Evolving paradigms in the treatment of relapsed/refractory multiple myeloma: increased options and increased complexity.

The use of modern therapies such as thalidomide, bortezomib and lenalidomide coupled with upfront high-dose therapy and autologous stem cell transplant (ASCT) has resulted in improved survival in patients with newly diagnosed multiple myeloma (MM). However, patients with relapsed/refractory multiple myeloma (RRMM) often have poorer clinical outcomes and might benefit from novel therapeutic strategies. Emerging therapies, such as deacetylase inhibitors, monoclonal antibodies and new proteasome inhibitors, appear promising and may change the therapeutic landscape in RRMM. A limited number of studies has shown a benefit with salvage ASCT in patients with RRMM, although there remains ongoing debate about its timing and effectiveness. Improvement in transplant outcomes has re-ignited a debate on the timing and possible role for salvage ASCT and allogeneic stem cell transplant in RRMM. As the treatment options for management of patients with RRMM become increasingly complex, physicians must consider both disease- and patient-related factors in choosing the appropriate therapeutic approach, with the goal of improving efficacy while minimizing toxicity.

Introduction

Multiple myeloma (MM) is a clonal B-cell disorder of terminally differentiated plasma cells that accounts for ≈ 10% of hematologic malignancies.[1, 2] It is the second most common hematologic malignancy in the United States, with an overall incidence rate of 4.4 cases per 100 000 population/year.[3] MM remains largely incurable, thus therapy is initiated when patients are symptomatic with the ultimate goal of improving patients' long-term outcomes.[4] Over the past 10–15 years, the introduction of modern therapies, such as immunomodulatory agents (IMiDs) and proteasome inhibitors (PIs), has led to significant improvements in overall survival (OS).[5] Five-year survival rates have improved, from 34.8% (1998–2001) to 44.6% (2006–2009), in both transplant-eligible and transplant-ineligible patients, primarily due to treatment advancements in newly diagnosed MM (NDMM).[6, 7, 8, 9]

Patients with relapsed and refractory multiple myeloma (RRMM) present a therapeutic challenge. This heterogeneous group of patients has been defined by the International Myeloma Workshop Consensus Panel as having either primary refractory, refractory, relapsed, both relapsed and refractory, or double refractory MM (Table 1).[10] In contrast to NDMM, RRMM responds poorly to IMiDs and PIs. Kumar et al.[11] showed a median survival of only 1.5 years in patients with relapsed MM, and their clinical course was typified by decreasing response duration with increasing number of salvage regimens. Patients with double-refractory MM, who are refractory to both IMiDs and PIs, have an even poorer prognosis, with a median OS and progression-free survival (PFS) of 9 and 5 months, respectively.[5]

Table 1

Definitions of relapsed and refractory disease in multiple myeloma

Category	Definition
Primary refractory multiple myeloma	Nonresponsive patients who have never achieved minimal response or better with no significant change in M protein concentration and no evidence of clinical progression
Refractory multiple myeloma	Nonresponsive while on primary or salvage therapy or progresses within 60 days of last therapy
Relapsed multiple myeloma	Previously treated myeloma that progresses and requires the initiation of salvage therapy but does not meet criteria for either primary refractory myeloma or relapsed and refractory myeloma categories
Relapsed and refractory multiple myeloma	Nonresponsive while on salvage therapy or progresses within 60 days of last therapy in patients who have achieved minimal response or better at some point previously before, then progress in their disease course (e.g., relapsed and refractory to bortezomib)
Double-refractory multiple myeloma	Disease refractory to both proteasome inhibitors and immunomodulatory drugs

Current treatment strategies are based on both patient- and disease-related factors, including pre-existing toxicities, comorbidities, prior response, aggressiveness of relapse and cytogenetics.[12, 13, 14] This review provides an overview of the challenges in the care of patients with RRMM, current treatment options, possible role of hematopoietic stem cell transplant (HSCT), emerging therapies and a proposed strategy for the treatment of patients with RRMM.

The challenge in RRMM—inherent drug resistance and clonal evolution

The genomic complexity and clonal evolution of MM over the course of treatment are thought to contribute to drug resistance and relapse (Figure 1).[15] These occur through either the re-emergence of the dominant clone, linearly acquired mutations within the dominant clone or evolution of a prediagnostic clone with newly acquired mutations.[16] Recent advances in genome sequencing have provided evidence of both clonal heterogeneity and shifting clonal dominance over time.[16, 17] Clonal heterogeneity may explain decreased duration of response (DOR) and would warrant sequential therapy with alternate agents. In contrast, patients with a long DOR to their last therapy may have developed a clonal re-emergence and might be responsive to a previous therapy.[18, 19]

Figure 1

Hypothetical example of clonal evolution and heterogeneity in patients with multiple myeloma (MM) over the disease course. MM is characterized by clonal heterogeneity. As the disease progresses and patients receive various therapies, different clones may emerge and become dominant, thus contributing to treatment resistance typical of patients with relapsed/refractory MM. (a) The clonal distribution at diagnosis, prior to treatment with lenalidomide plus dexamethasone. (b) An emergence of clone 2 at relapse, prior to re-treatment with lenalidomide plus dexamethasone. (c) Emergence of a lenalidomide-resistant clone 3. BMSC=bone marrow stromal cells.

