Результаты среди пациентов MM, получающих раннее вмешательство в процесс развития MRD после автономной трансплантации стволовых клеток.

Abbreviations

Autologous Stem Cell Transplantation

Biochemical Progression

Confidence interval

Clinical Progression

Hypercalcemia Renal failure Anemia, and Bone lesions

Early Intervention

Extramedullary Disease

Hazard ratio

High‐Risk Cytogenetic Abnormalities

International Myeloma Working Group

Lactate Dehydrogenase

Multiple myeloma

Minimal Residual Disease

MRD Progression

normal plasma cells

Overall Survival

progressive disease/progression

Progression‐Free Survival

Revised International Staging System

Time to Next Treatment

Ultra‐High‐Risk Cytogenetic Abnormalities

Introduction

Multiple myeloma (MM) is the second most common hematologic malignancy and originates from clonal plasma cells [ 1 ]. Despite advances in treatment, MM remains incurable, and nearly all patients with this malignancy eventually relapse after first‐line therapy. Over the past two decades, the definition of disease progression (PD) in MM has evolved because of the introduction of novel, highly sensitive detection methods. Traditionally, progression is defined as either an increase in monoclonal protein levels or the emergence of clinical symptoms [ 2 ]. However, the integration of minimal residual disease (MRD) assessment into the International Myeloma Working Group (IMWG) consensus criteria in 2016 resulted in refined response evaluations, thus offering a more accurate measure of residual disease burden during or after therapy [ 3 ]. Nevertheless, the current definition of PD in the era of MRD testing remains somewhat limited.

MRD negativity has emerged as one of the most powerful predictors of improved outcomes in patients with MM [ 4 , 5 ]. A large analysis [ 6 ] revealed that MRD negativity is associated with prolonged progression‐free survival (PFS) in both newly diagnosed MM patients and relapsed/refractory MM patients. Additionally, an expected but important observation was that the loss of MRD negativity often heralded a relapse [ 7 , 8 ]. Owing to the increasingly common use of MRD assessments, 30%–60% of patients will experience MRD progression without meeting the IMWG criteria for PD. The median time from MRD progression to PD has been reported to be between 1 and 3 years [ 9 , 10 , 11 , 12 ]. This evolving relapse pattern offers a unique window for early therapeutic intervention. As such, MRD‐driven strategies—such as initiating salvage therapy upon MRD reappearance—are being actively investigated in clinical trials [ 13 ]. While MRD assessment holds promise as a gateway to achieving an operational cure [ 14 ], there is a lack of evidence supporting standardized treatment approaches for MRD progression in the absence of clinical PD.

At present, there is no consensus regarding the optimal management of MRD progression before PD, and treatment decisions are often individualized in clinical practice [ 15 ]. To better understand the effects of different treatment strategies following MRD progression on patient outcomes, we conducted a retrospective analysis of data from our institutional database to examine MM patients who experienced MRD progression but did not meet the criteria for PD. The aim of this study was to explore the associations between treatment timing and outcomes following MRD progression, providing preliminary data to guide the design of future prospective trials.

Materials and Methods

Patient Selection and Definition

We conducted a retrospective analysis of 261 NDMM patients who achieved MRD negativity following initial treatment at our institution between December 2010 and January 2024. All patients received induction therapy with novel agents followed by autologous stem cell transplantation (ASCT) and maintenance therapy. Serial MRD assessments were performed using multiparametric flow cytometry at the end of induction, every three months during the first year post‐ASCT, and every 6 to 12 months thereafter, regardless of disease status or IMWG response. In patients without treatment intervention, repeat MRD testing occurred at variable intervals of 1–12 months. Among these patients, 97 experienced MRD progression (MRD‐P), including 46 who did not meet the IMWG criteria [ 3 ] for biochemical progression (BP) or clinical progression (CP). After 4 patients were excluded because of a lack of MRD data post‐MRD‐P or irregular follow‐up, a total of 42 patients were ultimately included for analysis. The decision to initiate treatment following MRD‐P was made by the treating physician and on the basis of personal preferences. This study was approved by the Ethics Committee of the First Affiliated Hospital of Sun Yat‐sen University (Approval Number: [2025]236) and was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants.

In accordance with the IMWG definitions, CP is defined as the development of new plasmacytomas or progression on the basis of the CRAB criteria. BP was defined as a measurable increase in serum or urine monoclonal protein that led to the initiation of a new line of therapy without meeting the criteria for CP. Patients were stratified into two groups on the basis of the timing of intervention: the early intervention (EI) group included patients who changed treatment at the time of MRD‐P without meeting the IMWG‐defined progression criteria, and the progressive disease (PD) group included patients who deferred changing treatment until biochemical or clinical progression occurred.

