[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Figure 1.
Article Identification and Selection
Article Identification and Selection

AlloSCT indicates allogeneic stem cell transplantation; MRD, minimal residual disease; RRMM, relapsed/refractory multiple myeloma.

Figure 2.
Overall Effect of MRD Status on PFS and OS
Overall Effect of MRD Status on PFS and OS

A, Tests for heterogeneity indicated no significant differences between the studies for both PFS and B, OS. C, Kaplan-Meier curves for PFS and D, OS. Data were adjusted to account for the different proportions of patients in each study being MRD-positive and MRD-negative. The sizes of the forest plot squares are proportional to the amount of information each trial contains (the inverse of the variance); the horizontal lines represent the 95% CIs. CI indicates confidence interval; MRD, minimal residual disease; OS, overall survival; PFS, progression-free survival.

Figure 3.
In CR Patients, Effect of MRD Status on PFS and OS
In CR Patients, Effect of MRD Status on PFS and OS

A and C, Forest plots and survival curves for patients in CR showing the overall effect of MRD status on PFS. B and D, forest plots and survival curves for patients in CR showing the overall effect of MRD status on OS. MRD-negative patients had better outcomes. Tests for heterogeneity indicated no significant differences between the studies for both PFS and OS. Kaplan-Meier curves for PFS (C) and OS (D); data were adjusted to account for the different proportions of patients in each study being MRD-positive and MRD-negative. The sizes of the Forest plot squares are proportional to the amount of information each trial contains (the inverse of the variance); the horizontal lines represent the 95% CIs. CI, confidence interval; MRD, minimal residual disease; OS, overall survival; PFS, progression-free survival.