Risk stratification in RRMM

The Mayo Stratification for Myeloma and Risk-Adapted Therapy (mSMART) recommendations provide a road map for risk assessment in NDMM and RRMM.[20] These recommendations suggest that prior cytogenetic abnormalities, as determined by fluorescence in situ hybridization or gene expression profiling, play an important role in risk stratification.[20, 21, 22, 23] Time to relapse should also be considered in determination of risk. Patients who relapse after 24 months of primary therapy are considered standard risk and are usually re-treated with the primary regimen. Patients who relapse after <12 months are classified as high risk, and new agents are incorporated into their therapy. Of note, mSMART does not specifically address the treatment challenges that develop as a result of phenotypic diversity and clonal heterogeneity in RRMM.

Current treatment options for RRMM

Patients with RRMM and rapidly increasing monoclonal (M) protein concentration with or without associated symptoms should be considered for salvage therapy. Current treatment standards for RRMM include (1) salvage chemotherapy, (2) salvage autologous stem cell transplant (ASCT), (3) allogeneic HSCT and (4) post-transplant consolidation/maintenance therapy. Given the concern for acquired drug resistance and clonal evolution of disease, upfront clinical trials incorporating emerging therapies, with/or without addition of stem cell transplant, are being increasingly employed.

Salvage chemotherapy options in RRMM

Monotherapy and combination therapies

IMiDs

IMiDs, including thalidomide, pomalidomide and lenalidomide, possess antimyeloma effects via binding to cereblon, a critical component of the E3 ubiquitin ligase complex. This results in enhanced ubiquitination and degradation of Aiolos (IKZF3) and Ikaros (IKZF1), which are important for myeloma cell survival (Figure 2a).[24]

Figure 2

MOA of agents approved or under development for MM. Agents approved or under development for MM target key biological pathways that drive MM cell proliferation and survival. (a) Approved agents include proteasome inhibitors (proteasome inhibitors target the proteasome, which plays a role in the normal degradation and clearance of intracellular misfolded and unfolded proteins. This inhibition leads to protein accumulation and eventual apoptosis), IMiDs (the CRBN E3 ubiquitin ligase complex marks protein with ubiquitin for degradation. The binding of an IMiD to this complex leads to the degradation of two key proteins, Aiolos (IKZF3) and Ikaros (IKZF1), ultimately killing MM cells) and DAC inhibitors (DAC inhibitors target proteins in the nucleus and cytoplasm. HDACs deacetylate target nuclear proteins implicated in gene regulation, including histones and tumor suppressor genes. DACs, which target cytoplasmic proteins, namely HDAC6, play a role in protein metabolism through the formation of aggresomes that transport proteins to be degraded by lysosomes. DAC inhibitors target HDAC6, blocking aggresome formation and subsequent protein degradation, thus leading to protein accumulation and apoptosis). (b) Agents under development: CAR-T cells (CAR-T cells are engineered to recognize target tumor cells and induce cell death), mAbs (mAbs utilize antibody-dependent cellular toxicity (targeting of cell surface proteins such as CS1 and CD38) to induce apoptosis; antibody drug conjugates (e.g., indatuximab ravtansine) target cells expressing the recognized receptor, leading to receptor internalization and release of cytotoxic chemotherapy and cell death), oncolytic virotherapy (viruses stimulate MM apoptosis through many complex mechanisms, including direct virus-mediated cytotoxicity and indirect enhancement of immune responses) and KSP inhibitors (KSPs facilitate early mitosis by separating microtubules. KSP inhibitors block this process, thereby serving as antimitotic agents in rapidly dividing MM cells). Adapted with permission from Novartis Pharmaceuticals Corporation. A=antigen; ADCC=antibody-dependent cell-mediated cytotoxicity; CAR=chimeric antigen receptor; CRBN=cereblon; HDAC=histone deacetylase; HSP90=heat-shock protein 90; i=inhibitor; IMiD=immunomodulatory drug; KSP=kinesin spindle protein; mAb=monoclonal antibody; MM=multiple myeloma; MOA=mechanism of action; NK=natural killer; TCR=T-cell receptor.

Thalidomide was the first IMiD evaluated in patients with RRMM. A systematic review of phase II studies demonstrated the single-agent efficacy of thalidomide in RRMM, with an overall response rate (ORR; defined as partial response (PR) or better) of 30% and a median OS of 14 months.[25] PFS and 3-year survival probabilities were significantly improved when patients with RRMM were treated with thalidomide/dexamethasone compared with conventional chemotherapy (PFS, 17 vs 11 months (P=0.02); 3-year survival probability, 60% vs 26% (P=0.002)).[26] Thalidomide-associated peripheral neuropathy (PN) and venous thromboembolism are the main side effects seen with prolonged use.

