In the last two decades, the treatment approach to multiple myeloma (MM) has been redefined by the development of a number of active compounds of several drug-classes. While only few patients were able to achieve a complete remission (CR) with conventional chemotherapy, approximately half of treated patients achieved a CR with multi-drug regimens based on non-cross resistant anti-MM agents such as immunomodulatory drugs (IMiDs), proteasome inhibitors (PIs) and monoclonal antibodies (mAbs), with or without high-dose melphalan (HDM) and autologous stem-cell transplantation (ASCT) [1
As compared to non-CR patients, the achievement of a CR has been associated to significantly longer progression-free survival (PFS) and overall survival (OS) [7
]. Even in CR patients, however, treatment intensification with HDM-ASCT and maintenance therapy were associated to better survival outcomes [9
It is now clear that the real prognostic value of CR relies on the absence of minimal residual disease (MRD). Since in 2016 the International Myeloma Working Group (IMWG) updated the response criteria by including new response categories based on the assessment of MRD both inside and outside the bone marrow [12
], interest has emerged in MRD as a surrogate endpoint for survival outcomes and in its incorporation in clinical trials.
In this review, we discuss how MRD assessment is currently being performed, what previous studies have taught us about the role of MRD in the treatment of MM, and how MRD could change the management of MM patients in the next future.
3. Current Evidence on the Role of MRD Assessment in MM
Several studies support the use of MRD for response monitoring in MM (Table 2
). The positive correlation between MRD negativity and prolonged PFS and OS was shown by two meta-analyses including several clinical trials in which MRD was assessed with a sensitivity level of 10−4
by both immunophenotypic and molecular techniques [21
The evidence that MRD status stratified patients in CR into 2 groups (CR-MRD negative and CR-MRD positive, the former having significantly longer PFS and OS) led to the idea that “conventional” CR was no more a clinically meaningful endpoint. Lahuerta and colleagues demonstrated that MRD-negative patients by MFC at a sensitivity of 10−4
in a conventionally defined CR had better PFS and OS as compared to MRD-positive CR patients. They also showed that the patients in the latter group had similar outcomes to patients in PR [61
]. In light of this observation, MRD negativity superseded the former CR definition as an endpoint, and the latest clinical trials have already incorporated the MRD negativity rate as a clinical endpoint.
As reported in other hematologic diseases, also in MM the higher the sensitivity for the definition of MRD undetectability, the better is the outcome. Both NGS and NGF confirmed this result: patients who achieved MRD negativity at a sensitivity of 10−6
showed a prolonged PFS, as compared to those who were MRD negative at 10−5
or higher [17
Whether MRD negativity could abrogate the high risk associated with unfavorable cytogenetic abnormalities detected by fluorescence in situ hybridization (FISH) or biological characteristics such as the International Staging System (ISS) is still a matter of debate. The achievement of MRD negativity, as well as of sustained MRD negativity, is less frequent in high-risk patients than in standard-risk patients. In the IFM 2009 and EMN02 trials, del(17p)-positive patients achieved MRD negativity in 11% and 7% of cases, respectively [17
]. However, it was reported that high-risk patients (defined by ISS, Revised ISS [R-ISS] and FISH) who reached MRD negativity had a survival rate comparable to that of standard-risk patients. Paiva et al. showed that the TTP in MRD-negative patients was similar irrespective of FISH status (not reached; P = 0.70), whereas FISH status still had a significant impact on TTP in MRD-positive patients (standard risk, median TTP 15 months vs. 12 months for MRD-positive high-risk patients; P = 0.02) [51
]. The EMN02 study confirmed this evidence: patients with high-risk MM at diagnosis (defined by ISS and FISH) and persistent residual disease after treatment had a dismal outcome (median PFS of 7 months for patients with ISS stage III and 15 months for those with high risk by FISH) [52