1.
Munshi  NC, Anderson  KC.  Minimal residual disease in multiple myeloma.  J Clin Oncol. 2013;31(20):2523-2526.PubMedGoogle ScholarCrossref
2.
Anderson  KC.  Oncogenomics to target myeloma in the bone marrow microenvironment.  Clin Cancer Res. 2011;17(6):1225-1233.PubMedGoogle ScholarCrossref
3.
Mailankody  S, Korde  N, Lesokhin  AM,  et al.  Minimal residual disease in multiple myeloma: bringing the bench to the bedside.  Nat Rev Clin Oncol. 2015;12(5):286-295.PubMedGoogle ScholarCrossref
4.
Biran  N, Ely  S, Chari  A.  Controversies in the assessment of minimal residual disease in multiple myeloma: clinical significance of minimal residual disease negativity using highly sensitive techniques.  Curr Hematol Malig Rep. 2014;9(4):368-378.PubMedGoogle ScholarCrossref
5.
Rawstron  AC, de Tute  RM, Haughton  J, Owen  RG.  Measuring disease levels in myeloma using flow cytometry in combination with other laboratory techniques: Lessons from the past 20 years at the Leeds Haematological Malignancy Diagnostic Service.  Cytometry B Clin Cytom. 2016;90(1):54-60.PubMedGoogle ScholarCrossref
6.
Rawstron  AC, Child  JA, de Tute  RM,  et al.  Minimal residual disease assessed by multiparameter flow cytometry in multiple myeloma: impact on outcome in the Medical Research Council Myeloma IX Study.  J Clin Oncol. 2013;31(20):2540-2547.PubMedGoogle ScholarCrossref
7.
Rajkumar  SV, Harousseau  JL, Durie  B,  et al; International Myeloma Workshop Consensus Panel 1.  Consensus recommendations for the uniform reporting of clinical trials: report of the International Myeloma Workshop Consensus Panel 1.  Blood. 2011;117(18):4691-4695.PubMedGoogle ScholarCrossref
8.
Swedin  A, Lenhoff  S, Olofsson  T, Thuresson  B, Westin  J.  Clinical utility of immunoglobulin heavy chain gene rearrangement identification for tumour cell detection in multiple myeloma.  Br J Haematol. 1998;103(4):1145-1151.PubMedGoogle ScholarCrossref
9.
Gregory  WM.  Adjusting survival curves for imbalances in prognostic factors.  Br J Cancer. 1988;58(2):202-204.PubMedGoogle ScholarCrossref
10.
Peto  R.  Why do we need systematic overviews of randomized trials?  Stat Med. 1987;6(3):233-244.PubMedGoogle ScholarCrossref
11.
Whitehead  A, Whitehead  J.  A general parametric approach to the meta-analysis of randomized clinical trials.  Stat Med. 1991;10(11):1665-1677.PubMedGoogle ScholarCrossref
12.
Roussel  M, Lauwers-Cances  V, Robillard  N,  et al.  Front-line transplantation program with lenalidomide, bortezomib, and dexamethasone combination as induction and consolidation followed by lenalidomide maintenance in patients with multiple myeloma: a phase II study by the Intergroupe Francophone du Myélome.  J Clin Oncol. 2014;32(25):2712-2717.PubMedGoogle ScholarCrossref
13.
Bakkus  MH, Bouko  Y, Samson  D,  et al.  Post-transplantation tumour load in bone marrow, as assessed by quantitative ASO-PCR, is a prognostic parameter in multiple myeloma.  Br J Haematol. 2004;126(5):665-674.PubMedGoogle ScholarCrossref
14.
Dal Bó  S, Pezzi  A, Amorin  B, Valim  V, Isabel Bittencourt  R, Silla  L.  Detection of minimal residual disease by flow cytometry for patients with multiple myeloma submitted to autologous hematopoietic stem cell transplantation.  ISRN Hematol. 2013;2013:847672. doi:10.1155/2013/847672PubMedGoogle ScholarCrossref
15.
Davies  FE, Forsyth  PD, Rawstron  AC,  et al.  The impact of attaining a minimal disease state after high-dose melphalan and autologous transplantation for multiple myeloma.  Br J Haematol. 2001;112(3):814-819.PubMedGoogle ScholarCrossref
16.
Ferrero  S, Ladetto  M, Drandi  D,  et al.  Long-term results of the GIMEMA VEL-03-096 trial in MM patients receiving VTD consolidation after ASCT: MRD kinetics’ impact on survival.  Leukemia. 2015;29(3):689-695.PubMedGoogle ScholarCrossref
17.
Korthals  M, Sehnke  N, Kronenwett  R,  et al.  Molecular monitoring of minimal residual disease in the peripheral blood of patients with multiple myeloma.  Biol Blood Marrow Transplant. 2013;19(7):1109-1115.PubMedGoogle ScholarCrossref
18.
Korthals  M, Sehnke  N, Kronenwett  R,  et al.  The level of minimal residual disease in the bone marrow of patients with multiple myeloma before high-dose therapy and autologous blood stem cell transplantation is an independent predictive parameter.  Biol Blood Marrow Transplant. 2012;18(3):423-431.e3.PubMedGoogle ScholarCrossref
19.
Ludwig  H, Greil  R, Masszi  T,  et al.  Bortezomib, thalidomide and dexamethasone, with or without cyclophosphamide, for patients with previously untreated multiple myeloma: 5-year follow-up.  Br J Haematol. 2015;171(3):344-354.PubMedGoogle ScholarCrossref
20.
Ludwig  H, Viterbo  L, Greil  R,  et al.  Randomized phase II study of bortezomib, thalidomide, and dexamethasone with or without cyclophosphamide as induction therapy in previously untreated multiple myeloma.  J Clin Oncol. 2013;31(2):247-255.PubMedGoogle ScholarCrossref
21.
Martinez-Lopez  J, Fernández-Redondo  E, García-Sánz  R,  et al; GEM (Grupo Español Multidisciplinar de Melanoma)/PETHEMA (Programa para el Estudio de la Terapéutica en Hemopatías Malignas) cooperative study group.  Clinical applicability and prognostic significance of molecular response assessed by fluorescent-PCR of immunoglobulin genes in multiple myeloma. Results from a GEM/PETHEMA study.  Br J Haematol. 2013;163(5):581-589.PubMedGoogle ScholarCrossref
22.