MRD And Molecular Risk Assessment

MRD was assessed in bone marrow aspirate samples using multiparameter flow cytometry (MFC) according to institutional protocols. Bone marrow aspirates (3–5 mL) were collected in EDTA anticoagulant tubes, and whenever feasible the first‐pull aspirate was preferentially used to minimize peripheral blood hemodilution. MRD was evaluated in BM samples (1.2 × 10 7 cells per sample). Samples were processed by ammonium chloride red blood cell lysis followed by washing with phosphate‐buffered saline before antibody staining. Because this retrospective study spanned a relatively long enrollment period, MRD detection methods evolved over time. Between 2010 and 2013, MRD was evaluated using a single‐tube 8‐color panel (cKappa/cLambda/CD138/CD117/CD56/CD19/CD38/CD45). Beginning in 2014, a standardized two‐tube 8‐color panel was implemented, consisting of tube 1 (cKappa/cLambda/CD19/CD20/CD138/CD45/CD38/CD56) and tube 2 (CD19/CD20/CD138/CD45/CD56/CD38/CD117/CD81). A limited number of early MRD assessments (5 patients, 21 tests) were performed using the single‐tube assay, which had an analytical sensitivity of 2 × 10 −5 . Although these patients were initially monitored using the earlier assay, MRD progression (conversion from MRD negativity to MRD positivity) in all five patients was detected after 2014 using the standardized two‐tube assay. For the standardized two‐tube 8‐color panel, flow cytometry data were analyzed using a sequential gating strategy including selection of stable events (CD45 vs. Time), removal of doublets (FSC‐Height vs. FSC‐Area and SSC‐Height vs. SSC‐Area), exclusion of debris (FSC vs. SSC), identification of plasma cells based on CD38 and CD138 expression, and identification of aberrant plasma cells according to abnormal expression of plasma cell‐associated markers combined with immunoglobulin light‐chain restriction. MRD levels were calculated as the proportion of clonal plasma cells among total leukocytes. At least 2 × 10 6 events were analyzed, and MRD negativity was defined using the standard criteria of 20 events of phenotypically aberrant clonal plasma cells. The sensitivity or the limit of detection (LOD) for this assay was validated to 20 cells in 2 × 10 6 events (1 × 10 −5 ), and the reproducibility or lower LOD is 50 cells in 2 × 10 6 events. MRD testing followed standardized laboratory procedures with routine instrument calibration, antibody validation, and internal quality control [ 3 ]. High‐risk cytogenetic abnormalities (HRCAs) included deletion of 17p and amplification of 1q, t(4;14), t(14;16), and t(14;20). The ultrahigh‐risk (UHRCA) patients had more than one HRCA.

MRD progression was defined as detection of ≥ 1 neoplastic plasma cell among 10 5 nucleated cells after a previously documented MRD‐negative status, confirmed by repeat MRD testing. MRD monitoring was performed after MRD‐P in routine clinical practice at variable intervals (approximately 1–12 months). MRD‐P was not defined based on a single MRD‐positive result in our study. For patients in the EI group, MRD conversion required two consecutive MRD‐positive results before treatment intervention was considered. Patients in the PD group underwent serial MRD monitoring (typically every 3–12 months) until clinical or biochemical progression occurred. Because of the retrospective study design, eight patients without intervention had repeat MRD testing at intervals of 1–12 months. All MRD progression cases required ≥ 2 MRD‐positive results confirming persistent MRD positivity and MRD levels at progression were well above the assay detection threshold in all cases. MRD kinetics could not be calculated because of rapid intervention timing and variable monitoring intervals.

Endpoints and Definitions

For patients with MRD‐P without concomitant disease progression (PD), we assessed the cumulative incidence of PD following MRD‐P. The clinical impact of early intervention was evaluated by comparing the time to next treatment (TTNT), overall survival (OS), and cumulative incidence of MRD re‐negativity between the two groups. The time to progression was defined as the interval from MRD‐P to the first IMWG‐defined biochemical or clinical progression. TTNT was defined as the interval from MRD‐P to second‐line treatment failure (defined as subsequent progression requiring third‐line therapy), death, or the last follow‐up. OS was defined as the time from MRD‐P to death or the last follow‐up. Treatment decisions in the absence of IMWG‐defined progression were made at the discretion of the treating physician. Therapeutic approaches include both intensive and low‐intensity regimens. Low‐intensity therapy involves adjustments to the maintenance regimen or a transition to doublet chemotherapy, whereas intensive therapy consists of three‐ or four‐drug regimens, with some patients undergoing salvage second‐line ASCT. The follow‐up deadline was December 2024.

