻⁵ negative, and 11% showed an opposite pattern (Figure 3A). For LC-MS and MRD at 10⁻⁶, the overall agreement was 71% (Cohen κ = 0.35), with the highest agreement after cycle 18 (78%). Most of the discrepancies arose from LC-MS–positive and NGS 10⁻⁶–negative results (Figure 3B). Importantly, none of the 4 LC-MS–negative but NGS-positive patients experienced disease progression, whereas 5 progression events occurred among the 17 LC-MS–positive/NGS-negative patients. However, the differences in survival were not statistically significant (data not shown). Detailed results for sensitivity, specificity, positive predictive value, and negative predictive value at each time point are provided in supplemental Table 2.

Achieving EXENT negativity as a best response was associated with superior PFS (median not reached vs 55.0 months; HR, 0.29; 95% CI, 0.14-0.58; P = .0002; Figure 4A) and OS when compared to EXENT-positive patients (median not reached vs 123.6 months; HR, 0.21; 95% CI, 0.06-0.68; P = .004; Figure 4B). Similarly, PFS in patients with LC-MS negativity was significantly longer than in patients with LC-MS positivity (median not reached vs 73.7 months; HR, 0.22; 95% CI, 0.08-0.63; P = .002; Figure 4C), whereas there were no significant differences in OS (median not reached vs 123.6 months; HR, 0.36; 95% CI, 0.08-1.63; P = .17; Figure 4D). A sensitivity analysis for PFS demonstrated similar effects of MS negativity across all 3 pooled trials (supplemental Figures 3-5). The PFS curves showed clear separation when patients were grouped by the depth of response measured via MS, with EXENT-positive patients having the shortest PFS, those who were EXENT negative/LC-MS positive having intermediate PFS, and those with LC-MS negativity having the longest PFS (Figure 5A). For OS, patients with EXENT negativity had a similar prognosis, regardless of the LC-MS status (Figure 5B). A similar association was observed for PFS based on the depth of MRD negativity in the bone marrow, with the longest survival among patients who were negative at a threshold of 10⁻⁶ (supplemental Figure 6). Negativity at both 10⁻⁵ and 10⁻⁶ cutoffs was associated with longer PFS (median not reached vs 55.0 months [HR, 0.31; 95% CI, 0.13-0.75; P = .006] and median not reached vs 73.7 months [HR, 0.36; 95% CI, 0.32-0.83; P = .01]); whereas there were no significant differences for OS when patients were stratified by the MRD status in the bone marrow (supplemental Figure 7).

MS negativity using EXENT was achieved in 28 of 43 (65%) patients with high-risk cytogenetics and in 39 of 54 (72%) standard risk patients. The corresponding numbers for LC-MS were 14 of 43 (33%) and 21 of 54 (39%; supplemental Figure 8), respectively. For both high-risk and standard-risk patients, PFS was longer in those who achieve MS negativity (standard-risk, EXENT [HR, 0.21; 95% CI, 0.08-0.54; P = .0003]; high-risk, EXENT [HR, 0.27; 95% CI, 0.09-0.85; P = .02]; standard-risk, LC-MS [HR, 0.28; 95% CI, 0.06-1.28; P =.08]; high-risk, LC-MS [HR, 0.18; 95% CI, 0.04-0.79; P = .01]). Notably, achieving LC-MS negativity appeared to mitigate the adverse prognosis associated with high-risk cytogenetics (supplemental Figure 9).

We assessed the prognostic implications of combining the evaluation of measurable disease in the bone marrow and peripheral blood. Among patients evaluable by both modalities, a complementarity was observed between EXENT and MRD at a threshold of 10⁻⁵ and 10⁻⁶, with patients with double-negative results achieving significantly longer PFS than those negative in only 1 of the analyzed modalities (10⁻⁵: median not reached vs 70.5 months; HR, 0.29; 95% CI, 0.13-0.67; P = .002; Figure 6A; 10⁻⁶: median not reached vs 49.7 months; HR, 0.15; 95% CI, 0.04-0.57; P = .001; Figure 6B); this was not observed for OS (supplemental Figure 10). When analyzing OS and PFS, MRD negativity at a threshold of 10⁻⁶ and LC-MS negativity have comparable prognostic value for PFS and OS without complementarity (Figure 6C). Nevertheless, patients who were negative by both LC-MS and MRD at a threshold of 10⁻⁶ had an excellent prognosis, with only 1 PFS event among 19 patients who reached such a deep response (Figure 6D).