Martinez-Lopez  J, Lahuerta  JJ, Pepin  F,  et al.  Prognostic value of deep sequencing method for minimal residual disease detection in multiple myeloma.  Blood. 2014;123(20):3073-3079.PubMedGoogle ScholarCrossref
23.
Martínez-Sánchez  P, Montejano  L, Sarasquete  ME,  et al.  Evaluation of minimal residual disease in multiple myeloma patients by fluorescent-polymerase chain reaction: the prognostic impact of achieving molecular response.  Br J Haematol. 2008;142(5):766-774.PubMedGoogle ScholarCrossref
24.
Paiva  B, Vidriales  MB, Cerveró  J,  et al; GEM (Grupo Español de MM)/PETHEMA (Programa para el Estudio de la Terapéutica en Hemopatías Malignas) Cooperative Study Groups.  Multiparameter flow cytometric remission is the most relevant prognostic factor for multiple myeloma patients who undergo autologous stem cell transplantation.  Blood. 2008;112(10):4017-4023.PubMedGoogle ScholarCrossref
25.
Paiva  B, Martinez-Lopez  J, Vidriales  MB,  et al.  Comparison of immunofixation, serum free light chain, and immunophenotyping for response evaluation and prognostication in multiple myeloma.  J Clin Oncol. 2011;29(12):1627-1633.PubMedGoogle ScholarCrossref
26.
Paiva  B, Gutiérrez  NC, Rosiñol  L,  et al; PETHEMA/GEM (Programa para el Estudio de la Terapéutica en Hemopatías Malignas/Grupo Español de Mieloma) Cooperative Study Groups.  High-risk cytogenetics and persistent minimal residual disease by multiparameter flow cytometry predict unsustained complete response after autologous stem cell transplantation in multiple myeloma.  Blood. 2012;119(3):687-691.PubMedGoogle ScholarCrossref
27.
Puig  N, Sarasquete  ME, Balanzategui  A,  et al.  Critical evaluation of ASO RQ-PCR for minimal residual disease evaluation in multiple myeloma. A comparative analysis with flow cytometry.  Leukemia. 2014;28(2):391-397.PubMedGoogle ScholarCrossref
28.
Rawstron  AC, Davies  FE, DasGupta  R,  et al.  Flow cytometric disease monitoring in multiple myeloma: the relationship between normal and neoplastic plasma cells predicts outcome after transplantation.  Blood. 2002;100(9):3095-3100.PubMedGoogle ScholarCrossref
29.
San Miguel  JF, Almeida  J, Mateo  G,  et al.  Immunophenotypic evaluation of the plasma cell compartment in multiple myeloma: a tool for comparing the efficacy of different treatment strategies and predicting outcome.  Blood. 2002;99(5):1853-1856.PubMedGoogle ScholarCrossref
30.
Sarasquete  ME, García-Sanz  R, González  D,  et al.  Minimal residual disease monitoring in multiple myeloma: a comparison between allelic-specific oligonucleotide real-time quantitative polymerase chain reaction and flow cytometry.  Haematologica. 2005;90(10):1365-1372.PubMedGoogle Scholar
31.
Fukumoto  K, Fujisawa  M, Suehara  Y,  et al.  Prognostic impact of immunophenotypic complete response in patients with multiple myeloma achieving better than complete response.  Leuk Lymphoma. 2016;57(8):1786-1792.PubMedGoogle ScholarCrossref
32.
Rawstron  AC, Gregory  WM, de Tute  RM,  et al.  Minimal residual disease in myeloma by flow cytometry: independent prediction of survival benefit per log reduction.  Blood. 2015;125(12):1932-1935.PubMedGoogle ScholarCrossref
33.
Munshi  NC, Anderson  KC, Bergsagel  PL,  et al. Consensus recommendations for risk stratification in multiple myeloma: report of the International Myeloma Workshop Consensus Panel 2. Blood. 20115;117(18):4696-700.
34.
de Tute  RM, Rawstron  AC, Gregory  WM,  et al.  Minimal residual disease following autologous stem cell transplant in myeloma: impact on outcome is independent of induction regimen.  Haematologica. 2016;101(2):e69-e71.PubMedGoogle ScholarCrossref
35.
Paiva  B, van Dongen  JJM, Orfao  A.  New criteria for response assessment: role of minimal residual disease in multiple myeloma.  Blood. 2015;125(20):3059-3068.PubMedGoogle ScholarCrossref
36.
Flanders  A, Stetler-Stevenson  M, Landgren  O.  Minimal residual disease testing in multiple myeloma by flow cytometry: major heterogeneity.  Blood. 2013;122(6):1088-1089.PubMedGoogle ScholarCrossref
37.
San-Miguel  JF, Paiva  B, Gutierrez  C. New tools for diagnosis and monitoring of multiple myeloma. Am Soc Clin Oncol Educ Book. 2013. Available at: http://meetinglibrary.asco.org/sites/meetinglibrary.asco.org/files/edbook_pdf/2013_EdBook.pdf. Accessed March 2016.
38.
Jasielec  J, Dytfeld  D, Griffith  KA,  et al.  Minimal residual disease status predicts progression-free survival in newly diagnosed multiple myeloma (NDMM) patients treated with carfilzomib, lenalidomide, and low-dose dexamethasone (KRd).  Blood. 2014;124(21). abstract 2127.Google Scholar
39.
Korde  N, Mailankody  S, Roschewski  M,  et al.  Minimal residual disease (MRD) testing in newly diagnosed multiple myeloma (MM) patients: a prospective head-to-head assessment of cell-based, molecular, and molecular-imaging modalities.  Blood. 2014;124(21). abstract 2105.Google Scholar
40.
Takamatsu  Murata R, , Zheng J, et al.  Prognostic value of sequencing-based minimal residual disease detection in multiple myeloma.  Blood. 2014;124(21). abstract 2003.Google Scholar
41.
Avet-Loiseau  H, Corre  J, Maheo  S,  et al.  Identification rate of myeloma-specific clonotypes in multiple diagnostic sample types from patients with multiple myeloma using next-generation sequencing method.  Blood. 2014;124(21). abstract 2036.Google Scholar
42.
Fernández de Larrea  C, Delforge  M, Davies  F, Bladé  J.  Response evaluation and monitoring of multiple myeloma.  Expert Rev Hematol. 2014;7(1):33-42.PubMedGoogle ScholarCrossref
Original Investigation
January 2017

