High-risk multiple myeloma: how to treat at diagnosis and relapse?

A 48-year-old man with newly diagnosed (ND), Revised International Staging System (R-ISS) III (ISS III plus del(17p)), Bence-Jones κ multiple myeloma (MM) sought treatment and consultation for MM that had been diagnosed in another institution. The patient had an active lifestyle, and the workup showed mild anemia and small lytic lesions in the pelvis and femora as myeloma-defining events. His Eastern Cooperative Oncology Group performance status was 1. He was treated with 6 induction cycles of lenalidomide, bortezomib, and dexamethasone (RVd), achieving a very good partial response, followed by high-dose melphalan and autologous stem cell transplantation (HDM-ASCT), achieving stringent complete remission (sCR) with minimal residual disease (MRD) positivity. He rejected a second ASCT and proceeded to consolidation with 2 cycles of lenalidomide, carfilzomib, and dexamethasone, achieving sCR and MRD negativity. Maintenance with lenalidomide was prescribed. Twelve months after starting maintenance therapy, relapse occurred with reappearance of the M-component in urine. He was included in a clinical trial and treated with B-cell maturation antigen (BCMA) chimeric antigen receptor T (CAR-T) cells, and a new sCR and MRD negativity were achieved. The patient continues in follow-up.

Table 1 summarizes the most relevant patient- and disease-based factors to define high-risk patients.

For a long time, chronological age influenced treatment decisions, and the outcome was poor for the elderly. The International Myeloma Working Group (IMWG), through a pooled analysis including 869 ND elderly patients enrolled in clinical trials, built a simplified geriatric score based on age, comorbidities, and cognitive and physical conditions to distinguish among fit (score  =  0), intermediate fitness (score  =  1), and frailty (score ≥ 2). Frail patients showed a significantly shorter overall survival (OS; 57% at 3 years) than unfit (76% at 3 years) and fit patients (84% at 3 years).¹  Many clinical trials are using an IMWG-modified frailty model to identify frail patients because of the importance of their identification for the treatment decision-making process.² 

The presence of extramedullary disease (EMD) or plasma cell leukemia (PCL), primary or secondary, is infrequent, but these are considered high-risk features not only because the plasma cells escape from the bone marrow environment but also because patients with EMD or PCL are difficult to treat, with poor outcomes with the current therapies (3-year survival rate is 35% for EMD³  and median OS is 12 months for PCL⁴ ).

The IMWG currently recommends the detection of t(4;14), t(14;16) and del (17/17p) in selected plasma cells by interphase fluorescent in situ hybridization for the identification of high-risk patients.⁵  In newly diagnosed MM (NDMM), the presence of at least 1 high-risk cytogenetic abnormality (CA) is associated with a median OS of 24.5 months, significantly shorter than the 50.5 months observed when there are not any CAs (P  <  .001). This is far from complete and requires being updated.

The gain of the long arm of chromosome 1 (+1q) is a frequent CA observed in approximately 30% of NDMMs and associated with poor outcome. A retrospective study conducted in 201 patients with NDMM treated with RVd reported that patients harboring +1q had a shorter median progression-free survival (PFS; 41.9 months) and OS (not reached) compared with those without +1q (PFS of 65.1 months, P  =  .002 and OS not reached, P  =  .003). The negative impact on survival with +1q may be more profound if there is amplification of 1q, defined by the presence of 4 or more copies of chromosome 1q (median PFS of 25.1 months) or in association with other high-risk CAs (median PFS of 34.6 months). The poor prognosis of patients who have +1q with a coexisting deletion of chromosome 1p (del(1p)) has also been described.⁶ 

Although it is well accepted that del(17p) is the CA with more prognostic impact in MM, some questions are under debate: (1) What is the optimal threshold to predict poor prognosis? (2) Is TP53 an optimal molecular target? (3) What about mono- or biallelic deletion and/or inactivation of TP53 through mutations? Patients with a “double-hit” biallelic inactivation of TP53 are at high risk, especially if this abnormality coexists with ISS 3 (1.5-year PFS of 33%) in ND patients.⁷  The Intergroupe Francophone du Myélome (IFM) has recently reported in a large series of patients with NDMM that the presence of isolated del(17p) was also associated with a poor outcome, although the poorest outcome was reported for the double-hit patients.⁸ 

In summary, the optimal identification of high-risk MM based on CA is under construction, but in clinical practice, the CA recommended by the IMWG, together with abnormalities of chromosome 1 and mutational status of TP53, if possible, would be necessary to define the risk at baseline. At the moment of the relapse, it would be optimal to repeat these evaluations because of the clonal evolution and the potential acquisition of new CAs not detectable at baseline.

The ISS risk model, including albumin and β2-microglobulin levels, was improved with the incorporation of 2 well-known disease-related prognostic biomarkers, CA and serum lactate dehydrogenase (LDH) levels, which are associated with a higher proliferative activity, resulting in the R-ISS model.