This is, to our knowledge, the largest data set reporting the assessment of MRD in the peripheral blood using 2 different MS assessment methods in association with MRD in bone marrow by NGS in patients with NDMM. We demonstrate, across multiple carfilzomib-containing induction and consolidation strategies with or without ASCT, that MS negativity with EXENT is associated with improved PFS and OS, and that LC-MS negativity identifies patients with a very high probability of long-term disease control (>85% progression free at 5 years). MS by EXENT platform and MRD by NGS at 10⁻⁵ or 10⁻⁶ in the bone marrow were complementary, adding MS by EXENT to MRD by NGS significantly improved the prognostic value, with respect to PFS (5-year PFS of 79% and 82% at 10⁻⁵ or 10⁻⁶, respectively, for patients with double negativity), and the 5-year PFS was 89% among patients with best response of LC-MS and NGS 10⁻⁶ negativity.

Our study clearly demonstrates a continuum of the depth of responses, that becomes apparent when different methods of M protein evaluation are applied. These results demonstrate that peripheral blood MRD testing by MS adds prognostic value with increased sensitivity. Although most patients had undetectable disease by standard immunofixation, the proportion of negative cases decreased substantially with the use of more sensitive techniques. In this analysis, all 3 protocols implemented intensive treatment with carfilzomib-based quadruplet and triplet regimens, yet only 36% reached negativity by LC-MS (only EXENT-negative samples were tested by LC-MS). These results suggest that even among patients treated with modern, highly active regimens, achieving LC-MS negativity remains a relatively uncommon event, emphasizing its potential role in identifying patients with the deepest responses. More importantly, the log-fold increases in the LoD of MS from EXENT to LC-MS were associated with superior PFS, similar to what has been shown previously for log-fold improvements in MRD 10⁻⁵ to 10⁻⁶ thresholds in the bone marrow.⁶ 

EXENT, although less sensitive than LC-MS, is an automated and validated analytical method with a short turnaround time, making it suitable for widespread implementation. Consistent with previous reports, there was ∼70% concordance between EXENT and 10⁻⁵ MRD, with a slightly higher prevalence of MS-positive/MRD-negative results compared with the reverse.^14,15,21 These findings differ from the recently published results of the Spanish group, which analyzed EXENT in comparison with bone marrow flow cytometry–based MRD with an LoD exceeding 10⁻⁵, in which the proportion of MRD-positive/MS-negative cases was slightly higher than MRD-negative/MS-positive cases.²² Importantly, our results demonstrate complementarity of the 2 modalities in their association with PFS (at both 10⁻⁵ and 10⁻⁶ thresholds), with patients with double negativity achieving longer PFS than those negative in only 1 assay. This finding reiterates our previous results and aligns with other reports.^15,22,23 

The longest PFS in our study was observed in patients who achieved LC-MS negativity. Notably, of the 4 PFS events reported in this group, only 2 were owing to myeloma progression, whereas the other 2 were attributed to deaths from secondary malignancies, with no evidence of disease progression. This highlights the need to account for competing risks in long-term survival analyses, as nonmyeloma-related mortality may become a more relevant factor in patients achieving deep responses. Overall, our results suggest that achieving LC-MS negativity may be associated with exceptionally long PFS after first-line treatment. This highlights LC-MS as a highly promising technique for assessing exceptionally deep responses. Its utility is limited at this time, because currently the results still require manual analysis despite automated sample handling. Further technical advancements in this technique are eagerly anticipated. Until then, LC-MS may be most applicable to patients who achieve MRD negativity at the 10⁻⁶ level in the bone marrow.

