Pursuing a Curative Approach in Multiple Myeloma: A Review of New Therapeutic Strategies

Multiple myeloma (MM) is still considered an incurable hematologic cancer and, in the last decades, the treatment goal has been to obtain a long-lasting disease control. However, the recent availability of new effective drugs has led to unprecedented high-quality responses and prolonged progression-free survival and overall survival. The improvement of response rates has prompted the development of new, very sensitive methods to measure residual disease, even when monoclonal components become undetectable in patients’ serum and urine. Several scientific efforts have been made to develop reliable and validated techniques to measure minimal residual disease (MRD), both within and outside the bone marrow. With the newest multidrug combinations, a good proportion of MM patients can achieve MRD negativity. Long-lasting MRD negativity may prove to be a marker of “operational cure”, although the follow-up of the currently ongoing studies is still too short to draw conclusions. In this article, we focus on results obtained with new-generation multidrug combinations in the treatment of high-risk smoldering MM and newly diagnosed MM, including the potential role of MRD and MRD-driven treatment strategies in clinical trials, in order to optimize and individualize treatment.


Introduction
Multiple Myeloma (MM) is a hematologic malignancy caused by the outgrowth of monoclonal plasma cells that leads to end-organ damage [1]. In 2018, MM accounted for 1.2% of all cancer diagnoses and 1.6% of all cancer deaths in Europe [2]. The median overall survival (OS) of newly diagnosed (ND) MM patients improved from 3.9 years for patients diagnosed between 2004 and 2007 to 6.3 years for those diagnosed between 2008 and 2012 to a median OS that is not yet reached in patients diagnosed after 2012 [3]. The introduction of new drug classes like proteasome inhibitors (PIs), immunomodulatory drugs (IMiDs), and, more recently, monoclonal antibodies (mAbs) has been the main determinant of the observed OS improvement, together with an improved supportive care. Nevertheless, the main cause of death in MM patients is still the development of drug-resistant disease [4]. Although obtaining deep responses is a universally recognized predictive factor of good outcome [5], long-term disease control, rather than disease eradication, is still the aim of MM treatment in current clinical practice, since the available data show that even patients achieving minimal residual disease (MRD) negativity relapse. This confirms the so-far incurable nature of MM. Recent data, comparing the survival of young MM patients treated between 2005 and 2015 to that of young patients affected by curable hematologic diseases (e.g., diffuse large B cell lymphoma (DLBCL) and Hodgkin  [45] 421 None of the risk factors: -M-protein >2g/dL -BMPCs >20% -FLCr >20
The current standard of care for SMM is periodical monitoring, with a suggested frequency based on patient risk to identify the possible evolution to symptomatic MM in due time and avoid severe organ damage. While this strategy suits well low-risk SMM patients who are unlikely to progress to MM, it may be questionable in high-risk SMM. In this setting, open questions are: (1) Can these patients benefit from an early therapy aiming at delaying the very likely evolution to MM? (2) Is there a possibility that early treatment may actually cure the disease? The latter hypothesis is based upon the very good outcome observed in patients with symptomatic MM and good prognosis [50], as well as upon a lower genomic complexity during the early phases of the disease that, together with a lower tumor burden, might suggest a higher possibility of cure [51]. Moreover, better treatment feasibility is expected in asymptomatic patients in good conditions. This rationale led to the design of clinical trials for the treatment of high-risk SMM (Table 3).     In the phase III randomized QuiRedex study, 119 high-risk SMM patients [52] received lenalidomide-dexamethasone (Rd) vs. observation. After a median follow-up of 75 months, the median TTP was not reached in the Rd group (n = 57) vs. 23 months in the observation group (n = 62, HR 0.24, p < 0.0001). An advantage in OS in the Rd arm was also detected (HR 0.43, p = 0.024). Interestingly, survival was similar in the two groups for patients who had previously received subsequent lines of therapy at the progression to active disease (HR 1.34, p = 0.50). The Rd combination showed acceptable levels of toxicity: Grade ≥3 adverse events (AEs) were infection (6%), asthenia (6%), neutropenia (5%), and skin rash (3%). During treatment, two patients treated with lenalidomide died of infection. A higher rate of second primary malignancies (SPMs) was detected in the Rd group (10%) vs. the observation group (2%). Of note, progression was defined by classical CRAB criteria (hyperCalcemia, Renal failure, Anemia, and Bone lesions) and advanced imaging techniques at screening were not performed, thus suggesting that the study also included patients who would currently be classified as having symptomatic MM.