Association of Minimal Residual Disease With Superior Survival Outcomes in Patients With Multiple Myeloma: A Meta-analysis

Author Affiliations
  • 1Dana-Farber Cancer Institute, Harvard Medical School, Boston, Massachusetts
  • 2Unit for Genomics in Myeloma, Institut Universitaire du Cancer, Toulouse, France
  • 3St James's University Hospital, Leeds, United Kingdom
  • 4University of Leeds, Leeds, United Kingdom
  • 5Celgene Corporation, Summit, New Jersey
  • 6Celgene International, Boudry, Switzerland
  • 7Excerpta Medica, Amsterdam, the Netherlands
JAMA Oncol. 2017;3(1):28-35. doi:10.1001/jamaoncol.2016.3160
Key Points

Question  What is the utility of minimal residual disease (MRD) assessment in predicting survival outcomes in patients with newly diagnosed multiple myeloma (MM)?

Findings  In a meta-analysis of published literature (1990-2016) achievement of MRD-negative status following treatment was associated with a significant improvement in progression-free survival (hazard ratio [HR], 0.41; 95% CI, 0.36-0.48; N = 1273). Overall survival was also significantly prolonged in MRD-negative patients (HR, 0.57; 95% CI, 0.46-0.71; N = 1100). Focusing on patients achieving conventional complete response, MRD-negative status showed similar improvement in survival.

Meaning  This large-cohort meta-analysis identifies MRD status as a marker of long-term survival outcome in patients with MM and provides evidence to support the integration of MRD assessment as an end point in clinical trials.

Abstract

Importance  Numerous studies have evaluated the prognostic value of minimal residual disease (MRD) in patients with multiple myeloma (MM). Most studies were small and varied in terms of patient population, treatment, and MRD assessment methods.

Objective  To evaluate the utility of MRD detection in patients with newly diagnosed MM.

Data Sources  A Medline search was conducted for articles published in English between January 1990 and January 2016.

Study Selection  Eligible studies reported MRD status and progression-free survival (PFS) or overall survival (OS) in 20 or more patients following treatment. Among 405 articles identified, 21 met the initial eligibility criteria and were included in the analysis.

Data Extraction and Synthesis  Information on patient characteristics, treatment, MRD assessment, and outcomes were extracted using a standard form.

Main Outcomes and Measures  The impact of MRD status on PFS and OS was assessed by pooling data from relevant trials. Data were adjusted to allow for different proportions of patients with MRD in different studies, and analyzed using the Peto method. Forest plots were created based on Cox model analysis. Other prespecified research questions were addressed qualitatively.

Results  Fourteen studies (n = 1273) provided data on the impact of MRD on PFS, and 12 studies (n = 1100) on OS. Results were reported specifically in patients who had achieved conventional complete response (CR) in 5 studies for PFS (n = 574) and 6 studies for OS (n = 616). An MRD-negative status was associated with significantly better PFS overall (hazard ratio [HR], 0.41; 95% CI, 0.36-0.48; P < .001) and in studies specifically looking at CR patients (HR, 0.44; 95% CI, 0.34-0.56; P < .001). Overall survival was also favorable in MRD-negative patients overall (HR, 0.57; 95% CI, 0.46-0.71; P < .001) and in CR patients (HR, 0.47; 95% CI, 0.33-0.67; P < .001). Tests of heterogeneity found no significant differences among the studies for PFS and OS.

Conclusions and Relevance  Minimal residual disease-negative status after treatment for newly diagnosed MM is associated with long-term survival. These findings provide quantitative evidence to support the integration of MRD assessment as an end point in clinical trials of MM.

Introduction

A substantial proportion of patients with multiple myeloma (MM) can now expect to achieve clinical complete response (CR) as a result of recent therapeutic advances.1,2 These advances include the combined use of immunomodulatory drugs (thalidomide, lenalidomide, or pomalidomide) and proteasome inhibitors (bortezomib or carfilzomib), along with high-dose therapy with autologous stem cell transplantation (ASCT) in eligible individuals. Complete response rates are likely to continue to increase with the incorporation of novel combinations of therapies.3 Nevertheless, most patients who achieve CR eventually relapse,1 suggesting that a small but clinically relevant population of myeloma cells not detected by current techniques persists. Assays with greater sensitivity have been developed to detect minimal residual disease (MRD), including multiparameter flow cytometry (MFC), allele-specific oligonucleotide quantitative polymerase chain reaction (ASO-qPCR), and next-generation sequencing (NGS) techniques.3,4

Potential applications of MRD assessment in MM treatment are numerous.1,3,5,6 It is already considered an important prognostic factor.7 Minimal residual disease testing could be used to monitor response to therapy; the presence or absence of MRD may also inform subsequent treatment decisions, including consolidation and maintenance.7 Historically, owing to the complexity of conventional MRD assays, evaluations were limited to a small number of patients. Recent development of MFC- and NGS-based methods has allowed for MRD assessment in larger studies. To understand the real impact of MRD on outcomes from small-to-medium-sized studies, we performed a meta-analysis of all published data regarding the utility of MRD detection in patients with newly diagnosed MM (NDMM).