Lenalidomide is a more potent thalidomide derivative. Phase I and II trials demonstrated single-agent efficacy of lenalidomide in patients with RRMM, with PR rates of 24–29%.[27] MM-009/MM-010 phase III trials demonstrated superior PFS and OS in patients with RRMM receiving lenalidomide/dexamethasone compared with placebo/dexamethasone (Table 2).[28, 29]

Table 2

Selected published clinical trials for the treatment of relapsed/refractory multiple myeloma

Study phase, N	Treatment	ORR (⩾PR), %	PFS, months	OS, months	Safety, percentage of grade ⩾3 AEs
APEX[38] Phase III, N=669	BTZ vs D	38 vs 18 (P<0.001)	6.2 vs 3.5 (P <0.001)	80% vs 66% (P=0.003) 1-year survival rate	Thrombocytopenia (30 vs 6), neutropenia (14 vs 1), anemia (10 vs 11)
MM-009[29] Phase III, N=353	Len/D vs Pbo/D	60.2 vs 24 (P<0.001)	11.1 vs 4.7 (P<0.001)	29.6 vs 20.2 (P<0.001)	Neutropenia (41 vs 5), thrombocytopenia (15 vs 7), VTE (15 vs 4)
MM-010[28] Phase III, N=351	Len/D vs Pbo/D	60.2 vs 24.0 (P<0.001)	11.3 vs 4.7 (P<0.001)	NR vs 20.6 (P=0.03)	Neutropenia (30 vs 2), thrombocytopenia (11 vs 6), VTE (11 vs 5)
MMY-3021[37] Phase III, N=222	Subcutaneous BTZ vs Intravenous BTZ	42 for both (4 cycles)	10.4 vs 9.4 (P=0.39)	72.6% vs 76.7% (P=0.50) 1-year survival rate	Thrombocytopenia (13 vs 19), neutropenia (18 vs 18), anemia (12 vs 8)
PX-171-003-A1[41] Phase II, N=266	Carfilzomib	23.7	3.7	15.6	Thrombocytopenia (29), anemia (24), lymphopenia (20)
MM-002[31] Phase II, N=221	Pom/D vs Pom	33 vs 18	4.2 vs 2.7	16.5 vs 13.6	Neutropenia (41 vs 48), anemia (22 vs 24), thrombocytopenia (19 vs 22)
MM-003[32] Phase III, N=302	Pom/Low D vs High D	31 vs 10	4.0 vs 1.9 (P<0.0001)	12.7 vs 8.1 (P=0.03)	Neutropenia (48 vs 16), anemia (33 vs 37), thrombocytopenia (22 vs 26)
PANORAMA 1[50] Phase III, N=768	Panobinostat/BTZ/D vs Pbo/BTZ/D	61 vs 55 (P=0.09)	12.0 vs 8.1 (P<0.0001)	33.6 vs 30.4 (P=0.26)	Thrombocytopenia (67 vs 31), Lymphopenia (53 vs 40), diarrhea (26 vs 8)
ASPIRE[54] Phase III, N=792	Carfilzomib/Len/Low D vs Len/Low D	87.1 vs 66.1 (P<0.001)	26.3 vs 17.6 (P=0.0001)	NR	Diarrhea (4 vs 4), fatigue (8 vs 6)

Abbreviations: AE=adverse event; BTZ=bortezomib; D=dexamethasone; High=high-dose; Len=lenalidomide; Low=low dose; NR=not reached; ORR=overall response rate; OS=overall survival; Pbo=placebo; PFS=progression-free survival; Pom=pomalidomide; PR=partial response; VTE=venous thromboembolism.

Pomalidomide is another potent derivative of thalidomide. A phase I study of single-agent pomalidomide in patients with RRMM was efficacious (ORR, 21% PFS, 4.6 months; OS, 18.3 months).[30] The phase II  MM-002 study demonstrated an improvement in PFS with pomalidomide/low-dose dexamethasone compared with pomalidomide alone in patients with double-refractory MM (Table 2).[31] The pivotal phase III MM-003 trial demonstrated a significantly longer PFS and OS with pomalidomide/low-dose dexamethasone compared with high-dose dexamethasone (Table 2).[32] An ongoing phase III trial (OPTIMISMM, MM-007) is evaluating the safety and efficacy of pomalidomide in combination with bortezomib and dexamethasone (Table 3).[33]

Table 3

Selected ongoing clinical trials in the treatment of relapsed/refractory multiple myeloma

Study	Treatment	Primary end point	Estimated study completion date
NCT01985126[107] Phase II, N=126	Daratumumab	ORR	October 2016
OPTIMISMM/MM-007 (NCT01734928)[33] Phase III, N=782	Pom/BTZ/D vs Pbo/BTZ/D	PFS	January 2017
NCT02136134[114] Phase III, N=480	Daratumumab/BTZ/D vs Pbo/BTZ/D	PFS	March 2017
ELOQUENT-2 (NCT01239797)[115] Phase III, N=640	Elotuzumab/Len/D vs Pbo/Len/D	PFS	March 2018
ENDEAVOR (NCT01568866)[116] Phase III, N=898	Carfilzomib/D vs BTZ/D	PFS	March 2019
NCT01564537[100] Phase III, N=703	Ixazomib/Len/D vs Pbo/Len/D	PFS	May 2019
NCT02076009[108] Phase III, N=560	Daratumumab/Len/D vs Pbo/Len/D	PFS	September 2020

Abbreviations: BTZ=bortezomib; D=dexamethasone; Len=lenalidomide; ORR=overall response rate; Pbo=placebo; PFS=progression-free survival; Pom=pomalidomide.