The efficacy of lenalidomide was also shown in the phase II/III E3A06 study, in which lenalidomide was compared to observation in SMM [62]. After a median follow-up of 35 months in phase III of the trial, the overall response rate (ORR) was 50% in the R group and 0% in the observational group. One-year, 2-year, and 3-year PFS were respectively 98%, 93%, and 91% in the R group, favorably comparing with respectively 89%, 76%, and 66% in the observational arm (HR 0.28, p = 0.002). Among lenalidomide-treated patients, grade 3/4 non-hematologic toxicities occurred in 28% of the phase III patients, with hypertension and infections being most common toxicities. However, no difference in scores regarding the quality of life was noted between the lenalidomide and the observational groups. In this trial, SPMs were detected in 11.4% of lenalidomide-treated patients vs. 3.4% of patients in the observational group.
In this setting, another attractive drug is the second-generation PI ixazomib, which is characterized by a convenient oral administration and shows good safety results. In a phase I study, ixazomib associated with dexamethasone showed good tolerability and high response rate (ORR 64%, PR 57%, and VGPR 14%) [55]. A phase II study exploring the entirely oral triplet ixazomib-lenalidomide-dexamethasone confirmed the good tolerability profile and efficacy of this combination, with a 58% of ≥VGPR (CR 19%, VGPR 34%) [58].
MAbs were also evaluated for the treatment of SMM. The anti-SLAMF7 elotuzumab as single agent showed a low response rate (ORR 10%, minimal response (MR) 29%), with a 2-year PFS of 69%, while first data of the combination with Rd showed a ≥VGPR rate of 43% [56,57].
The phase II CENTAURUS trial evaluated daratumumab alone in 123 patients with three different dose schedules and durations (long intense, intermediate, short intense; Table 3) [60]. At a median follow-up of 25.9 months, ≥VGPR rates were higher in the long intense and intermediate arms compared to the short intense arm (29%, 25%, and 18%, respectively). The 24-month PFS rates were 90%, 82%, and 75% in the three arms. Grade ≥3 AEs occurred in 44% (long intense), 27% (intermediate), and 15% (short intense) of patients. The most frequent grade 3-4 AEs were hypertension and hyperglycemia. The subcutaneous formulation of daratumumab is also being explored in a randomized phase III trial against active monitoring in this setting (NCT03301220). Another anti-CD38 mAb, isatuximab, is under evaluation (NCT02960555).
More intense regimens using three-or four-drug combinations ±ASCT were used in the high-risk SMM setting, aiming at the eradication of MM.
In a small cohort of 12 high-risk SMM patients, Korde and colleagues demonstrated that carfilzomib-lenalidomide-dexamethasone (KRd) induced deep responses (≥CR 100%) and MRD negativity (92% by MFC); after a median follow-up of 15.9 months, none of the patients progressed to MM [53]. Interestingly, the same regimen administered in the NDMM setting produced a lower rate of deep responses (≥CR 56%), suggesting that high-risk SMM patients can be more sensitive to treatment [53].
Another ongoing randomized phase II study (HO147SMM) is comparing KRd to Rd, but no data are available yet.
The addition of the anti-CD38 mAb daratumumab to KRd induction and consolidation is being evaluated in the ASCENT study (NCT03289299), which is now recruiting patients. A randomized comparison of daratumumab-Rd vs. Rd in the context of high-risk SMM is also ongoing (NCT03937635).