Methods
Literature Search and Article Selection

A Medline search was performed for articles published in English between January 1990 and January 2016, using the MeSH terms “multiple myeloma” AND “neoplasm, residual” and the nonMeSH terms “MRD,” “myeloma,” and “minimal residual disease.” Eligible articles included those that reported on controlled trials, randomized controlled trials, or patient cohort studies with MRD status and survival outcomes progression-free survival (PFS) or overall survival (OS) in 20 or more NDMM patients following therapy. Patients could have received any type of treatment except allogeneic stem cell transplantation (alloSCT), and MRD could be assessed by any method (MFC, ASO-qPCR, or NGS), but analysis was restricted to techniques with a limit of detection of 0.01% or lower. Trials were excluded if they: included only patients with relapsed/refractory MM (RRMM) or smoldering myeloma; assessed MRD in apheresis product; or reported on the same study population used in an already-included trial.

Data Extraction

If primary data were not accessible, survival graphs from relevant trials were carefully measured and a computer program was written to reconstruct the individual survival and censoring times from these measurements. Articles were scrutinized to ensure that all P values, CIs, hazard ratios (HRs), numbers of events/deaths, and median survival times, and durations of patient follow-up matched those reported. There was a PFS curve but not an OS curve for 1 study.8 However, P values and percentages at particular times were provided for the OS data, which enabled censored values to be used from the PFS curves; it was therefore possible to use the additional information from the article to derive the survival times. For the pooled analysis, data were adjusted to allow for the different proportions of patients with MRD in the different studies. P values are for adjusted log-rank χ2 tests.

Statistical Analysis

For a pooled analysis of all studies reporting survival data, PFS and OS curves were generated.9 This method adjusts for the different proportions of MRD positivity and negativity in each study, thereby avoiding inappropriate bias potentially generated by studies with high or low proportions of MRD positivity. The method produces an adjusted log-rank χ2 statistic to evaluate the significance of any differences between MRD positivity and negativity. It also provides a nonproportional hazards-based equivalent to performing a Cox model analysis stratified by study or group. If the hazards are proportional, the results will be similar to such a Cox model analysis, which was the case in all such analyses in this report.

The overview methodology described in detail by Peto10 was applied. In brief, for PFS and OS, the expected number (E) of events was derived in the MRD-positive and MRD-negative groups for each study, assuming no difference between the MRD groups. This was compared with the observed number (O) of events and the differences (O-Es) were then tested for heterogeneity to see whether the scatter of results was unexpected. The sum of [O-E]2/variance should be distributed as χ2n-1 if the scatter is random, where n is the number of studies.

Hazard ratio forest plots were then generated using the inverse variance weighting method, as described in detail by Whitehead and Whitehead.11 Cox proportional hazards model analysis was performed for each study, generating HR and CIs, and the required variance. An overall Cox model analysis was run on the whole data set, stratified by study to generate similar statistics for the total of all the studies combined. The 95% CIs are shown for the individual trials. For the overall result, 95% CIs are also given (open diamonds in the forest plot). The proportional hazards assumption was checked for the Cox model analyses using log-log plots and Schoenfeld residuals and any departures from proportionality were extremely minor.

There were no PFS events in the MRD-negative group in 1 study12 and no OS events in the MRD-negative group in another study,8 making it impossible to derive CIs and variance for Cox model HRs. In these 2 cases, an odds ratio approach was used to derive CIs and variance, incorporating the correlations between odds ratios and HRs which were all strong (r > .988).

Statistical analyses were performed with Stata (version 13.0, Stata Corp) or purpose-written Digital Visual Fortran (version 6.0A, Digital Equipment Corp) statistical software. Hypothesis tests were 2-sided.

Results
Literature Search

The initial search yielded 405 articles, and 25 additional articles were identified from the reference sections of recently published articles on the topic. After applying eligibility criteria 21 studies were included in the qualitative assessments (Figure 1).8,12-30 Of the 21 articles identified, 13 involved patients with NDMM and in 9 articles it was not reported whether the population was limited to NDMM patients. Sixteen articles involved ASCT-eligible patients and 1 involved ASCT-ineligible patients; the remaining 4 studies included both ASCT-eligible and ASCT-ineligible patients. The primary MRD assays that were evaluated were MFC (n = 9), PCR (n = 11), or NGS (n = 1).

Fourteen studies (n = 1273) reported information on the impact of MRD on PFS and 12 (n = 1100) assessed the impact of MRD on OS; these studies were therefore included in the overall quantitative meta-analysis (eTable in the Supplement). Twelve publications reported conventional CR7 at the time of MRD measurement.6,8,19,21-27,31 However, further investigation identified potential duplication of data across some studies and led to the exclusion of 5 additional articles from the quantitative analysis in CR patients.21-24,27

Impact of MRD Status on Survival Outcomes

The overall prognostic value of MRD status in terms of PFS was assessed in 14 studies involving 1273 patients (660 MRD-negative, 613 MRD-positive).8,12-14,16-18,24,25,28-31 The impact of MRD status on OS was assessed in 12 studies involving 1100 patients (599 MRD-negative, 501 MRD-positive).6,8,13,14,16-19,24,25,28,31 Compared with MRD positivity, MRD negativity was associated with better PFS (HR, 0.41; 95% CI, 0.36-0.48; P < .001) (Figure 2A) and OS (HR, 0.57; 95% CI, 0.46-0.71; P < .001) (Figure 2B). Median PFS was 54 months for MRD-negative patients and 26 months for MRD-positive patients (Figure 2C); median OS was 98 and 82 months, respectively (Figure 2D). Tests of heterogeneity found no significant differences among the studies for OS (χ2 = 8.81; df = 11; P = .64) but significant differences among the studies for PFS (χ2 = 42.1; df = 13; P < .001). This was a result of 2 very small studies,12,16 which showed unusually large differences; the Roussel et al12 study also had no events occurring in MRD negative patients. When these 2 studies were excluded the test for heterogeneity was no longer significant (χ2 = 10.1; df = 11; P = .53).