PIs

PIs (i.e., bortezomib and carfilzomib) alter the ability of the proteasome to degrade intracellular proteins that have been targeted for destruction, leading to altered protein homeostasis and plasma cell apoptosis (Figure 2a).[34, 35, 36] Bortezomib was the first PI developed for the treatment of MM. Two phase III trials have demonstrated the efficacy of bortezomib in patients with RRMM (Table 2).[37, 38] In the APEX trial, patients treated with i.v. bortezomib had significantly higher rates of ORR, PFS and 1-year survival compared with high-dose dexamethasone.[38] However, there was a significantly higher rate of PN with bortezomib (all grades: 36% vs 9% P<0.01). The MMY-3021 trial demonstrated that s.c. bortezomib was comparable in efficacy to i.v. bortezomib and resulted in significantly reduced PN (38% vs 53% P=0.04).[37]

Carfilzomib is a second-generation PI that has proven efficacy in RRMM. In the phase II study PX-171-003-A0, carfilzomib achieved an ORR of 16.7%, with a median DOR of 7.2 months.[39] The phase II study PX-171-004 was designed to assess the effects of carfilzomib in bortezomib-naive patients who had received only 1–3 prior lines of therapy. One cohort was treated with 20 mg/m2 per cycle, while a second cohort was treated with 20 mg/m2 in cycle 1 followed by 27 mg/m2 in subsequent cycles. The ORR was 42.4 and 52.2% in cohorts 1 and 2, respectively. The median DOR was 13.1 months in cohort 1 and was not reached in cohort 2.[40] In the phase II study PX-171-003-A1, patients who received a median of 5 prior lines of therapy attained an ORR of 23.7%, with a median DOR of 7.8 months, with carfilzomib.[41] Higher doses of carfilzomib with prolonged infusions are under investigation.[42]

Deacetylase inhibitors

Histone deacetylases (HDACs) have been identified as a relevant therapeutic target in MM (Figure 2a).[34, 43] HDACs mediate epigenetic silencing of tumor suppressor genes in MM cells,[44] and overexpression of HDACs has been shown to be a marker of poor prognosis in patients with MM.[45] In addition to histones, HDACs are known to regulate proteins associated with gene expression, DNA repair and replication, cell cycle, cytoskeleton organization, and chaperone activity.[46] Of significant interest is HDAC6, which regulates α-tubulin-mediated transport of the aggresome to the lysosome, leading to protein degradation. DAC inhibitors (DACi) block HDAC6 activity and subsequent protein catabolism, leading to altered protein homeostasis and cell death (Figure 2a).[34, 47] It has been hypothesized that dual targeting of the proteasome and aggresome pathways through PIs and DACi may be effective in patients with RRMM.[34] Furthermore, DACi have been shown to increase expression of epigenetically silenced tumor suppressor genes in vitro.[44] Panobinostat, a potent pan-DACi, demonstrated synergistic activity when combined with bortezomib or lenalidomide in preclinical studies.[47, 48] Key phase II (PANORAMA 2) and 3 (PANORAMA 1) clinical trials evaluated panobinostat/bortezomib/dexamethasone in patients with RRMM (Table 2).[49, 50] PANORAMA 2 showed that the addition of panobinostat to dexamethasone and bortezomib elicited responses in patients who were previously refractory to bortezomib.[49] In PANORAMA 1, there was a significant improvement in PFS with panobinostat/bortezomib/dexamethasone compared with placebo/bortezomib/dexamethasone (Table 2).[50] PANORAMA 1 was the first phase III trial in a decade to demonstrate a significant and clinically relevant efficacy for an agent with a novel mechanism of action in RRMM. Panobinostat was recently approved by the US Food and Drug Administration for the treatment of relapsed MM patients who had received ⩾2 prior regimens, including bortezomib and an IMiD.

Combination therapies in RRMM

A number of clinical trials have been designed to test the combinatorial effects of novel agents. In a phase II trial, patients with relapsed MM or RRMM were treated with lenalidomide, bortezomib and dexamethasone (VRd); 6% of the patients had received prior lenalidomide. The ORR was 64%, with a median PFS of 9.5 months and OS of 30 months. Common toxicities were sensory neuropathy, fatigue and neutropenia.[51] Another phase II trial compared cyclophosphamide, pomalidomide and dexamethasone (CyPomD) with pomalidomide and dexamethasone in patients with lenalidomide-refractory RRMM. At a median follow-up of 15 months, the ORR was superior in the CyPomD group (65% vs 39% P=0.03), with improved PFS (9.2 vs 4.4 months; P=0.04).[52] A multicenter phase I/II trial of carfilzomib, pomalidomide and dexamethasone (CarPomD) was conducted in patients who had received a median of 6 prior lines of therapy. The ORR was 70%, and the median duration of survival and PFS were 17.7 months and 9.7 months, respectively.[53] In the phase III trial ASPIRE, patients with relapsed MM were randomized to carfilzomib, lenalidomide and dexamethasone versus lenalidomide and dexamethasone. PFS (26.3 vs 17.6 months; P<0.001) and 24-month OS rates (73.3% vs 65% P=0.04) were significantly improved with carfilzomib.[54] Other combination therapies that include pegylated liposomal doxorubicin, bendamustine and melphalan with dexamethasone, IMiDs and PIs are listed in Table 4. Collectively, these studies suggest that combination therapies with novel agents may improve the disease course in RRMM.