Treatment of Symptomatic NDMM
The first efforts aiming at a curative approach in NDMM were done by the University of Arkansas group in the 1990s, developing a program called Total Therapy (TT) using a series of non-cross-resistant induction regimens, 2 cycles of high-dose chemotherapy, followed by ASCT and maintenance treatment [63]. Toxicity concerns and the unavailability of novel agents hindered the success of this approach, although the long-term follow-up of treated patients (median 21 years) showed a plateau in the survival curves with an estimated cure rate of 9% based on PFS data and of 18% based on the duration of CR [64].
Currently, general treatment approaches in NDMM patients are tailored upon their eligibility for high-dose therapy (HDT) and ASCT [65].

ASCT-Eligible Patients
The current therapeutic approach includes sequential induction therapy and ASCT ± consolidation, followed by maintenance. Induction is typically administered for 4-6 cycles prior to ASCT [66]. The introduction of the PI bortezomib increased the response rate compared to classical chemotherapy [67], and is now a backbone of treatment. The addition of a third drug to the bortezomib-dexamethasone (Vd) combination (i.e., thalidomide [VTd], cyclophosphamide [VCd], lenalidomide [VRd], doxorubicin [PAD]) increased the depth of response [68]. In a head-to-head comparison, VTd was superior to VCd as induction prior to HDT-ASCT in terms of ORR (92% vs. 83%) and ≥VGPR rate (66% vs. 56%) [69], demonstrating that even with first-generation novel agents, the combination of a PI and an IMiD was beneficial.
No randomized trial directly compared VRd vs. VTd induction, although a recent integrated analysis of French and Spanish trials was performed (VRd: PETHEMA, GEM 2012, and IFM 2009; VTd: GEM2005 and IFM 2013-04) [79]. In the Spanish studies, after 6 cycles of induction, the ≥VGPR rate was 66.3% vs. 51.2% (p = 0.003) in VRd vs. VTd groups. In the French studies, after 4 cycles of induction, the ≥VGPR rate was similar between VRd vs. VTd groups (57.1% vs. 56.5%). The safety profile of VRd was better than that of VTd in both Spanish and French studies, with a lower rate of polyneuropathy (PNP).
High-dose melphalan (200 mg/m 2 , MEL200) followed by ASCT is currently a standard approach in transplant-eligible patients, due to the longer PFS showed in randomized clinical studies comparing ASCT vs. novel agent-based therapy [50,74,80,81], but the role of double vs. single ASCT is still an open issue. The EMN02/HO95 phase III trial showed a benefit in the double ASCT arm in terms of PFS (3-year PFS 73% vs. 64% in double vs. single ASCT); this effect was particularly evident in the high cytogenetic risk group, where an OS benefit was also noticed [82]. Similarly, in a meta-analysis including three phase III trials, after a median follow-up of 10 years, double ASCT was significantly better than single ASCT in terms of PFS and OS. Consistent with the EMN02/HO95 data, the benefit was particularly evident in the high-risk group [82], suggesting that, in this patient population, a double ASCT is advisable. Nevertheless, the STAMINA trial did not show any difference in PFS or OS of patients receiving double vs. single ASCT. It is always difficult to perform comparisons between different trials, but the better and prolonged induction (VRD) used in the majority of the patients enrolled in the STAMINA study (whereas 3-4 cycles of VCD were used in the EMN02 study) and the lower compliance to the second ASCT procedure reported in the same study can partially explain the different results [83].
Many trials explored consolidation regimens with the rationale to deepen patient response. In the most recently published PETHEMA study, VRd induction, ASCT, and VRd consolidation produced a ≥CR rate of 58% (46% sCR, 12% CR) [75]. These data are consistent with the IFM phase II and phase III studies using VRd consolidation. In the IFM2009 study, VRd consolidation after VRd induction and ASCT showed a similar trend, with the ≥CR rate increasing from 27% during the induction phase, to 47% after ASCT to 50% after consolidation (sCR 40%, CR 10%) ( Table 4) [70]. Response deepened over time, as well as MRD negativity. In the PETHEMA study, MRD by NGF with a cut-off sensitivity of 3 × 10 −6 progressively increased from 34.5% post-induction to 53.4 % post-ASCT, to 58% after consolidation [75]. The phase III STAMINA and EMN02/HO95 trials were designed to evaluate the role of consolidation vs. no consolidation in a randomized fashion. In the STAMINA trial, the 38-month probability for PFS was respectively 58% with single ASCT + VRd consolidation, 58% with tandem ASCT and no consolidation, and 53% with single ASCT and no consolidation, with no statistical differences [83]. In the EMN02/HO95 study, VRd consolidation after ASCT/bortezomib-melphalan-prednisone (VMP) showed a PFS advantage, with a 5-year PFS of 48% in the VRd consolidation arm and 41% in the no consolidation arm [83,84].