Our results also provide important insights into the kinetics of MRD evaluations using different methods related to M protein clearance. Although negative results by both EXENT and LC-MS were observed as early as after 4 cycles of induction therapy, the highest rates of negativity were reported later, at cycle 12 and at later time points. This aligns with previous reports and likely reflects the prolonged clearance of M protein, primarily because of immunoglobulin G recirculation.^24,25 

This analysis has some limitations. Firstly, the patients were not uniformly treated, as this is a pooled analysis of 3 different clinical trials with varying follow-up times. However, the patients’ characteristics were generally similar across the 3 studies and all patients received carfilzomib-based therapy. There were no between-group differences in the proportion of patients reaching MS negativity or in PFS. The survival analyses based on the best response may be susceptible to immortal time bias. To minimize this effect, we excluded patients with samples available only after cycle 4 from the analysis. As a result, there was only one PFS event within the first 12 months of observation, whereas most EXENT- and LC-MS–negative results occurred within the first 12 months of treatment. Another limitation is the currently relatively short median follow-up for this study (55.9 months), which limits our ability to predict PFS and OS at more extended time periods (ie, 8-10 years from initiation of therapy). Longer follow-up of patients included in this analysis will help to identify patients who may achieve a functional cure. Finally, positron emission tomography–computed tomography scans were not routinely performed in any of the 3 trials. Future studies on multimodal MRD evaluation should incorporate imaging techniques, particularly for patients with extramedullary disease.

Based on these results, MS-based measurement of residual disease in the peripheral blood may be incorporated into deep response assessment algorithms. Given the significant difference in PFS between patients with double negativity (EXENT and MRD by NGS) and those achieving only 1 negative result, a noninvasive, automated MS test may help guiding the decision to perform bone marrow evaluation only in EXENT-negative patients. In this analysis, omitting the bone marrow biopsy could have been considered for ∼15% of patients. Looking ahead, LC-MS could be incorporated into functional cure definitions, which are still in development.

In summary, the results pooled from these 3 prospective studies demonstrate that sensitive measurement of M protein using MS is a powerful method for response assessment in MM, with strong prognostic performance. In the era of MRD-driven clinical trials, efforts should be made to incorporate MS testing into future treatment designs as well.

The authors thank Gabriella Lakos and Emmanuel Okwelogu (Binding Site, part of Thermo Fisher Scientific, Rochester, New York) for their feedback on the manuscript text and study design.

This study was supported by Binding Site, part of Thermo Fisher Scientific.

Contribution: T.K., B.A.D., and A.J.J. designed the study; D.B. performed laboratory evaluations; T.K., B.A.D., and A.J.J. analyzed the results; T.K. made the figures; T.K., B.A.D., J.H.C., A.P., D.B., and A.J.J. drafted the first version of the manuscript; and all authors collected data and edited and approved the final version of the manuscript.

Conflict-of-interest disclosure: T.K. reports honoraria from Johnson & Johnson. B.A.D. reports research funding from Amgen and GlaxoSmithKline (GSK); reports consulting fees from Sanofi, Janssen, Canopy, and COTA; and is an independent reviewer of clinical trial for Bristol Myers Squibb. J.H.C.’s spouse is employed by AbbVie, Inc. A.P. reports honoraria from Janssen, Roche, and Amgen; and reports clinical trial support from Sanofi and GSK. D.B. is a current employee of Binding Site, part of Thermo Fisher Scientific. D.D. reports speaker honoraria and participation on advisory boards for Amgen and Celgene (Bristol Myers Squibb); and reports conference fees paid by Amgen. A.J.J. serves on advisory boards and consults with honoraria for AbbVie, Amgen, Bristol Myers Squibb, GSK, Janssen, and Sanofi-Aventis. K.J. declares no competing financial interests.

Correspondence: Andrzej J. Jakubowiak, The University of Chicago, 5841 S Maryland Ave, M/C 2115, Chicago, IL, 60637; email: ajakubowiak@bsd.uchicago.edu.