Statistical Analysis

Statistical analyses were performed using SPSS software (version 26.0; IBM) or R version 4.1.2. Survival and the cumulative incidence of MRD re‐negativity were compared between groups by using the Kaplan–Meier method. The Cox proportional hazards model was used to compare patients in different groups and to estimate HRs and 95% confidence intervals (CIs). The data are expressed as the mean (± standard deviation) or median (interquartile range), as appropriate. Normally distributed continuous variables were compared using Student's t ‐test, and nonnormally distributed continuous variables were compared using the Mann–Whitney U test. Categorical variables were analyzed using the χ2 test or Fisher's exact test as appropriate. For all analyses, a p value < 0.05 was considered to indicate statistical significance.

Results

Baseline Characteristics of MM Patients With MRD ‐P

The baseline characteristics of the patients are presented in Table 1 . A total of 42 patients with MRD‐P were included in the final analysis. The median age at onset of MRD‐P was 56 years, and 59.5% of the patients were male. Among the patients, 45.3% had HRCAs, and 11.9% had 2+ HRCAs. All the patients received novel agent‐based induction therapy (92.9% triplet, 7.1% quadruplet regimens), and 7.1% underwent tandem ASCT.

Upon MRD‐P detection, 9 patients (21.4%) initiated immediate treatment (EI group), 25 patients (59.5%) deferred intervention until biochemical/clinical progression (PD group), and 8 patients were still under observation and did not alter their treatment. For analyses, the 8 patients under observation were included in the PD group and censored at the time of last follow‐up, as they represent a ‘watch‐and‐wait’ strategy similar to the PD group approach. Baseline characteristics were comparable between the EI and PD groups, with no significant differences in age, sex, HRCAs, UHRCAs, R‐ISS stage, or induction/maintenance therapies (Table 1 ). The percentage of residual normal plasma cells (NPCs) at initial MRD‐P detection was similar between the groups (median 92.4% vs. 93.0%, p = 0.904). However, at the time of treatment initiation, the NPC proportions differed significantly between groups: the EI group maintained high NPCs (median 90.2%), whereas the PD group presented markedly lower NPCs (median 7.3%, p < 0.001). This decline reflects progressive bone marrow infiltration in the PD group during the observation period.

Treatment Approaches and Outcomes in MM Patients With MRD ‐P

The treatment approach and MRD re‐negativity outcomes for patients with MRD‐P are shown in Figure 1 . The median follow‐up was 25.0 months. Low‐intensity therapy involves switching from lenalidomide or thalidomide to VD or KP regimens for maintenance, whereas intensive therapy involves switching from maintenance therapy to induction regimens such as PAD, Dara+KD/+VD/+RD/+PD, Dara+KPD, or KAD ± salvage ASCT.

A total of 42 patients were included, with 9 in the EI group receiving either low‐intensity therapy (6 patients) or intensive therapy (3 patients). In this group, 66.7% (4/6) of patients receiving low‐intensity therapy and 100% (3/3) of those receiving intensive therapy achieved MRD re‐negativity, but given the small sample sizes, the difference was not statistically significant ( p = 0.257). The remaining 33 patients (excluding 8 in the observation‐only group) deferred treatment changes until progression occurred. Among the 12 patients who progressed to BP, 5 received low‐intensity therapy, and 7 received intensive therapy. No patients in the low‐intensity therapy group achieved MRD re‐negativity (0/5), whereas 42.9% (3/7) of those receiving intensive therapy achieved MRD re‐negativity; there was a statistically significant difference between the therapy subgroups ( p = 0.047). All 13 patients with CP received intensive therapy, resulting in an MRD re‐negativity rate of 15.4% (2/13). Overall, the rate of MRD re‐negativity was significantly greater in the EI group than in the PD group (77.8% vs. 20.0%, p = 0.004). This heterogeneity in treatment approaches reflects real‐world clinical practice but introduces complexity in interpreting whether observed outcome differences result from intervention timing, specific therapeutic choices, or an interaction between both factors.