MRD Superior to Conventional CR for Survival Prediction

To evaluate the impact of MRD status on PFS in patients who had achieved conventional CR, data were pooled from 5 studies involving 574 patients (396 MRD-negative, 178 MRD-positive).8,25,26,28,31 For OS, data were pooled from 6 studies involving 616 patients (430 MRD-negative, 186 MRD-positive).8,19,25,26,28,31 In patients achieving CR, the presence of MRD predicted shorter PFS (HR, 0.44; 95% CI, 0.34-0.56; P < .001) (Figure 3A) and OS (HR, 0.47; 95% CI, 0.33-0.67; P < .001) (Figure 3B). Median PFS was 56 months for MRD-negative patients and 34 months for MRD-positive patients (Figure 3C) and median OS was 112 and 82 months, respectively (Figure 3D); PFS rates were 70% and 46% at 3 years, 48% and 27% at 5 years, and 37% and 14% at 7 years, respectively. Similarly, the OS rate was higher for MRD-negative patients compared with MRD-positive patients at 3 years (94% vs 80%), 5 years (80% vs 61%), and 7 years (67% vs 47%). Tests of heterogeneity found no significant differences among the studies for PFS (χ2 = 2.68; df = 4; P = .61) and OS (χ2 = 4.22; df = 5; P = .62).

Among the published analyses that were not restricted to CR patients, the impact of MRD on outcomes was less clear.12-14,16-18,29,30 One study found no significant difference in outcomes between patients with or without detectable MRD.14 Others noted that MRD status did not correlate with standard response criteria.17,18 In the study conducted by Rawstron et al6 it was noted that 34 (26%) of 246 MRD-negative patients did not achieve conventional CR, including 29 (12%) who had less than very good partial response (VGPR). Patients who were MRD-negative but failed to achieve CR had similar PFS and OS as those who were MRD-positive. Further analyses by this group suggested that log reduction in MRD (assessed as a continuous variable, rather than using a threshold for MRD positivity vs negativity), negated the significance of response in multivariate analyses for both PFS and OS.32

None of the trials directly compared the ability of 2 different treatment approaches to induce MRD-negative status. However, 5 studies evaluated MRD status before and after ASCT.6,12,17,18,20 All 5 indicated that ASCT increased the proportion of patients with MRD-negative status.

Prognostic Value of MRD Status in Relation to Other Prognostic Factors

Eleven articles reported results from univariate and/or multivariate analyses regarding the ability of MRD status to predict outcomes.13,16,18,21-26,28,30 In all 11 trials, MRD was shown to be a significant predictor of outcomes. Notably, only 6 articles mentioned cytogenetics: high-risk cytogenetics, defined as any t(4;14), t(14;16), or del(17p),33 was a negative predictor of PFS (or time to progression [TTP] or event free survival [EFS]) in 3 reports,18,24,26 and OS in 2 reports.18,26 In the study by Paiva et al26 the combination of MRD status and cytogenetics was highly predictive of TTP, and the combination of MRD status, cytogenetics, and age was predictive of OS. Only 1 study reported that MRD status predicted PFS and OS in patients with unfavorable cytogenetics (defined as gain[1q], del[1p32], t[4;14], t[14;20], t[14;16], and del[17p]).6 Our meta-analysis of these latter studies6,26 indicated that the best OS is seen in patients with favorable cytogenetics who achieve MRD negativity compared with patients who are either high-risk or MRD-positive; worst results are seen in patients with high-risk cytogenetics who remain MRD-positive (P < .001) (eFigure in the Supplement). In a more recent analysis, cytogenetics (favorable vs unfavorable vs unknown/not evaluable) and log reduction in MRD were the only significant predictors of both PFS and OS in multivariate analysis.32

Impact of Maintenance Therapy on MRD

Ten studies mentioned maintenance therapy,6,8,12,15,17,18,25-27,29 but only 2 specifically evaluated MRD status after maintenance therapy. In 1 article, lenalidomide maintenance therapy was reported to increase response status in 4 patients and MRD status in 5 patients.12 In the MM-IX study, more MRD-positive patients became MRD-negative during thalidomide maintenance compared with patients on no maintenance (8/29 [28%] vs 1/29 [3%]).6 Furthermore, more MRD-negative patients remained MRD-negative with thalidomide maintenance than with no maintenance (24/25 [96%] vs 11/16 [69%]; P = .026).

Discussion

This large-cohort meta-analysis confirms that MRD status has prognostic value and is a valid surrogate marker for both PFS and OS in patients with MM, including those who had achieved a CR. All studies, irrespective of the therapies used, uniformly confirmed the impact of MRD status on outcome, indicating that the predictive value of MRD status was independent of the type of treatment used. This is consistent with the results of a recent study34 demonstrating that the depth of MRD is the determining factor for subsequent outcome. Findings from this meta-analysis provide quantitative evidence to support the conceptual basis for integrating MRD assessment into the treatment of MM.35

One of the main strengths of this analysis of pooled data from different clinical trials is the method used to generate the PFS and OS curves. These curves were adjusted for each study or group to allow for different proportions of patients with MRD positivity and negativity in the different studies, using methods described in detail elsewhere.9 This approach avoids the creation of curves that were biased inappropriately by studies with very high or very low proportions of patients with MRD positivity.