The R-ISS emerged from a large series of patients with NDMM, and 3 subgroups were defined: R-ISS I (n  =  871), including ISS stage I, no high-risk CA, and normal LDH level with a 5-year OS rate of 82%; R-ISS III (n  =  295), including ISS stage III and high-risk CA or high LDH level with a 5-year OS of 40%; and R-ISS II (n  =  1894), including all the other possible combinations and with a 5-year OS of 62%.⁹  This staging system is still valid, although there are some limitations: (1) most patients were assigned to the R-ISS II, including those with high-risk CA but not ISS III, (2) some other high-risk CAs such as +1q were not included, and (3) other, more complex genomic abnormalities such as mutations or inactivation of TP53 were not considered.

In addition to the above high-risk features, how do we recognize those patients with no apparent high risk at diagnosis but who progress within the first 12 to 18 months after an optimal first line of therapy? These patients are functional high risk with poor prognosis, and further investigations are required to unravel if there is a clonal selection or just an inadequate evaluation at diagnosis.

If the identification of high-risk patients with MM is challenging, its management is not easy either. So far, only a few clinical trials have been specifically conducted in this population because the definition is heterogeneous. In addition, high-risk subgroups in clinical trials are quite small to generate solid recommendations.

In 2021, we know that poor prognosis with the presence of high-risk features can be at least improved or even abrogated by the achieving a deep and sustained response over time, which can most likely be obtained through the use of novel therapeutic options.¹⁰ 

At least 3 large meta-analyses support the use of MRD for monitoring the response in MM because of its prognostic value. The most recent one included publications up to June 2019, showing that the achievement of undetectable MRD improved PFS (hazard ratio [HR], 0.33) and OS (HR, 0.45) in comparison with the presence of MRD. Moreover, its prognostic impact was sustained across the different subgroups, including those with some high-risk features such as elderly patients with NDMM, relapsed/refractory patients, or even the presence of high-risk CA.¹¹ 

Of note, the higher the sensitivity threshold for the MRD evaluation and the longer the sustained undetectable MRD over time, the higher the prognostic value.

In addition, the achievement of undetectable MRD can convert risk assessment in MM into something dynamic, and the high risk at baseline can be overcome when undetectable MRD is achieved. In the PETHEMA/GEM2012MENOS65 trial, 458 patients with NDMM had longitudinal assessment of MRD after 6 induction cycles with RVd, autologous transplantation, and 2 consolidation courses with RVd. The 3-year PFS rate for patients with R-ISS I, II, or III was comparable (95%, 94%, and 88%) if MRD was undetectable after treatment. By contrast, outcomes were progressively poor for patients with R-ISS I, II, and III when MRD was detectable, with a 3-year PFS of 62%, 53%, and 28%, respectively, and analogous results were observed when OS was considered. Similarly, outcome of patients with high-risk CA was abrogated when undetectable MRD was achieved.¹² 

One additional aspect needs to be incorporated in the MRD assessment: the MRD evaluation outside of the bone marrow through the use of functional imaging tools such as positron emission tomography/computerized tomography.¹³  Deauville scores to focal lesions less than 4 and bone marrow uptake showing the liver background (Deauville score <4) have been identified as the best cutoff to define positron emission tomography/computerized tomography negativity after therapy and complete metabolic response, as they have been described in at least 2 clinical trials, the FORTE and CASSIOPETT substudy of CASSIOPEIA trial.

The general approach described above is feasible for frail patients, but tolerability and quality of life are crucial to deliver treatments in order to reach depth responses. At least 1 clinical trial has been conducted in unfit and frail patients with NDMM according to the IMWG frailty index, using ixazomib and daratumumab plus very low dose of dexamethasone.¹⁴  Preliminary results are encouraging, with 1-year OS rates of 96% and 74% for unfit and frail patients, respectively. Subgroup analysis recently conducted in the phase 3 trials ALCYONE and MAIA have also shown how the addition of daratumumab to either bortezomib, melphalan, and prednisone (VMP) or Rd (lenalidomida and dexamethasone) has been able to significantly improve the outcome of frail patients compared with VMP or Rd alone, according to a modified IMWG frailty index (Table 2). In the relapsed-refractory setting, other subanalyses of phase 3 clinical trials have reported how carfilzomib at different doses and schedules or the combination of pomalidomide-dexamethasone plus isatuximab is feasible and able to improve the outcome of frail patients.^15,16  Although this information is obtained from clinical trials, the good toxicity profile of all novel agents makes it possible to maintain therapy in the frail population.