Prognosis and Survival in MM Patients With MRD ‐P

The median time from initial MRD‐P detection to first progression (biochemical or clinical) was 10.3 months (95% CI: 7.2–15.8 months) for all patients. Notably, compared with those who waited, patients who received early intervention experienced a significantly longer time from MRD‐P to progression: 34.8 months (95% CI: 15.2 months‐not reached) in the EI group vs. 6.6 months (95% CI: 5.6–22.9 months) in the PD group ( p = 0.029) (Figure 2A ). Among the 25 patients in the PD group who eventually progressed, 12 progressed to biochemical progression (median 5.7 months), and 13 progressed to clinical progression (median 10.8 months). Among those with clinical progression, plasmacytoma (extramedullary disease) was observed in 3 patients (23.1%). The median time from MRD‐P to plasmacytoma‐associated CP was 22.9 months, with a median overall survival of 41.0 months. No significant survival difference was observed compared with patients without plasmacytoma ( p = 0.473), likely due to the limited sample size. All patients received intensive therapy, and MRD re‐negativity was achieved in 1 of 3 patients (33.3%). Additionally, the median duration from MRD‐P to second‐line treatment failure was 34.8 months in the EI group (95% CI: 34.8 months), whereas the median duration in the PD group was 25.6 months (95% CI: 15.6 months). This difference was not statistically significant ( p = 0.14) (Figure 2B ). In terms of overall survival, compared with the PD group, the EI group showed a clear advantage, with the median OS not yet reached (95% CI: 56.8 months‐not reached), with a median OS of 41.0 months (95% CI: 23.9 months‐not reached) ( p = 0.034) (Figure 2C ). Moreover, the cumulative incidence of MRD re‐negativity was significantly greater in the EI group than in the PD group ( p < 0.001) (Figure 2D ). In the EI group, the median time to achieve MRD re‐negativity was 12.6 months (95% CI: 10.3 months). In the PD group, the median time to MRD re‐negativity was 75.9 months (95% CI: 75.9 months).

The analysis revealed several factors influencing MRD re‐negativity following MRD‐P in MM patients (Table 2 ). Both univariate and multivariate analyses revealed that UHRCA at diagnosis and early intervention following MRD‐P were significant predictors of MRD re‐negativity. Specifically, early intervention significantly improved MRD re‐negativity, with a hazard ratio (HR) of 0.07 (95% CI: 0.01–0.36; p = 0.001) in the univariate analysis and HR = 0.08 (95% CI: 0.01–0.51; p = 0.007) in the multivariate analysis. Conversely, patients with UHRCA at diagnosis were less likely to achieve MRD re‐negativity, with an HR of 9.84 (95% CI: 2.10–46.20; p = 0.004) in the univariate analysis and an HR = 9.11 (95% CI: 1.45–57.32; p = 0.019) in the multivariate analysis. While NPC percentage < 80% at relapse was initially associated with MRD re‐negativity in the univariate analysis (HR = 3.45; 95% CI: 1.03–11.54; p = 0.045), this association did not remain significant in the multivariate analysis (HR = 1.07; 95% CI: 0.26–4.44; p = 0.923). Furthermore, the intensity of treatment after MRD‐P did not significantly affect MRD re‐negativity (HR = 0.94; 95% CI: 0.27–3.19; p = 0.917).

Discussion

In patients with MM, second‐line treatment is typically initiated upon meeting the IMWG criteria for PD [ 16 ]. However, with the increasing use of MRD assessments in clinical practice, it has become clear that MRD‐P often occurs before PD [ 12 , 15 ]. Previous studies have shown that the loss of MRD negativity frequently precedes relapse. However, there is a lack of clarity regarding the need for and optimal timing of another line of treatment once MRD positivity emerges.

Is salvage therapy necessary upon the loss of MRD negativity in our pursuit of surgical cures for MM? MRD‐driven therapeutic strategies are being increasingly explored in clinical trials. For example, the REMNANT study [ 17 ], a Norwegian randomized phase 2/3 trial involving 391 transplant‐eligible (TE) patients, is assessing whether treating MRD progression after first‐line therapy prolongs PFS and OS compared with treating relapse after disease progression. However, definitive data are still lacking, as such data are difficult to obtain in routine clinical practice. Nevertheless, our center has been at the forefront of dynamic MRD monitoring in MM patients receiving novel‐agent induction therapy, ASCT, and maintenance therapy in China [ 18 , 19 ]. Therefore, our data offer real‐world insights into the different patterns of first relapse. This study investigated the clinical implications of early intervention at the onset of MRD‐P in MM patients following ASCT.

In our study, 46 out of 97 (47.4%) patients experienced MRD‐P before they met the criteria for PD. These findings are consistent with the results of two simultaneously published analyses. In the FORTE study ( N = 474) [ 9 ], 70 patients experienced MRD resurgence, with a median of 22.3 months from MRD resurgence to PD, whereas 22 patients experienced MRD progression before meeting the IMWG criteria for PD. Similarly, the GEM2012MENOS65/GEM2014MAIN study ( N = 458) [ 11 ] reported that 71 patients experienced MRD resurgence, with a median of 39 months to PD, and 31 patients experienced MRD progression before PD. The period between MRD resurgence and disease progression presents an opportunity for testing early intervention strategies.

Our findings indicate that 9 patients started treatment immediately (EI group) upon MRD‐P. A comparison of these two groups revealed that patients in the EI group had a significantly longer time to first PD, resulting in notable improvements in OS. Additionally, there was a trend toward a longer TTNT in the EI group. MRD is considered the root cause of recurrence and relapse; therefore, it is essential to achieve and maintain an MRD‐negative status in the pursuit of a functional cure. Furthermore, it is critical to reestablish MRD negativity after MRD positivity, as this significantly affects patient outcomes. MRD re‐negativity rates were higher in the EI group, with multivariate analysis suggesting that early intervention may be associated with deeper, more durable responses and offering further support for the potential of a tumor cure.