In transplant-eligible patients, maintenance therapy is the standard approach after ASCT ± consolidation. A meta-analysis of three phase III trials randomizing patients to lenalidomide vs. observation/placebo showed a significant benefit in the lenalidomide arm in terms of PFS (median, 53 months vs. 24 months, HR 0.48; p = 0.001) and OS (not reached (NR) vs. 86 months, HR 0.75; p = 0.001) [85]. More recently, the Myeloma XI study confirmed the advantage of lenalidomide maintenance vs. observation after ASCT (median PFS 57 vs. 30 months, HR 0.48, p < 0.0001; 3-year OS 87.5% vs. 80.2%, HR 0.69, p = 0.01) [86]. Maintenance with lenalidomide can also further deepen the response, with 27-30% of MRD-positive patients becoming MRD-negative during treatment [87]. Besides its efficacy, the tolerability of continuous lenalidomide maintenance is an important issue. In the meta-analysis, about 30% of subjects receiving lenalidomide experienced a treatment-related AE that led to discontinuation. Moreover, a higher incidence of SPMs in the lenalidomide arm was reported, although it was outweighed by the advantage of a better disease control [85]. Although the optimal duration is currently considered to be until progressive disease, the median actual duration is generally around 2-3 years [85], with retrospective data showing a benefit in patients continuing the drug for at least 2 years [88,89]. However, there are currently no randomized prospective data showing evidence that lenalidomide until progressive disease is better than its administration for a prolonged but fixed duration [74].
Maintenance with lenalidomide alone showed conflicting results in high-risk patients [85,86], and the addition of PIs was suggested to be beneficial [90]. In a phase III trial [91], long-term treatment with bortezomib showed to abrogate the negative effect of deletion 17p [92][93][94]. Moreover, in a randomized study, the administration of the second-generation PI ixazomib as post-ASCT maintenance improved PFS compared to placebo and showed a similar effectiveness for both standard-and high-risk patients [95].
The high rate of deep responses (CR and MRD negativity) obtained after this sequential first-line treatment could further be improved by incorporating the second-generation irreversible PI carfilzomib or adding a fourth drug class, such as the anti-CD38 mAbs.
Despite similar MRD negativity rates, a lower number of early relapsing patients was noted in the KRd-ASCT arm than in the KRd-12 arm. This was observed in intermediate + high-risk patients, but not in standard-risk patients, suggesting that, despite the use of second-generation PIs upfront, ASCT could still play a role in this patient population [9].
Similarly, the addition of isatuximab to KRd is being investigated in the phase II GMMG-CONCEPT study. In the safety run-in phase (10 patients), the overall safety profile was consistent with those previously reported with KRd and isatuximab. Non-hematologic grade ≥3 AEs were treatment-unrelated cerebral vascular disorder (2 patients), self-limiting ventricular tachycardia (1), and diarrhea (1). Three patients experienced a grade 2 infusion-related reaction (IRR) during the first infusion of isatuximab [100].
Quadruplet regimens not including mAbs may allow to achieve deep responses in the majority of patients, preserving the opportunity to use mAbs after induction ± ASCT in patients not achieving MRD negativity.