This analysis did not account for the type of MRD test used in each study. Approaches to testing vary widely36; the sensitivity of different protocols also varies.4,27,36,37 However, this may represent a strength of the analysis as the results are method-agnostic, ie, it suggests that if MRD is undetectable with a certain level of sensitivity, the results have similar significance irrespective of the method used. The MFC assay is the most widely used method for MRD testing in MM thus far owing to its broad availability, short turnaround time, and relatively low cost.3 The main limitations of this technique are its lower sensitivity (up to 1 × 10−4 or 1 × 10−5) and lack of standardization among laboratories. The ASO-qPCR assay, although sensitive, is cumbersome and is being replaced by NGS-based MRD assessment which is more sensitive than MFC38,39 or ASO-qPCR,40 and feasible in up to 90% of MM patients.41 To assess whether differences in the method of MRD assessment across the studies would impact our findings, we performed additional analyses comparing HRs for OS and PFS according to the 2 major methods of MRD assessment, flow cytometry and PCR. The HR for OS in the MFC studies (n = 923) was 0.60 (95% CI, 0.47-0.76); in the PCR studies (n = 177) it was 0.44 (95% CI, 0.26-0.77). The HR for PFS (n = 1072) in the MFC studies was 0.44 (95% CI, 0.37-0.52); and in the PCR studies (n = 201) it was 0.27 (95% CI, 0.18-0.40). As expected, the HR is slightly greater in the PCR studies as it provides a more sensitive measurement.

The studies in this analysis included primarily NDMM patients, most of whom were undergoing ASCT. The applicability of the results of this analysis in other populations, such as those with transplant-ineligible NDMM, RRMM, or high-risk cytogenetic features, is unclear. In addition, the timing of MRD assessment varied among the studies. For example, among the 14 trials included in the overall PFS meta-analysis, 5 assessed MRD before ASCT and 12 assessed MRD after ASCT. Among the trials assessing MRD after ASCT, most assessed patients after 3 months (or day 100), but some continued to assess patients every 3 to 6 months thereafter. Despite these differences, all studies showed large and consistent effects of MRD, confirmed by the nonsignificant χ2 statistic for heterogeneity, suggesting that any methodological variations between studies have a relatively minor influence on the overall MRD effect. In addition, there is always a risk in meta-analyses that negative results are less likely to have been reported, eg, lack of effect of MRD status on OS and/or PFS. Lastly, this analysis did not isolate the prognostic effect of MRD from those of post-transplant treatments patients may have received. Future trials will need to focus on some of these questions to determine the clinical utility of MRD assessment as well as its ability to inform treatment decisions.

Assessment of MRD has several important potential applications in treating MM.1,42 In clinical trials, MRD assessment after initial treatment could be a useful surrogate end point for PFS and/or OS. It is in fact becoming an important component of the recommendations for uniform reporting of clinical trials.7 In clinical practice MRD testing may aid in prognostication; help make decisions regarding subsequent treatment, especially consolidation treatment; and, in the near future, guide the type and duration of maintenance therapy. Importantly, as the frequency of CR has increased, MRD negativity is emerging as a key end point for clinical studies. Integration of MRD testing into standard practice requires optimization and standardization of MRD assessment and standardization of its timing.4,7,42 Test standardization includes establishing optimal assay methods, timing of sample collection, sensitivity requirements, thresholds for MRD-positive status, and other factors.36 For example, recent evidence suggests that MRD quantitation may be more informative than MRD status: The MRC Myeloma IX trial32 demonstrated a 1-year survival benefit for each 1-log depletion in tumor burden by MFC. The questions to be addressed in the future include determining the impact of different treatment approaches on MRD status (eg, consolidation or maintenance therapy); and the prognostic importance of MRD status in relation to other known prognostic factors.

Conclusions

The results of this large analysis showed that MRD negativity, as determined by various high-sensitivity methods, predicted better PFS and OS in patients with MM, including those who had achieved CR. Minimal residual disease status is a marker of long-term outcomes in patients with MM. It should therefore be considered a new end point in clinical trials and clearly has a role as a surrogate marker of OS.

Back to top
Article Information

Corresponding Author: Nikhil C. Munshi, MD, Dana-Farber Cancer Institute, 450 Brookline Ave, Boston, MA 02215 (nikhil_munshi@dfci.harvard.edu).

Accepted for Publication: May 24, 2016.

Published Online: September 15, 2016. doi:10.1001/jamaoncol.2016.3160

Author Contributions: Dr Gregory had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. The authors are fully responsible for all content and editorial decisions.

Concept and design: Munshi, Avet-Loiseau, Sherrington, Georgieva, Anderson, Gregory.

Acquisition, analysis, or interpretation of data: All Authors.

Drafting of the manuscript: Munshi, Avet-Loiseau, Samur, Georgieva, Gregory.

Critical revision of the manuscript for important intellectual content: Munshi, Avet-Loiseau, Rawstron, Owen, Child, Thakurta, Sherrington, Georgieva, Anderson, Gregory.

Statistical analysis: Samur, Gregory.

Obtained funding: Sherrington.

Administrative, technical, or material support: Munshi, Rawstron, Sherrington, Anderson.

Study supervision: Munshi, Thakurta.

Conflict of Interest Disclosures: Dr Munshi reports consultancy positions and advisory committee memberships with Celgene, Takeda, Janssen, and Merck. Dr Anderson reports consultancy positions with Celgene, Takeda, and Gilead. Dr Avet-Loiseau reports consultancy positions and advisory committee memberships with Celgene and Takeda. Dr Rawstron: Janssen reports research funding from Gilead and research funding, consultancy, honoraria; Celgene, Roche, Abbvie, Genzyme, GSK – honoraria; BD biosciences – other remuneration. Dr Owen: Celgene – consultancy, honoraria, research funding; Janssen consultancy and honoraria from Takeda. Drs Thakurta and Sherrington report employment and equity ownership with Celgene. Dr Child reports research funding from Celgene. Dr Georgieva reports employment with Excerpta Medica. Dr Gregory reports consultancy with Celgene and honoraria from Janssen. No other disclosures are reported.

Funding/Support: The work was partly supported by NIH grants PO1-155258, and P50-100707 (NCM, HAL and KCA). and an MRC UK grant G0100132 (JAC) and partly funded by Celgene. Celgene provided financial support for statistical analysis and editorial assistance for this study.

Role of the Funder/Sponsor: The National Institutes of Health, the Medical Research Council, and Celgene had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Additional Contributions: The authors received editorial assistance in the preparation of this manuscript from Daniel Gilmartin, PhD, Excerpta Medica. He was not compensated for his contribution. The authors thank all the investigators and patients who participated in the clinical trials referred to in the paper.