Notably, clinical progression may manifest with biologically aggressive features such as extramedullary disease. In our cohort, 3 of 13 patients (23.1%) with clinical progression developed extramedullary disease (plasmacytoma). Plasmacytoma‐associated relapse is generally considered a high‐risk feature, reflecting aggressive disease biology and therapeutic resistance. However, no statistically significant difference in survival was observed compared with patients without plasmacytoma, likely due to the limited sample size. Larger prospective studies are warranted to better define the clinical impact of plasmacytoma in the context of MRD progression.

There are limited data on the intensity and choice of treatment regimens for MRD‐P [ 12 , 20 ]. The ideal strategy is to avoid overly intensive treatment to prevent toxicity while also avoiding overly weak treatment to prevent inefficacy and resistance. Finding this balance is challenging, and these findings underscore the need for standardized interventions in the face of MRD resurgence. In our study, some patients adjusted their maintenance therapy regimen by switching from immunomodulatory drugs to bortezomib or to a dual‐drug combination. Other patients received more intensive treatment, including triple‐ or quadruple‐drug regimens or sequential salvage ASCT. Our results showed that in the biochemical relapse group, treatment intensity affected prognosis. The overall treatment intensity in the early intervention group was lower than that in the PD group, but the outcomes were superior in the EI group. Within the EI group, the MRD re‐negativity rates were 66.7% (2/3) with low‐intensity regimens and 100% (3/3) with intensive regimens, although these subgroups were too small for meaningful statistical comparison. As shown in the MASTER study [ 21 ], some patients experienced MRD resurgence after treatment was stopped, but five participants who received combination therapy with lenalidomide and daratumumab experienced subsequent MRD reduction or stabilization. These findings suggest that in some patients, simply modifying their treatment regimen may be sufficient to achieve MRD negativity.

However, these findings must be interpreted cautiously. Treatment selection was influenced by physician assessment of disease characteristics, patient fitness, and prior responses, introducing confounding by indication. Patients in the EI group may have had better performance status and treatment tolerance, explaining their response to less intensive regimens, whereas patients in the “watch‐and‐wait” group may have had factors unsuitable for immediate intensification. We cannot determine whether early intervention benefits are independent of treatment intensity or whether specific regimens are uniquely effective when given early. These preliminary observations raise the hypothesis that early intervention may be associated with favorable outcomes across different treatment intensities, but adequately powered randomized trials with standardized treatment protocols are essential for identifying optimal regimen selection strategies.

We also acknowledge that early initiation of treatment and varying treatment intensities may not impact survival in all patients, and further research is needed to identify the characteristics of patients who will benefit from early treatment. These questions can be addressed only through a carefully designed prospective study. However, our data provide some valuable insights. Several factors were found to influence the likelihood of achieving MRD re‐negativity. The presence of HRCAs at diagnosis, particularly UHRCAs, is associated with a lower likelihood of MRD re‐negativity. For such high‐risk populations, more intensive and earlier treatment following MRD‐P may offer an opportunity to further improve prognosis [ 22 ]. Additionally, residual NPCs during the MRD phase play a critical role in MM. Previous studies have shown that myeloma patients with a high baseline NPC ratio present distinct indolent features, such as a low tumor burden. Furthermore, a high residual NPC ratio at diagnosis or relapse was independently associated with favorable survival outcomes [ 8 , 23 ]. In our study, while NPC proportions were comparable at MRD‐P detection (EI: 92.40% vs. PD: 93.00%, p = 0.904), the PD group experienced a significant decrease in NPC proportion during observation (median 7.30% at treatment initiation, p < 0.001), reflecting progressive marrow infiltration. Progressive NPC loss may impair bone marrow function, immune recovery, and treatment tolerance, potentially contributing to poor outcomes with delayed intervention. However, we cannot determine whether early intervention prevents NPC decline or whether patients with slower NPC decline should be selected for observation. Future studies should incorporate serial NPC monitoring and kinetics as prognostic markers and potential intervention triggers.

Prospective randomized trials are essential to definitively establish the role of early MRD‐driven intervention. Our retrospective findings provide hypothesis‐generating data supporting ongoing trials. The REMNANT study is specifically designed to determine whether, compared with deferring treatment until clinical progression, treating with MRD‐P after first‐line therapy prolongs PFS and OS. Our real‐world data demonstrate the feasibility and potential signals of benefit, reinforcing the importance of prospective validation. Only through rigorously designed randomized trials can the true efficacy, optimal timing, and appropriate patient selection for MRD‐driven interventions be definitively established.