Table 4

Selected combination chemotherapy trials for the treatment of relapsed/refractory multiple myeloma

Study phase, N	Treatment	ORR (⩾PR), %	PFS, months	OS	Safety (all grades), %
Pegylated liposomal doxorubicin
Orlowski et al.[117] Phase III, N=640	PLD/BTZ vs BTZ	44 vs 41	9.3 vs 6.5 (P<0.001)	At 15 months: 76% vs 65% (P=0.03)	Nausea (46 vs 37), diarrhea (43 vs 34), neutropenia (35 vs 20)
Berenson et al.[118] Phase II, N=40	PLD/BTZ/Len/Dex	49	9	NR	Fatigue (40), thrombocytopenia (35), neutropenia (35)
Bendamustine
Lau et al.[119] Phase II, N=30	Ben/Thal/Dex	46	19	7.2 months	Anemia (78), neutropenia (83), thrombocytopenia (65), pain (48), infection (48), neuropathy (35)
Lentzsch et al.[120] Phase I/II, N=29	Ben/Len/Dex	76	6.1	NR	Thrombocytopenia (83), neutropenia (79), anemia (59), leukopenia (59), fatigue (45), diarrhea (35), hypocalcemia (31), hypoglycemia (31), nausea (28)
Ludwig H et al.[121] Phase II, N=79	Ben/BTZ/Dex	61	9.7	25.6 months	Infection (66), thrombocytopenia (38), anemia (18)a
Melphalan
Palumbo A, et al.[122] N=24	Mel/Thal/Pre	42	9	14 months	Anemia (100), thrombocytopenia (100), neutropenia (100), neuralgia, (54), infection (21)

Abbreviations: Ben=bendamustine; BTZ=bortezomib; Dex=dexamethasone; Len=lenalidomide; NR=not reached; ORR=overall response rate; OS=overall survival; PFS=progression-free survival; PLD=pegylated liposomal doxorubicin; Pom=pomalidomide; Mel=Melphalan; Thal=Thalidomide; Pre=Prednisone.

Thrombocytopenia and anemia are grade ⩾3.

Salvage ASCT in RRMM

Initial studies revealed that ASCT following high-dose therapy resulted in prolonged control of disease and improved OS;[55, 56] however, additional studies are needed to determine the optimal timing of salvage ASCT in RRMM.[57] Several retrospective, registry-based or single-center experiences of salvage ASCT after a previous ASCT suggest a benefit with this approach, with an approximate ORRs of 65% and PFS and OS approaching 12 and 32 months, respectively (Table 5).[58, 59, 60, 61, 62, 63, 64, 65, 66] A phase III study evaluating the efficacy of salvage ASCT compared with conventional chemotherapy demonstrated improvement in PFS, but not OS. Patients with an adverse cytogenetic risk profile had poorer outcomes, suggesting that ASCT may not be beneficial in this subset of patients.[67] Michaelis et al. showed that 187 patients who underwent a second ASCT for disease relapse/progression, from 1995 to 2008 had a 1-year non-relapse mortality rate of 2%, a complete response (CR) rate of 25% and a PR rate of 43%, while the risk of relapse or progression was 51, 82 and 91% at 1, 3 and 5 years, respectively. Multivariate analysis showed that a longer interval (>36 months) from the first ASCT to relapse or progression predicted decreased risk of relapse or progression following the second transplant (relative risk, 0.63).[63] Overall, the use of salvage ASCT in patients with RRMM appears to improve PFS compared with conventional chemotherapy alone. Patients with more aggressive disease, and those with a prolonged response to the first ASCT, appear to benefit the most. Most of the initial studies published using salvage ASCT are limited by being mainly retrospective, having inherent biases in patient selection, and the use of non-modern induction therapies. Future prospective studies are necessary to assess the true impact of salvage ASCT in the context of current, more effective therapies.