]. Similarly, Paiva et al. showed that the initial R-ISS prognostic stratification was meaningful only in patients with persistent MRD, but not in MRD-negative patients [23
Many studies demonstrated that the favorable prognosis associated with MRD negativity is treatment-naïve. Perrot et al. observed no differences in terms of PFS for MRD-negative patients who received HDM/ASCT vs. those who did not receive ASCT [17
]; however, a higher number of patients in the ASCT group obtained MRD negativity, as compared to the non-ASCT group (79% vs. 65%, P < 0.001) [62
]. In transplant-ineligible patients, the addition of daratumumab to the standard combinations bortezomib-melphalan-prednisone and lenalidomide-dexamethasone increased the rates of MRD negativity in comparison with standard therapy (27–24% vs. 7%, respectively; P < 0.001 for both) and prolonged PFS and OS in the overall population. Again, the PFS was similar in MRD-negative patients, irrespectively of previous treatment [1
In the phase III CASSIOPEIA study, the addition of daratumumab to the standard induction regimen bortezomib-thalidomide-dexamethasone (D-VTd) induced significantly higher rates of MRD at 10−5
than VTd alone, with a PFS benefit in patients achieving MRD negativity [6
]. Interestingly, the addition of daratumumab not only prolonged the PFS of MRD-positive patients but also seemed to benefit MRD-negative patients [6
]. A longer follow-up is needed to confirm these data.
The achievement of MRD negativity is a treatment goal that could be pursued not only in young, transplant-eligible patients, but also in older ones. This was confirmed by the Spanish PETHEMA/GEM2010MAS65 study, in which age did not negatively affect the outcome of MRD-negative patients (median TTP was not reached for patients aged 65–75 years vs. >75 years; P = 0.74) [51
4. Incorporation of MRD Results into Clinical Practice
Despite its incorporation into the IMWG response criteria, there is no current evidence that MRD can be used to drive therapeutic choices and tailor patient treatment in standard clinical practice. Nonetheless, MRD has been recently adopted as a new primary endpoint in ongoing trials, which will inform whether MRD status might be used to modulate treatment strategies.
High-dose melphalan plus ASCT is a standard intensification strategy for the treatment of newly diagnosed (ND)MM patients younger than 70–75 years of age [63
]. The role of ASCT has been frequently challenged in the past, but all studies reported a PFS advantage for patients receiving ASCT, as compared to no ASCT, although with an inconsistent OS benefit [62
]. Whether MRD status could be used to decide between a transplant-based or a non-transplant-based approach in an era of highly effective induction regimens is still a matter of debate. In the IFM 2009 study, no difference in terms of PFS was found in MRD-negative patients irrespective of treatment received (ASCT or VRD), although a higher number of patients in the ASCT arm were MRD negative as compared to the VRD arm (30% vs. 21%) [17
]. In the FORTE study, the rate of MRD negativity (10−5
, by MFC) was similar (58% vs. 54%) in patients who received 4 induction cycles with carfilzomib-lenalidomide-dexamethasone (KRd) followed by ASCT and further 4 KRd consolidation cycles vs. 12 KRd cycles without ASCT [56
In a recently presented study, 4 daratumumab-KRd (D-KRd) induction cycles succeeded in converting to MRD negativity (10−5
, by NGS) 40% of treated patients. After HDM-ASCT, the MRD negativity rate increased up to 73% [59
], a proportion similar to that observed in another study presented at the ASH 2019 meeting by Landgren et al. [60
], in which 77% of patients who received 8 induction cycles of D-KRd without transplant were MRD negative (10−5
, by NGS). With the development of new multi-drug regimens, the benefit in terms of response rates conferred by HDM plus ASCT over no ASCT seemed to decrease, although a long-term follow-up is needed to determine potential differences in terms of PFS and OS. Altogether, these data suggest the hypothesis that patients who are able to achieve MRD negativity with the induction therapy may not need HDM-ASCT, thus supporting the development of controlled trials randomizing patients to ASCT vs. non-transplant-based strategies. In this light, results from the FORTE study will help determine the future of HDM-ASCT in the management of young MM patients.