We emphasize that our findings do not justify routine early intervention in all patients with MRD‐P. Treatment decisions should remain individualized on the basis of comprehensive clinical assessment until definitive evidence emerges. The current “watch‐and‐wait” approach has a valid rationale: early treatment initiation increases cumulative drug exposure, treatment costs, adverse event risk, and potential impact on quality of life. For patients with slowly rising or low‐level MRD, favorable cytogenetics, and preserved bone marrow function, deferring treatment may be appropriate and spare unnecessary therapy during sustained remission. The optimal approach likely requires individualized risk stratification on the basis of MRD kinetics (rate of rise, doubling time), the cytogenetic risk profile, the residual disease burden (MRD level, normal plasma cell proportion), and patient‐specific factors (age, comorbidities, treatment tolerance, preferences). Our study suggests that early intervention warrants further investigation in well‐designed trials that assess not only survival outcomes but also toxicity, quality of life, and cost‐effectiveness to fully inform clinical decision‐making.

Despite the valuable insights from this study, several important limitations must be acknowledged. First, the retrospective, nonrandomized design introduces inevitable selection bias, as treatment decisions were based on physician and patient preferences rather than randomization. Patients with aggressive MRD‐P features may have been preferentially assigned to early intervention, while those with indolent disease were more likely to be observed. Although baseline characteristics—such as HRCAs, UHRCAs, R‐ISS stage, and normal plasma cell proportions at MRD‐P detection—showed no significant differences between groups, critical unmeasured confounders remained, particularly MRD kinetics, absolute MRD burden at detection, and clinical judgment factors. These unmeasured variables may have influenced both treatment assignment and outcomes independently, precluding causal inference. Second, the limited cohort size ( N = 42) and highly imbalanced distribution (9 in the EI group vs. 33 in the PD group) resulted in wide confidence intervals, insufficient statistical power to adjust for multiple confounders simultaneously, and statistically fragile findings. Subgroup analyses are too small for meaningful interpretation, and validation in larger studies is essential. Third, treatment regimens varied widely, ranging from doublet maintenance modifications to intensive multidrug combinations with salvage ASCT. This heterogeneity, compounded by treatment selection on the basis of unmeasured disease and patient factors, introduces confounding by indication and prevents the isolation of timing effects from specific therapeutic choices. Fourth, MRD detection methods and monitoring schedules evolved during the study period. Early MRD assessments used a single‐tube assay with slightly lower sensitivity, whereas a standardized two‐tube assay was implemented in later years, and MRD progression events in these patients were detected using the standardized method. Variable MRD monitoring intervals (1–12 months) limited precise determination of MRD‐P timing and precluded systematic assessment of MRD kinetics, and the sensitivity limit of flow cytometry (1 × 10 −5 ) may have missed low‐level disease. However, because MRD progression required repeated MRD positivity, the likelihood that MRD‐P events reflected assay variability is minimized. Finally, the study did not use predefined risk‐stratified treatment approaches based on emerging prognostic markers such as MRD kinetic parameters, MGUS‐like characteristics [ 24 , 25 ], or clonal evolution patterns, which could significantly impact treatment response. Given these limitations, our findings should be considered hypothesis‐generating rather than practice‐changing, and prospective randomized controlled trials with standardized protocols, adequate sample sizes, and comprehensive assessments of not only survival outcomes but also toxicity and quality of life are essential to definitively establish the role of early MRD‐driven intervention.

In conclusion, this retrospective study provides preliminary evidence suggesting the potential benefits of early intervention in patients with MRD‐P. Our data indicate that initiating treatment at MRD‐P onset was associated with prolonged survival and higher rates of MRD re‐negativity than deferring treatment until PD. However, these findings must be interpreted cautiously. We emphasize that these results do not justify routine early intervention in all patients with MRD‐P; rather, treatment decisions should remain individualized on the basis of a comprehensive risk assessment incorporating MRD kinetics, the cytogenetic profile, disease burden, and patient‐specific factors. Further prospective randomized controlled trials, such as the ongoing REMNANT study, are essential for establishing evidence‐based protocols for managing MRD‐P and identifying which patient subgroups benefit most from early intervention, ultimately paving the way for more effective and personalized treatment strategies in multiple myeloma patients.

Author Contributions

Xiaozhe Li: data curation (equal), formal analysis (equal), funding acquisition (equal), investigation (equal), methodology (equal), project administration (equal), resources (equal), software (equal), supervision (equal), validation (equal), visualization (equal), writing – original draft (equal), writing – review and editing (equal). Meilan Chen: investigation (equal), resources (equal). Lifen Kuang: investigation (equal), resources (equal). Junru Liu: investigation (equal), resources (equal). Beihui Huang: investigation (equal), resources (equal). Juan Li: conceptualization (equal), data curation (equal), formal analysis (equal), funding acquisition (equal), investigation (equal), methodology (equal), project administration (equal), resources (equal), software (equal), supervision (equal), validation (equal), visualization (equal), writing – original draft (equal), writing – review and editing (equal).