In a small group of patients, the addition of the second-generation oral PI ixazomib to Rd (Ixa-Rd) during induction followed by ASCT or by ixazomib maintenance induced a good response, with 63% of patients achieving ≥VGPR and 12% MRD negativity. However, responses were not as deep as those reached in patients treated with upfront daratumumab or carfilzomib, making Ixa-Rd less appealing from a curative perspective [78].

ASCT-Ineligible Patients
ASCT-ineligible patients are a heterogeneous population. Scores predicting mortality and the risk of treatment toxicity in elderly patients have been assessed. Evidence from clinical trials [101] suggested that frailty-adapted therapies should be applied and that mainly fit patients can benefit from strategies aiming at the deepest possible response, due to higher toxicities with similar therapies in intermediate-fit/frail patients that in the end hamper the effectiveness of treatment itself [102,103].
The standard first-line treatment schemes for elderly patients are VMP, Rd, and VRd. In the phase III VISTA trial, VMP was superior to melphalan-prednisone (MP) in terms of CR rate, PFS, and OS (median 56 months vs. 43 months) [104,105].
Continuous Rd significantly increased PFS and OS compared to MPT and also prolonged PFS (but not OS) compared to Rd18 (median PFS 26 months for Rd vs. 21 months for Rd18 and 21.9 months for MPT; 4-year estimated OS 59% for Rd vs. 56% for Rd18 and 51% for MPT). Rd was also generally better tolerated than MPT [106]. In a phase III clinical trial specifically designed for intermediate-fit patients, according to the IMWG frailty score, continuous Rd was compared to Rd induction for 9 cycles followed by R maintenance alone at lower doses: PFS was superimposable, with a better tolerability with Rd-R [107].
VRd was also prospectively compared to Rd in the SWOGS0777 trial (Table 5), which, however, was not restricted to elderly patients (median age 63 years) [108]. The addition of bortezomib to Rd resulted in significantly improved PFS (43 months vs. 30 months in the Rd group; p = 0.0018) and OS (75 months vs. 64 months in the Rd group; p = 0.025). Regarding safety, the VRd combination showed higher rates of grade ≥3 AEs (82 vs. 75%), neurological toxicities (33% vs. 11%), and discontinuation (23% vs. 10%). The high neurological toxicity could be due to the two-weekly intravenous infusion of bortezomib used in this trial. In a small phase II study [109], a modified VRd, including lower lenalidomide doses (15 mg) and weekly subcutaneous bortezomib ("VRd lite"), produced a median PFS of 35.1 months and fewer toxic effects. Studies exploring the upfront use of anti-CD38 mAbs in transplant-ineligible patients showed deep responses also in this setting. In the ALCYONE trial, the quadruplet daratumumab-VMP (Dara-VMP) was compared to VMP showing a clear advantage in PFS (HR 0.50, 95% CI 0.38-0.65, p < 0.001) [27]. At least CR rates were 42 vs. 24% and MRD negativity rates by NGS were 22.3% vs. 6.2%, respectively. Safety issues mostly consisted of IRRs (overall 27%, grade ≥3 5%) and a high incidence of infections (grade ≥3 pneumonia 11% vs. 4% in the Dara-VMP vs. VMP arms).
Similarly, in the phase III randomized MAIA study, Dara-Rd significantly prolonged PFS as compared to Rd (HR 0.56, 95% CI 0.43-0.73, p < 0.001), with ≥CR rates of 47% vs. 24% and MRD negativity in 24% vs. 7% patients, respectively. The safety profile was similar in the two arms, but the daratumumab group experienced a higher incidence of neutropenia and infections (including pneumonia) than the Rd group. As in the ALCYONE study, IRRs were reported in the daratumumab arm (overall 40%, mostly of grades 1-2 with an incidence of grade ≥3 IRRs of 2.7%) [110].
An ongoing phase I study is investigating isatuximab, in combination with VRd (Isa-VRd): the first report on 22 patients showed good tolerability, with 46% of grade ≥3 AEs, mostly hematologic. Besides, response rates are promising, with MRD negativity rates (10 −6 ) by NGS of 33% and by NGF 18% [111].