Table 5

Selected studies using salvage ASCT or Allo-SCT in the treatment of relapsed/refractory multiple myeloma

Clinical study	N	Efficacy post-ASCT2	Post-ASCT2 TRM incidence
ASCT
Cook G et al.[59]	106 case-matched pairsa,b,c (ASCT2 vs conventional chemotherapy)	ORR: 64% 4-year OS rate: 32% vs 22% (P<0.001)	Within 100 days: 7% 1 year: 7% 5 years: 12%
Gonsalves WI et al.[60]	98a,b (ASCT2 after ASCT1 performed between 1994 and 2009)	ORR: 86% Median PFS: 10.3 months Median OS: 33 months	4%
Jimenez-Zepeda et al.[61]	81a,b (ASCT2 performed between 1992 and 2009)	ORR, day 100: 97% Median PFS: 16.4 months Median OS: 53 months	Within 100 days: 2.6%
Lemieux E et al.[62]	81b,d (HDT + ASCT2 performed between 1995 and 2009 after frontline or tandem ASCT)	ORR: 93% Median PFS: 18 months Median OS: 4 years	0
Michaelis LC et al.[63]	187a,b,c (ASCT2 performed between 1995 and 2008)	1-, 3- and 5-year respective PFS rates: 47, 13 and 5% 1-, 3- and 5-year respective OS rates: 83, 46 and 29%	1 year: 2% 3 years: 4%
Morris C et al.[64]	7452b,c,e (ASCT2 after ASCT1 performed between 1993 and 2002) n=2655 (planned ASCT2) n=4797 (unplanned ASCT2)	Median OS: 61 months (planned) vs 51 months (unplanned)	No ASCT2 before relapse/TRM: HR, 1.00 0–6 months to ASCT2: HR, 3.69 (P<0.001) 6–12 months to ASCT2: HR, 2.97 (P<0.001) >12 months to ASCT2: HR, 11.30 (P<0.001)
Sellner L et al.[66]	200b,e (ASCT2 after ASCT1 and HDT + melphalan reinduction therapy between 1995 and 2010)	ORR, day 100: 80% Median PFS: 15.2 months Median OS: 42.3 months	Within 100 days: 3%
Allo-SCT
Bensinger W et al.[75]	80d (allo-SCT after HDT±modified TBI)	ORR: 59% 4.5-year PFS probability: 20% 4.5-year OS probability: 24%	Within 100 days: 44%
Bjorkstrand BB et al.[76]	189 case-matched pairs (allo-SCT vs ASCT)	ORR: 86% vs 72% (ASCT vs allo-SCT; P=0.001) Median OS: 34 vs 18 months (ASCT vs allo-SCT; P=0.001)	3 years: 41% vs 13% (allo-SCT vs ASCT; P=0.0001)
Qazilbash MH et al.[79]	40f (Allo-SCT vs ASCT2 between 1992 and 2004) n=14 allo-SCT n=26 ASCT2	ORR: 69% vs 64% Median PFS: 7.3 vs 6.8 months Median OS: 13 vs 29.5 months	Within 100 days: 11% vs 7% Overall: 27% vs 14%
Mehta J et al.[78]	42 case-matched pairs (allo-SCT between 1992 and 2006 vs ASCT2)	ORR: 62% vs 81% (P=0.05) 3-year PFS probability: 31±10% vs 72±9% (allo-SCT vs ASCT2; P=0.03) 3-year OS probability: 54±8% vs 29±9% (ASCT2 vs allo-SCT; P=0.01)	1-year probability: 43±8% vs 10±5% (allo-SCT vs ASCT2; P=0.0001)
Efebera Y et al.[80]	51f (allo-SCT between 1996 and 2006)	2-year PFS rate: 19% 2-year OS rate: 32%	1 year: 25%
Coman T et al.[86]	52b,d (Len after allo-SCT between 2006 and 2009)	ORR: 83% Median PFS: 18 months Median OS: 30.5 months	4%

Abbreviations: Allo-SCT=allogeneic stem cell transplant; ASCT1=initial autologous SCT; ASCT2=second autologous SCT; HDT=high-dose chemotherapy; HR=hazard ratio; Len=lenalidomide; ORR=overall response rate; OS=overall survival; PFS=progression-free survival; TBI=total body irradiation; TRM=transplant-related mortality.

Excluded patients participating in a tandem ASCT program.

Retrospective study.

Registry-based study.

Multicenter study.

Included patients participating in a tandem ASCT program.

Reduced-intensity myeloablative conditioning was performed prior to allo-SCT.

Allogeneic stem cell transplant (Allo-SCT) in RRMM

Allo-SCT is potentially curative in patients with MM; however, its role and timing are still subject to debate.[68, 69, 70] Following allografting, the immune reaction of donor T cells against myeloma-specific antigens has resulted in the achievement of complete remissions after the discontinuation of immunosuppression or after the infusion of donor T lymphocytes in patients with recurrent disease post-transplant.[71, 72, 73] A complete molecular remission rate up to 50% has been reported following allo-SCT in MM.[74]

Allo-SCT with myeloablative conditioning has been reported to result in long-term PFS, with a plateau in survival curves suggesting possible curative benefit in some patients.[75, 76] However, transplant-related mortality is an important challenge to this approach.[75, 77, 78] A retrospective case-matched analysis by the European Group for Blood and Marrow Transplantation compared 189 myeloma patients treated with allo-SCT with an equal number of patients who received ASCT. Results showed inferior median survival with allo-SCT compared with ASCT, with a median survival of 18 and 34 months, respectively (P=0.001), mainly attributed to increased transplant-related mortality in allo-SCT (41 vs 13% for ASCT, P=0.0001), not compensated for by a lower rate of relapse and progression (Table 5).[76] Transplant-related mortality may be decreased in patients who receive a reduced-intensity or nonmyeloablative conditioning, allo-SCT following an initial ASCT.[79, 80] In this tandem approach, cytoreduction is achieved with ASCT, and allo-SCT generates a graft-versus-myeloma effect that may be curative. However, results from published studies using this approach appear conflicting in patients with NDMM,[81, 82, 83] attributed to patient heterogeneity, different conditioning regimens and graft-versus-host disease (GVHD) prophylaxis used. Importantly, benefits from an allo-SCT may only become apparent after a long-term follow-up, which most of the published studies lack. However, it remains to be determined whether allo-SCT should be offered as part of the initial therapy in patients with high-risk disease like RRMM or only as a salvage therapy.