Funding

This study was supported by the National Natural Science Foundation Youth Science Foundation (No. 82200221), Guangzhou Municipal Science and Technology Project (2025A04J5219), and the National Natural Science Foundation (No. 82470197 and No. 82270209).

Disclosure

No permission was required, as all figures, tables, and content in this manuscript are original and created by the authors. No copyrighted material from other sources has been reproduced.

Ethics Statement

This study was approved by the Ethics Committee of the First Affiliated Hospital of Sun Yat‐sen University (Approval Number: [2025]236) and was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants or their legal guardians before inclusion in the study.

Consent

Written consent was obtained from all patients or their legal guardians, ensuring their voluntary participation in the study. The authors affirm that all patients included in this study provided written consent for the publication of their data and findings in this manuscript. All patient information was anonymized to protect privacy in accordance with ethical guidelines.

Conflicts of Interest

The authors declare no conflicts of interest.

Acknowledgements

We would like to express our sincere gratitude to the patients who participated in this study, as well as the medical and research staff at The First Affiliated Hospital of Sun Yat‐sen University for their invaluable assistance in data collection and patient management.

Data Availability Statement

The data generated in this study are available upon request from the corresponding author.

References

1 N. Van de Donk , C. Pawlyn , and K. L. Yong , “ Multiple Myeloma ,” Lancet 397 ( 2021 ): 410 – 427 . 33516340 10.1016/S0140-6736(21)00135-5

2 S. Goldman‐Mazur , A. Visram , P. Kapoor , et al., “ Outcomes After Biochemical or Clinical Progression in Patients With Multiple Myeloma ,” Blood Advances 7 ( 2023 ): 909 – 917 . 35413102 10.1182/bloodadvances.2022007082 PMC10025108

3 S. Kumar , B. Paiva , K. C. Anderson , et al., “ International Myeloma Working Group Consensus Criteria for Response and Minimal Residual Disease Assessment in Multiple Myeloma ,” Lancet. Oncology 17 ( 2016 ): e328 – e346 . 27511158 10.1016/S1470-2045(16)30206-6

4 M. Mohty , H. Avet‐Loiseau , F. Malard , and J. L. Harousseau , “ Potential Future Direction of Measurable Residual Disease Evaluation in Multiple Myeloma ,” Blood 142 ( 2023 ): 1509 – 1517 . 37471603 10.1182/blood.2023020284

5 C. Wijnands , S. Noori , N. Donk , M. M. VanDuijn , and J. F. M. Jacobs , “ Advances in Minimal Residual Disease Monitoring in Multiple Myeloma ,” Critical Reviews in Clinical Laboratory Sciences 60 ( 2023 ): 518 – 534 . 37232394 10.1080/10408363.2023.2209652

6 M. Cavo , J. San‐Miguel , S. Z. Usmani , et al., “ Prognostic Value of Minimal Residual Disease Negativity in Myeloma: Combined Analysis of POLLUX, CASTOR, ALCYONE, and MAIA ,” Blood 139 ( 2022 ): 835 – 844 . 34289038 10.1182/blood.2021011101 PMC8832474

7 M. Mohan , S. Kendrick , A. Szabo , et al., “ Clinical Implications of Loss of Bone Marrow Minimal Residual Disease Negativity in Multiple Myeloma ,” Blood Advances 6 ( 2022 ): 808 – 817 . 34807986 10.1182/bloodadvances.2021005822 PMC8945288

8 J. Martinez‐Lopez , N. Lopez‐Muñoz , A. Chari , et al., “ Measurable Residual Disease (MRD) Dynamics in Multiple Myeloma and the Influence of Clonal Diversity Analyzed by Artificial Intelligence ,” Blood Cancer Journal 14 ( 2024 ): 131 . 39112458 10.1038/s41408-024-01102-x PMC11306767

9 C. Guerrero , N. Puig , M. T. Cedena , et al., “ Predictors of Unsustained Measurable Residual Disease Negativity in Transplant‐Eligible Patients With Multiple Myeloma ,” Blood 143 ( 2024 ): 597 – 603 . 38048552 10.1182/blood.2023022083

10 S. Kumar , “ Don't Let the Genie Out of the Bottle! ,” Blood 143 ( 2024 ): 564 – 566 . 38358852 10.1182/blood.2023023293 PMC10873532

11 M. D'Agostino , G. Bertuglia , D. Rota‐Scalabrini , et al., “ Predictors of Unsustained Measurable Residual Disease Negativity in Patients With Multiple Myeloma ,” Blood 143 ( 2024 ): 592 – 596 . 38048557 10.1182/blood.2023022080