The good results from the upfront use of second-generation PIs in the transplant-eligible setting encouraged its exploration in several clinical trials for the treatment of elderly patients. Carfilzomib associated with melphalan and prednisone (KMP) showed promising results in a phase I/II study, with 90% ORR and 58% ≥VGPR rates and about 8% of grade ≥3 cardiovascular AEs [112]. However, in the phase III CLARION study, KMP failed to outperform VMP in terms of PFS, OS and MRD negativity rates [113]. The safety profile was different between the two arms, with KMP inducing more acute renal failure (any grade 13.9% vs. 6.2%), more cardiac failure (any grade 10.8% vs. 4.3%), and less peripheral neuropathy (grade ≥2 2.5% vs. 35.1%) than VMP.
The association of carfilzomib with cyclophosphamide dexamethasone (KCd) was evaluated in two phase I/II studies, the first adopting the once-weekly carfilzomib schedule and the second the twice-weekly schedule [114,115]. Both trials demonstrated a high efficacy profile (median PFS 35.7 and 35.5 months, respectively; 3-year OS: 72% and 75%) with acceptable toxicity. Overall toxicities mainly occurred during the induction phase and the incidence of non-hematologic AEs was similar to that observed with the KMP combination. KCd showed a lower myelotoxicity than KMP and VMP. Of note, few AEs emerged during maintenance. The once-and twice-weekly schedules were compared in a meta-analysis, with no significant differences in terms of efficacy and toxicities, and with a benefit also observed in high-risk patients [116].

Future Perspectives
Patient fitness is one of the first factors to consider when planning the treatment strategy. Despite the manageable profile of some effective combinations, frail patients can unlikely tolerate full-dose combinations that may induce high MRD negativity. In these patients, disease control rather than cure may be the more realistically achievable goal. Nevertheless, disease control lasting for a few years, even without achieving CR or MRD negativity, could allow very elderly patients to have the same survival of age-matched healthy subjects, considering their actual life expectancy. On the other hand, in fit patients, the outcome-limiting factor is usually disease progression, and a curative approach aiming at sustained MRD negativity could be pursued. This approach should incorporate baseline risk evaluation and dynamic risk evaluation (MRD achievement and duration) during treatment (Figure 1). Baseline risk factors such as International Staging System (ISS), cytogenetics, lactate dehydrogenase (LDH) levels [117], extramedullary disease [118], circulating plasma cells [119], TP53 mutations [94], and many others can help define our therapeutic strategy. For instance, the use of double ASCT and long-term treatment with a PI plus IMiDs maintenance could be beneficial in the presence of high-risk cytogenetics [92]. The dynamic evaluation of patient risk after the start of treatment can also help tune treatment intensity. Many MRD-driven therapeutic choices are under investigation in clinical trials. One possibility is to evaluate treatment escalation in patients who do not achieve MRD negativity at a pre-specified time point. In particular, this could be the approach for high-risk aggressive MM, mirroring a strategy such as the one used for acute leukemia, where achieving MRD is the goal to achieve cure. Another possibility is to evaluate treatment withholding in patients with sustained MRD negativity. This could be the option with standard-risk disease, where the disease behavior is more similar to that of chronic leukemias. In MRD-negative patients, if a reappearance of MRD is detected, restarting prior therapy if previously interrupted, or starting a different second-line therapy before the development of an overt relapse, can also be explored. Of note, the deferral of treatment is currently recommended, even at the reappearance of a monoclonal component (biochemical relapse), if the increase of the monoclonal component is slow [120]. Treating the reappearance of MRD might be a further step for prolonged disease control but its usefulness should be demonstrated in well-designed trials.