Post-transplant consolidation/maintenance therapy in RRMM

Modern therapies with comparatively fewer side effects may be employed as consolidation/maintenance therapy after an autologous and/or allo-SCT. Offidani et al.[84] showed in a phase II study evaluating the effectiveness of bortezomib-based induction therapy, followed by consolidation/maintenance therapy in patients with RRMM, 37% CR, 34.5% very good PR (VGPR) and 4.5% PR with ORR 76%. Maintenance lenalidomide after induction therapy with liposomal doxorubicin, vincristine, dexamethasone and lenalidomide led to an ORR of 53%, a median PFS of 10.5 months and a median OS of 19 months.[85]

In addition, these modern therapies modulate T-cell responses, and play a pivotal role in graft-versus-myeloma effects. Following relapse after allo-SCT, use of lenalidomide with or without dexamethasone has shown efficacy, partially through an immunomodulatory GVHD effect to enhance intrinsic anti-MM activity.[86] Following reduced-intensity or nonmyeloablative conditioning allografting, a cohort of patients with progressive disease treated with lenalidomide developed a flare of GVHD, with improved outcomes.[87, 88]

Preclinical studies also have highlighted the immunomodulatory role of bortezomib on GVHD following an allo-SCT.[89, 90] Bortezomib induced selective depletion of alloreactive T lymphocytes, decreased the production of Th1 cytokines and allowed the emergence of a suppressor T-cell subset.[91, 92] Bortezomib in combination with tacrolimus and methotrexate has been found to be effective for GVHD prophylaxis, following reduced-intensity or nonmyeloablative conditioning allografts from human leukocyte antigen–mismatched unrelated donors.[93] Bortezomib with and without dexamethasone has been shown to be an effective consolidation therapy for prevention of disease relapse, or as a salvage regimen following relapse after an allo-SCT; however, the benefits need to be balanced with potential risks such as increase in infectious complications, worsening of GVHD, and increased neurotoxicity. In a cohort of patients with MM,[94, 95] the use of bortezomib and thalidomide after disease progression following reduced-intensity or nonmyeloablative conditioning allo-SCT and infusion of donor T lymphocytes showed durable responses;[96] the combination of post-transplant immunotherapy with infusion of donor T lymphocytes and novel agents resulted in increased CR (>50%) in patients with only PR after allo-SCT.[97] Treatment of relapsing patients with modern therapies following allo-SCT may enhance the graft-versus-myeloma effect while directly inhibiting tumor growth.[86] Future studies are necessary to determine the efficacy and safety of these strategies in RRMM.

Emerging therapies in RRMM

Novel therapies that target different mechanisms of action, including immunotherapy with monoclonal antibodies, are promising, and will expand our therapeutic armamentarium in the fight against MM. Their favorable safety profiles as monotherapy in patients with RRMM will enable combinatorial use with ASCT/allo-SCT to further improve long-term disease control.

New PIs

Oral PIs (i.e., ixazomib and oprozomib) are currently in clinical development. These agents show promising activity in NDMM as well as in patients with RRMM and are more easily administered (Table 3). Ixazomib has demonstrated enhanced proteasomal inhibition.[98] In a phase II trial evaluating single-agent ixazomib in patients with relapsed MM, the ORR was 34%. Common toxicities included thrombocytopenia, fatigue and nausea, while the incidence of PN was relatively low (18%).[99] An ongoing phase III trial is evaluating the efficacy of ixazomib/lenalidomide/dexamethasone compared with placebo/lenalidomide/dexamethasone in RRMM (Table 3).[100] Also ixazomib maintenance is being investigated post allo-SCT in high-risk MM (BMT CTN 1302). Oprozomib has also demonstrated antitumor activity in several phase I clinical studies.[101, 102] Gastrointestinal toxicities were common, and PN was rare.