References
1.
Munshi  NC, Anderson  KC.  Minimal residual disease in multiple myeloma.  J Clin Oncol. 2013;31(20):2523-2526.PubMedGoogle ScholarCrossref
2.
Anderson  KC.  Oncogenomics to target myeloma in the bone marrow microenvironment.  Clin Cancer Res. 2011;17(6):1225-1233.PubMedGoogle ScholarCrossref
3.
Mailankody  S, Korde  N, Lesokhin  AM,  et al.  Minimal residual disease in multiple myeloma: bringing the bench to the bedside.  Nat Rev Clin Oncol. 2015;12(5):286-295.PubMedGoogle ScholarCrossref
4.
Biran  N, Ely  S, Chari  A.  Controversies in the assessment of minimal residual disease in multiple myeloma: clinical significance of minimal residual disease negativity using highly sensitive techniques.  Curr Hematol Malig Rep. 2014;9(4):368-378.PubMedGoogle ScholarCrossref
5.
Rawstron  AC, de Tute  RM, Haughton  J, Owen  RG.  Measuring disease levels in myeloma using flow cytometry in combination with other laboratory techniques: Lessons from the past 20 years at the Leeds Haematological Malignancy Diagnostic Service.  Cytometry B Clin Cytom. 2016;90(1):54-60.PubMedGoogle ScholarCrossref
6.
Rawstron  AC, Child  JA, de Tute  RM,  et al.  Minimal residual disease assessed by multiparameter flow cytometry in multiple myeloma: impact on outcome in the Medical Research Council Myeloma IX Study.  J Clin Oncol. 2013;31(20):2540-2547.PubMedGoogle ScholarCrossref
7.
Rajkumar  SV, Harousseau  JL, Durie  B,  et al; International Myeloma Workshop Consensus Panel 1.  Consensus recommendations for the uniform reporting of clinical trials: report of the International Myeloma Workshop Consensus Panel 1.  Blood. 2011;117(18):4691-4695.PubMedGoogle ScholarCrossref
8.
Swedin  A, Lenhoff  S, Olofsson  T, Thuresson  B, Westin  J.  Clinical utility of immunoglobulin heavy chain gene rearrangement identification for tumour cell detection in multiple myeloma.  Br J Haematol. 1998;103(4):1145-1151.PubMedGoogle ScholarCrossref
9.
Gregory  WM.  Adjusting survival curves for imbalances in prognostic factors.  Br J Cancer. 1988;58(2):202-204.PubMedGoogle ScholarCrossref
10.
Peto  R.  Why do we need systematic overviews of randomized trials?  Stat Med. 1987;6(3):233-244.PubMedGoogle ScholarCrossref
11.
Whitehead  A, Whitehead  J.  A general parametric approach to the meta-analysis of randomized clinical trials.  Stat Med. 1991;10(11):1665-1677.PubMedGoogle ScholarCrossref
12.
Roussel  M, Lauwers-Cances  V, Robillard  N,  et al.  Front-line transplantation program with lenalidomide, bortezomib, and dexamethasone combination as induction and consolidation followed by lenalidomide maintenance in patients with multiple myeloma: a phase II study by the Intergroupe Francophone du Myélome.  J Clin Oncol. 2014;32(25):2712-2717.PubMedGoogle ScholarCrossref
13.
Bakkus  MH, Bouko  Y, Samson  D,  et al.  Post-transplantation tumour load in bone marrow, as assessed by quantitative ASO-PCR, is a prognostic parameter in multiple myeloma.  Br J Haematol. 2004;126(5):665-674.PubMedGoogle ScholarCrossref
14.
Dal Bó  S, Pezzi  A, Amorin  B, Valim  V, Isabel Bittencourt  R, Silla  L.  Detection of minimal residual disease by flow cytometry for patients with multiple myeloma submitted to autologous hematopoietic stem cell transplantation.  ISRN Hematol. 2013;2013:847672. doi:10.1155/2013/847672PubMedGoogle ScholarCrossref
15.
Davies  FE, Forsyth  PD, Rawstron  AC,  et al.  The impact of attaining a minimal disease state after high-dose melphalan and autologous transplantation for multiple myeloma.  Br J Haematol. 2001;112(3):814-819.PubMedGoogle ScholarCrossref
16.
Ferrero  S, Ladetto  M, Drandi  D,  et al.  Long-term results of the GIMEMA VEL-03-096 trial in MM patients receiving VTD consolidation after ASCT: MRD kinetics’ impact on survival.  Leukemia. 2015;29(3):689-695.PubMedGoogle ScholarCrossref
17.
Korthals  M, Sehnke  N, Kronenwett  R,  et al.  Molecular monitoring of minimal residual disease in the peripheral blood of patients with multiple myeloma.  Biol Blood Marrow Transplant. 2013;19(7):1109-1115.PubMedGoogle ScholarCrossref
18.
Korthals  M, Sehnke  N, Kronenwett  R,  et al.  The level of minimal residual disease in the bone marrow of patients with multiple myeloma before high-dose therapy and autologous blood stem cell transplantation is an independent predictive parameter.  Biol Blood Marrow Transplant. 2012;18(3):423-431.e3.PubMedGoogle ScholarCrossref
19.
Ludwig  H, Greil  R, Masszi  T,  et al.  Bortezomib, thalidomide and dexamethasone, with or without cyclophosphamide, for patients with previously untreated multiple myeloma: 5-year follow-up.  Br J Haematol. 2015;171(3):344-354.PubMedGoogle ScholarCrossref
20.
Ludwig  H, Viterbo  L, Greil  R,  et al.  Randomized phase II study of bortezomib, thalidomide, and dexamethasone with or without cyclophosphamide as induction therapy in previously untreated multiple myeloma.  J Clin Oncol. 2013;31(2):247-255.PubMedGoogle ScholarCrossref
21.
Martinez-Lopez  J, Fernández-Redondo  E, García-Sánz  R,  et al; GEM (Grupo Español Multidisciplinar de Melanoma)/PETHEMA (Programa para el Estudio de la Terapéutica en Hemopatías Malignas) cooperative study group.  Clinical applicability and prognostic significance of molecular response assessed by fluorescent-PCR of immunoglobulin genes in multiple myeloma. Results from a GEM/PETHEMA study.  Br J Haematol. 2013;163(5):581-589.PubMedGoogle ScholarCrossref
22.
Martinez-Lopez  J, Lahuerta  JJ, Pepin  F,  et al.  Prognostic value of deep sequencing method for minimal residual disease detection in multiple myeloma.  Blood. 2014;123(20):3073-3079.PubMedGoogle ScholarCrossref
23.
Martínez-Sánchez  P, Montejano  L, Sarasquete  ME,  et al.  Evaluation of minimal residual disease in multiple myeloma patients by fluorescent-polymerase chain reaction: the prognostic impact of achieving molecular response.  Br J Haematol. 2008;142(5):766-774.PubMedGoogle ScholarCrossref
24.
Paiva  B, Vidriales  MB, Cerveró  J,  et al; GEM (Grupo Español de MM)/PETHEMA (Programa para el Estudio de la Terapéutica en Hemopatías Malignas) Cooperative Study Groups.  Multiparameter flow cytometric remission is the most relevant prognostic factor for multiple myeloma patients who undergo autologous stem cell transplantation.  Blood. 2008;112(10):4017-4023.PubMedGoogle ScholarCrossref
25.
Paiva  B, Martinez-Lopez  J, Vidriales  MB,  et al.  Comparison of immunofixation, serum free light chain, and immunophenotyping for response evaluation and prognostication in multiple myeloma.  J Clin Oncol. 2011;29(12):1627-1633.PubMedGoogle ScholarCrossref
26.
Paiva  B, Gutiérrez  NC, Rosiñol  L,  et al; PETHEMA/GEM (Programa para el Estudio de la Terapéutica en Hemopatías Malignas/Grupo Español de Mieloma) Cooperative Study Groups.  High-risk cytogenetics and persistent minimal residual disease by multiparameter flow cytometry predict unsustained complete response after autologous stem cell transplantation in multiple myeloma.  Blood. 2012;119(3):687-691.PubMedGoogle ScholarCrossref
27.
Puig  N, Sarasquete  ME, Balanzategui  A,  et al.  Critical evaluation of ASO RQ-PCR for minimal residual disease evaluation in multiple myeloma. A comparative analysis with flow cytometry.  Leukemia. 2014;28(2):391-397.PubMedGoogle ScholarCrossref
28.
Rawstron  AC, Davies  FE, DasGupta  R,  et al.  Flow cytometric disease monitoring in multiple myeloma: the relationship between normal and neoplastic plasma cells predicts outcome after transplantation.  Blood. 2002;100(9):3095-3100.PubMedGoogle ScholarCrossref
29.
San Miguel  JF, Almeida  J, Mateo  G,  et al.  Immunophenotypic evaluation of the plasma cell compartment in multiple myeloma: a tool for comparing the efficacy of different treatment strategies and predicting outcome.  Blood. 2002;99(5):1853-1856.PubMedGoogle ScholarCrossref
30.
Sarasquete  ME, García-Sanz  R, González  D,  et al.  Minimal residual disease monitoring in multiple myeloma: a comparison between allelic-specific oligonucleotide real-time quantitative polymerase chain reaction and flow cytometry.  Haematologica. 2005;90(10):1365-1372.PubMedGoogle Scholar
31.
Fukumoto  K, Fujisawa  M, Suehara  Y,  et al.  Prognostic impact of immunophenotypic complete response in patients with multiple myeloma achieving better than complete response.  Leuk Lymphoma. 2016;57(8):1786-1792.PubMedGoogle ScholarCrossref
32.
Rawstron  AC, Gregory  WM, de Tute  RM,  et al.  Minimal residual disease in myeloma by flow cytometry: independent prediction of survival benefit per log reduction.  Blood. 2015;125(12):1932-1935.PubMedGoogle ScholarCrossref
33.
Munshi  NC, Anderson  KC, Bergsagel  PL,  et al. Consensus recommendations for risk stratification in multiple myeloma: report of the International Myeloma Workshop Consensus Panel 2. Blood. 20115;117(18):4696-700.
34.
de Tute  RM, Rawstron  AC, Gregory  WM,  et al.  Minimal residual disease following autologous stem cell transplant in myeloma: impact on outcome is independent of induction regimen.  Haematologica. 2016;101(2):e69-e71.PubMedGoogle ScholarCrossref
35.
Paiva  B, van Dongen  JJM, Orfao  A.  New criteria for response assessment: role of minimal residual disease in multiple myeloma.  Blood. 2015;125(20):3059-3068.PubMedGoogle ScholarCrossref
36.
Flanders  A, Stetler-Stevenson  M, Landgren  O.  Minimal residual disease testing in multiple myeloma by flow cytometry: major heterogeneity.  Blood. 2013;122(6):1088-1089.PubMedGoogle ScholarCrossref
37.
San-Miguel  JF, Paiva  B, Gutierrez  C. New tools for diagnosis and monitoring of multiple myeloma. Am Soc Clin Oncol Educ Book. 2013. Available at: http://meetinglibrary.asco.org/sites/meetinglibrary.asco.org/files/edbook_pdf/2013_EdBook.pdf. Accessed March 2016.
38.
Jasielec  J, Dytfeld  D, Griffith  KA,  et al.  Minimal residual disease status predicts progression-free survival in newly diagnosed multiple myeloma (NDMM) patients treated with carfilzomib, lenalidomide, and low-dose dexamethasone (KRd).  Blood. 2014;124(21). abstract 2127.Google Scholar
39.
Korde  N, Mailankody  S, Roschewski  M,  et al.  Minimal residual disease (MRD) testing in newly diagnosed multiple myeloma (MM) patients: a prospective head-to-head assessment of cell-based, molecular, and molecular-imaging modalities.  Blood. 2014;124(21). abstract 2105.Google Scholar
40.
Takamatsu  Murata R, , Zheng J, et al.  Prognostic value of sequencing-based minimal residual disease detection in multiple myeloma.  Blood. 2014;124(21). abstract 2003.Google Scholar
41.
Avet-Loiseau  H, Corre  J, Maheo  S,  et al.  Identification rate of myeloma-specific clonotypes in multiple diagnostic sample types from patients with multiple myeloma using next-generation sequencing method.  Blood. 2014;124(21). abstract 2036.Google Scholar
42.
Fernández de Larrea  C, Delforge  M, Davies  F, Bladé  J.  Response evaluation and monitoring of multiple myeloma.  Expert Rev Hematol. 2014;7(1):33-42.PubMedGoogle ScholarCrossref
×