In the context of the therapeutic approach to MM, another open question concerns the need for maintenance therapy after induction/HDM-ASCT in MRD-negative patients and its optimal duration. Since MRD negativity at day + 100 after ASCT is associated with prolonged survival [67
], it is legitimate to speculate that MRD-negative patients may not need maintenance therapy after ASCT, thus benefiting from a treatment-free interval. In the Myeloma XI study, the best outcome in terms of PFS was reported in MRD-negative patients who still received lenalidomide maintenance. Previous studies also showed that lenalidomide maintenance can convert a significant number of MRD-positive patients to MRD negativity (44% in the EMN02 study, 32% In the Myeloma XI study) [52
]. On the basis of the data generated so far, it is not yet possible to advise against maintenance therapy for patients who are MRD negative after ASCT or to encourage treatment discontinuation for those patients who become MRD negative during maintenance. Results of ongoing clinical trials specifically designed to address these points are eagerly awaited. In the EMN17 Perseus trial (NCT03710603), patients with sustained MRD negativity (at least 2 negative samples 12 months apart) who receive maintenance treatment with daratumumab-lenalidomide have the opportunity to withhold daratumumab maintenance after at least 2 years and continue with lenalidomide only. In the SWOG S1803 (NCT04071457) study comparing daratumumab-lenalidomide to lenalidomide alone as maintenance after ASCT, patients who are MRD negative after 2 years of maintenance are randomized to treatment discontinuation vs. continuation, thus allowing the investigators to address this important issue.
So far, the exact frequency of MRD testing is unknown. MRD monitoring could provide clinically useful information not only because, as previously reported, MRD-positive patients can be converted to MRD-negativity during treatment but also because MRD-negative patients can become MRD positive over time, this being an early sign of relapsing MM. An analysis of MRD kinetics with serial MRD assessments (every 6 months) during lenalidomide maintenance showed that MRD reappearance from a previous MRD-negative sample predicted 4 months in advance a biochemical relapse (i.e., the reappearance of a monoclonal component in serum or urine) and 9 months in advance a clinical relapse [69
]. In the same study, however, 30% of patients with persistent MRD positivity did not experience any relapse. From a clinical perspective, these data could help clinicians restart treatment before the occurrence of clinical relapse, thus preventing the morbidity associated to MM proliferation [70
]. Altogether, this evidence prompts the development of further studies to explore the role of MRD assessment and kinetics and their correlation with survival outcomes.
Finally, given that MRD negativity is a strong prognostic factor for PFS and OS, another important open question that needs to be clarified by randomized trials is whether patients who are MRD positive at a specific time point—such as after HDM-ASCT or consolidation—would benefit from a treatment switch to non-cross-resistant drugs or from treatment intensification, in order to maximize the odds of achieving MRD negativity, particularly in light of the promising results obtained with immunotherapeutic strategies tested in the relapse setting.
Aside from the upfront setting, where the odds of achieving CR and MRD negativity are higher, MRD data also came from the relapse setting, in which the latest studies, in particular with antibodies and cellular products, reported unprecedented rates of CR and even of MRD negativity among heavily pretreated patients (Table 3
). The rate of MRD-negative patients (10−5
) treated at relapse with an anti-CD38 mAb (either daratumumab or isatuximab) in combination with IMiDs or PIs ranged between 5% and 30% [71
]. In a study on the CAR T-cell therapy bb2121, Raje et al. observed an impressive 94% of MRD negativity (NGS, 10−5
) among evaluable, heavily pre-treated RRMM patients. Despite this outstanding result, the median PFS of the study was only 12 months, with many MRD-negative patients who relapsed despite the achievement of a deep response [73
]. Although newer agents allow relapsed patients to obtain deep responses that are ultimately associated with prolonged survival, these data suggest that we still need to understand the value of MRD at relapse and how to comprehensively assess disease response in this setting.