Before using MRD in the treatment of MM, several questions need to be answered. The first question is in which patients we should test MRD (in CR or sCR patients only, or in VGPR patients). The rationale to test MRD in VGPR patients is that, due to the long half-life of serum immunoglobulin (~1 month), the complete clearance of monoclonal component could take months until all the cells producing it have been eradicated, especially in IgG cases [121]. In these cases, VGPR patients who are MRD-negative in the bone marrow achieve CR in the months following MRD testing. However, MM is a spatially heterogeneous disease and residual plasma cells in extramedullary sites can produce the monoclonal component in VGPR cases in the presence of MRD negativity in the bone marrow. If this is the case, MRD should probably be measured at sCR and the confirmation of bone marrow MRD negativity with imaging techniques should be performed. In the context of a MRD-driven therapy, it is also tricky to evaluate (a) the impact and the likelihood of "false negative" or "false positive" MRD results; (b) the right time point; (c) the reasonably achievable cut-off at a specific time point. For instance, in the transplant-eligible setting, the post-induction time point could be used to investigate different durations and/or intensifications of induction regimens, and to understand whether or not intensification with transplant is necessary. It is possible that, after different treatments, different cut-offs can be achievable. As an example, a 10 −5 negativity can be the reasonable goal after induction, but with prolonged intensification (transplant or further consolidation), a deepest MRD negativity should likely be the goal. This means that different MRD cut-offs at different time points should be considered in planning MRD-driven treatment strategies. Moreover, the question is if treatment decision can rely on a single MRD evaluation, or if, as for all the other response categories, MRD needs to be confirmed. This also affects the choice of the best time point (for instance, can we reasonably use a post-induction time point to make decisions on treatment intensification if we consider an induction with 4 cycles only?). Secondly, we should also consider the importance of MRD duration, particularly in the context of continuous therapy.
In the transplant-eligible setting, a further issue is the role of checking and the feasibility of pursuing MRD negativity in the peripheral blood stem-cell collection. Autografts contaminated with MM cells (MRD-positive autografts at a sensitivity of 10 −7 by NGS) predicted a worse PFS than MRD-negative autografts [122]. However, this effect was mitigated in patients receiving further treatment after ASCT. Indeed, it should be noted that ASCT, consolidation and maintenance, especially with drugs not used during induction, still have the potential to eradicate MRD in a substantial number of patients who are MRD-positive at post-induction time point [7,8,43,70,123].
While the achievement of MRD negativity is clearly predictive of good outcomes, some NDMM patients are characterized by an MGUS-like plasma cell compartment [124]. In these patients, long-term disease control can be accomplished without achieving deep responses, probably due to an immune control of the residual disease. Further research is needed to reliably identify this patient population.
Moreover, the restoration of a physiological immune system at the time of MRD assessment could also play a role in predicting the patients that will possibly not relapse [28].

Conclusions
In conclusion, we are living exciting times in the field of MM, with many new regimens and strategies in the pipeline and an increasing knowledge of the complexity of the disease. Even if we currently do not have any evidence that we are able to cure MM in the great majority of treated patients, the longer follow-ups of the recent studies will determine the percentage of subjects able to actually maintain a disease-free status for a very long time. New well-designed MRD-driven trials will help us determine if it will be worth aiming at the cure of the disease and what will be the best therapeutic approach to achieve it.

Author Contributions:
Substantial contributions to the conception or design: All the authors. Acquisition, analysis, or interpretation of data: All the authors. First draft: M.D., L.B., S.O., and F.G. Supervision: M.B. and F.G. Critical revision for important intellectual content: All the authors. Final approval of the version to be published: All the authors. Agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved: All the authors.

Funding:
No funding was provided for this contribution.
Conflicts of Interest: S.O. has received honoraria from Amgen, Celgene, and Janssen; and has served on the advisory boards for Adaptive Biotechnologies, Janssen, Amgen, and Takeda. M.B. has received honoraria from Sanofi, Celgene, Amgen, Janssen, Novartis, Bristol-Myers Squibb, and AbbVie; and has received research funding from Sanofi, Celgene, Amgen, Janssen, Novartis, Bristol-Myers Squibb, and Mundipharma. F.G. has received honoraria from Amgen, Bristol-Myers Squibb, Celgene, Janssen, Takeda, and served on the advisory boards for Amgen, Bristol-Myers Squibb, Celgene, Janssen, Roche, Takeda, and AbbVie. The other authors declare no competing financial interests.