12 L. J. Costa , E. Medvedova , B. Dhakal , et al., “ Implications of Isolated MRD Progression in Newly Diagnosed Multiple Myeloma Treated With Quadruplet Therapy ,” Blood Cancer Journal 15 ( 2025 ): 13 . 39910064 10.1038/s41408-025-01219-7 PMC11799218

13 M. Meseha , J. Hoffman , D. Kazandjian , O. Landgren , and B. Diamond , “ Minimal Residual Disease‐Adapted Therapy in Multiple Myeloma: Current Evidence and Opinions ,” Current Oncology Reports 26 ( 2024 ): 679 – 690 . 38676789 10.1007/s11912-024-01537-2 PMC11169024

14 O. Landgren and D. Kazandjian , “ Modern Myeloma Therapy + Sustained Minimal Residual Disease‐Negative = (Functional) Cure! ,” Journal of Clinical Oncology 40 ( 2022 ): 2863 – 2866 . 35772041 10.1200/JCO.22.00622 PMC9426854

15 R. Szalat , K. Anderson , and N. Munshi , “ Role of Minimal Residual Disease Assessment in Multiple Myeloma ,” Haematologica 109 ( 2024 ): 2049 – 2059 . 38328864 10.3324/haematol.2023.284662 PMC11215375

16 J. Laubach , L. Garderet , A. Mahindra , et al., “ Management of Relapsed Multiple Myeloma: Recommendations of the International Myeloma Working Group ,” Leukemia 30 ( 2016 ): 1005 – 1017 . 26710887 10.1038/leu.2015.356

17 F. B. Askeland , A. M. Rasmussen , and F. Schjesvold , “ Relapse From MRD Negativity as Indication for Treatment in Multiple Myeloma ‐ The Remnant Study ,” Blood 136 ( 2020 ): 21 – 22 .

18 Q. Sun , X. Li , J. Gu , et al., “ Prognostic Significance of the Stage at Which an MRD‐Negative Status Is Achieved for Patients With Multiple Myeloma Who Received ASCT ,” Frontiers in Oncology 12 ( 2022 ): 776920 . 35664799 10.3389/fonc.2022.776920 PMC9157574

19 J. Gu , J. Liu , M. Chen , B. Huang , and J. Li , “ Longitudinal Flow Cytometry Identified ‘Minimal Residual Disease’ (MRD) Evolution Patterns for Predicting the Prognosis of Patients With Transplant‐Eligible Multiple Myeloma ,” Biology of Blood and Marrow Transplantation 24 ( 2018 ): 2568 – 2574 . 30142420 10.1016/j.bbmt.2018.07.040

20 B. A. Derman , “ MRD‐Guided Treatment Cessation in Multiple Myeloma ,” Lancet. Haematology 10 ( 2023 ): e867 – e868 . 37776871 10.1016/S2352-3026(23)00240-5

21 L. J. Costa , S. Chhabra , E. Medvedova , et al., “ Minimal Residual Disease Response‐Adapted Therapy in Newly Diagnosed Multiple Myeloma (MASTER): Final Report of the Multicentre, Single‐Arm, Phase 2 Trial ,” Lancet. Haematology 10 ( 2023 ): e890 – e901 . 37776872 10.1016/S2352-3026(23)00236-3 PMC10836587

22 Y. Tao , S. Jin , D. Yang , et al., “ Real‐World Advantage and Challenge of Post‐Autologous Stem Cell Transplantation MRD Negativity in High‐Risk Patients With Double‐Hit Multiple Myeloma ,” BMC Cancer 24 ( 2024 ): 406 . 38565996 10.1186/s12885-024-12077-0 PMC10985970

23 W. Yan , L. Shi , J. Xu , et al., “ Clinical Implications of Residual Normal Plasma Cells Within Bone Marrow Across Various Disease Stages in Multiple Myeloma ,” Leukemia 38 ( 2024 ): 2235 – 2245 . 39095502 10.1038/s41375-024-02366-9

24 A. Gozzetti , P. Pacelli , D. Raspadori , et al., “ Long‐Term CR Multiple Myeloma Patients Show Cured or MGUS‐Like Minimal Residual Disease Pattern by Next Generation Flow ,” Reviews on Recent Clinical Trials 17 ( 2022 ): 92 – 96 . 35578845 10.2174/1574887117666220516145628

25 W. Yan , J. Zhou , Y. Liu , et al., “ The Monoclonal Gammopathy of Undetermined Significance‐Like (MGUS‐Like) Classifier Stratifies Prognosis in Multiple Myeloma: Results From Real‐World Data in Chinese Population ,” Clinical Lymphoma, Myeloma & Leukemia 25 ( 2025 ): e452–e456.e4 . 10.1016/j.clml.2025.03.002 40157884