Proteasome inhibitors (PIs), immunomodulatory drugs, and corticosteroids are the key treatment elements of patients with MM patients with high-risk CA. For transplant-eligible patients with NDMM, the question about bortezomib or carfilzomib as the optimal PI for high-risk patients remains under debate because the only phase 3 randomized trial comparing bortezomib with carfilzomib did not show any difference, but it did only include patients with t(4;14).¹⁷ 

The use of moAbs (monoclonal antibodies) targeting SLAMF7 and CD38 also has been evaluated in this setting. Although the addition of elotuzumab showed no significant benefit when combined with RVd in the phase 2 SWOG-1211 study,¹⁸  the addition of daratumumab has been shown to improve the outcome of patients with NDMM and RRMM (relapsed refractory multiple myeloma) with high-risk CA in a recent meta-analysis.¹⁹  HDM-ASCT continues to be the standard of care in high-risk NDMM because of its capacity to achieve a higher undetectable MRD rate, and tandem transplant is even considered for this population based on the positive data from the EMN02 trial, confirmed in the STAMINA trial at least in terms of PFS.^20,21  However, tandem transplant may not be necessary with the introduction of moABs, especially with the introduction of cell therapy. Maintenance with lenalidomide is the standard of care to improve the outcome of high-risk patients compared with observation, but it should be improved through the addition of PIs or moABs, with preliminary positive data.^22,23 

In the nontransplant-eligible, high-risk patients with NDMM, daratumumab, lenalidomide, and dexamethasone would be the first choice based on the results reported in the MAIA trial, with a median PFS of 45.3 months compared with 29.6 months in the Rd arm (HR, 0.57).²⁴ 

In the RR (relapse or refractory) setting, the same approach is valid, and the combination of choice for patients with high-risk CA would be those with the higher likely probability of achieving undetectable MRD (Table 2).

Of note, the novel drug melflufen flufenamide has shown promising efficacy in 45 patients with EMD (extramedullary disease) included in the HORIZON trial, with an overall response of 24% vs 30% in patients without EMD.²⁵  Selinexor also may have a role in treating patients with del(17p) based on its mechanism of action and available evidence in some clinical trials.²⁶  Belantamb mafodotin, a BCMA-conjugated moAb, produced a response rate of 33%, which is similar to that reported in patients with high-risk cytogenetics.²⁷  Further studies are required to confirm this efficacy. Table 2 shows the efficacy reported in patients with high-risk CA in the most relevant clinical trials in patients with NDMM and RRMM.

Although no specific trials were performed until very recently, new trials with BCMA-targeted CAR-T cells have focused on this subgroup of patients (Table 3). Some subgroup analysis in phase 3 trials focused on early relapses, defined as those occurring within the first 12 to 18 months after the previous therapy, showing that the addition of carfilzomib to Rd or daratumumab to Rd vs Rd in the ASPIRE and POLLUX trials resulted in a significant benefit for the triple combination compared with Rd. A similar effect has been recently reported with the addition of daratumumab to bortezomib or carfilzomib^28-31  (Table 2).

BCMA is an attractive and extensively studied target for immunotherapy in MM. BCMA-targeting CAR-T cells have demonstrated fast, high, and deep responses in patients with RRMM. The sample sizes across the different trials are so far rather small but have included a great proportion of patients with high-risk features, such as high-risk CA, R-ISS III, EMD, or high tumor burden.^32,33  Ide-cel (idecabtagene vicleucel), indeed, already has been approved by US Food and Drug Administration for RRMM after at least 4 rounds of therapy, including PIs, immunomodulatory drugs, and anti-CD38, and has been evaluated in 128 patients with RRMM after a median of 6 prior lines (84% triple refractory). The ORR (overall response rate) was 73%, including a complete remission rate of 33% and a median PFS of 8.8 months. These efficacy data were sustained in patients with high-risk features, such as EMD (n  =  50), high-risk CA (n  =  45), or high tumor burden (n  =  65)³⁴  (Table 2). Many other BCMA-targeted CAR-T cells are under investigation, and some clinical trials are focused in patients with high-risk features (Table 3). If results are positive, CAR-T cell therapy will rapidly move as the first choice in patients with high-risk features.

Beyond BCMA-targeted CAR-T cells, there are other therapeutic options, such as bispecific moABs targeting not only BCMA but also GPRC5D, FcRH5, and others, under evaluation in patients with RRMM, and their efficacy will be also evaluated in patients with high-risk features.

In summary, in 2021, the identification of high-risk patients with MM continues being an unmet medical need, and the definition should be revisited. Genomics will help us to improve the identification of these patients, as well as the therapeutic advances, to find the best option for them. Considering the achievement of undetectable and sustained MRD can abrogate the poor prognosis of high-risk features, the management of these patients should be response adapted and determined from exposition to sequential treatments based on drugs with a new and different mechanism of action. International effort and research in this regard are required.

María-Victoria Mateos: has received honoraria derived from lectures and advisory boards from Janssen, BMS-Celgene, Amgen, Takeda, Abbvie, Sanofi, Oncopeptides, Adaptive, Roche, Pfizer, Regeneron, GSK, Bluebird Bio, and Sea-Gen.

Borja Puertas Martínez: no competing financial interests to declare.

Verónica González-Calle: has received honoraria from Janssen and Celgene, received research funding from Janssen, and provided consulting or an advisory role for Prothena and Janssen.

María-Victoria Mateos: nothing to disclose.

Borja Puertas Martínez: nothing to disclose.

Verónica González-Calle: nothing to disclose.