Monoclonal antibodies

Elotuzumab is a humanized monoclonal antibody specifically targeting CS1, a glycoprotein highly expressed on the surface of MM cells. Binding of elotuzumab leads to recruitment of natural killer cells and tumor cell death via antibody-dependent cellular cytotoxicity (Figure 2b). Elotuzumab has recently been granted breakthrough status based on results from a phase II study evaluating elotuzumab/lenalidomide/dexamethasone in patients with RRMM. This combination therapy achieved an ORR of 84%, including 14% with stringent CR (sCR)/CR and 43% VGPR). Median PFS was 29 months. Common treatment-related adverse events included gastrointestinal symptoms, muscle spasms, fatigue and upper respiratory tract infection.[103] Notably, efficacy was greatest among lenalidomide-naive patients. A phase III trial (ELOQUENT-2) evaluating elotuzumab treatment in patients with RRMM showed an ORR in the elotuzumab group of 79% versus 66% in the control group (P<0.001), with a relative reduction of 30% in the risk of disease progression or death after a 2-year follow-up (Table 3).[104]

Daratumumab and SAR650984, humanized monoclonal antibodies specific for CD38, can target tumor cells for elimination via antibody-dependent cellular cytotoxicity, complement-dependent cytotoxicity and phagocytosis. Daratumumab may also initiate CD38-mediated signal transduction, leading to cell death. In preliminary studies, daratumumab has demonstrated promising activity in combination with lenalidomide and dexamethasone.[105] Most common adverse events were diarrhea and neutropenia, with no dose-limited toxicities. Median time to response was 4.3 weeks, with 75% PR, 15% CR and 30% VGPR. Daratumumab earned breakthrough designation based on phase I results that demonstrated notable single-agent activity with an acceptable safety profile in heavily pretreated patients.[106] On the basis of these data, two phase III studies are further evaluating daratumumab in RRMM (Table 3).[107, 108] SAR650984 has demonstrated efficacy as a single agent in a phase I study. SAR650984-associated toxicities included fatigue, nausea, pyrexia, cytopenias, cough and gastrointestinal symptoms. The median time to response was 4.6 weeks, with an ORR of 33% in the highest-dose cohort (⩾10 mg/kg).[109] In combination with lenalidomide and dexamethasone, the ORR was 64.5%, with 6% sCR, 26% VGPR and 32% PR; the median PFS was 6.2 months. Of note, over 95% of patients had received prior IMiD therapy. The ORR was 62.5% among patients who relapsed after or were refractory to prior lenalidomide-containing therapy.[109, 110]

Indatuximab ravtansine is a chimeric anti-CD138 monoclonal antibody conjugated to DM4, a maytansinoid cytotoxic agent. After binding to CD138+ myeloma cells, the conjugated antibody is internalized. DM4 is metabolized in the lysosome and released into the cytoplasm, where it inhibits tubulin polymerization, leading to cell death (Figure 2b). Preliminary results from a phase I/IIa trial of indatuximab in combination with lenalidomide and dexamethasone in 30 evaluable patients revealed an ORR of 78%, with 10% sCR/CR and 33% VGPR. The most common adverse events were diarrhea, fatigue, nausea and hypokalemia.[111]

Tabalumab, an anti-B-cell activating factor antibody, inhibits interactions between the myeloma cell and its microenvironment that are necessary for survival.[112] Programmed death 1-specific antibodies, such as pembrolizumab and pidilizmab, enhance the tumor-specific T-cell response.[113] Two clinical trials are currently evaluating the potential for chimeric antigen receptor (CAR) T cells that are engineered to target B-cell maturation antigen or CD138. The small-molecule inhibitors vemurafenib, CPI-0610, ibrutinib, and filanesib (Table 6) and the potent PI marizomib are also in early stages of development. Oncolytic virotherapy (Figure 2b and Table 6) is also emerging as a promising therapy for RRMM. New therapeutic agents currently under development for RRMM are included in Table 6.

Table 6

Emerging therapies

Name	Mechanism of action	Route administered	Trial, phase	Combinations
Proteasome inhibitors
Ixazomib	Reversible PI	p.o. or i.v.	Phase I/II[123] NCT01564537, phase III	Lenalidomide, dexamethasone
Oprozomib	Irreversible PI	p.o.	NCT01881789, phase Ib/II	Lenalidomide, dexamethasone, cyclophosphamide
Marizomib	Irreversible PI	i.v.	NCT02103335, phase I	Pomalidomide, dexamethasone
Monoclonal antibodies
Elotuzumab	Anti–CS1	i.v.	NCT01239797, phase III	Lenalidomide, dexamethasone
Daratumumab	Anti-CD38	i.v.	NCT02076009, phase III	Lenalidomide, dexamethasone
SAR650984	Anti-CD38	i.v.	NCT01749969, phase Ib	Lenalidomide, dexamethasone
Indatuximab ravtansine	Anti-CD138	i.v.	NCT01638936 phase I/IIa	Lenalidomide, dexamethasone
Tabalumab	Anti-BAFF	i.v.	NCT00689507, phase I	Bortezomib
Pembrolizumab	Anti–PD-1	i.v.	NCT02036502, phase I	Lenalidomide, dexamethasone
Pidilizumab	Anti–PD-1	i.v.	NCT02077959, phase I/II	Lenalidomide
Small molecules and signaling pathway inhibitors
Vemurafenib	BRAFV600E	i.v.	NCT01524978, phase II	None
CPI-0610	BET inhibitor	p.o.	NCT02157636, phase I	None
Ibrutinib	Btk inhibitor	p.o.	NCT01962792, phase I/IIb	Carfilzomib, dexamethasone
Filanesib	KSP inhibitor	i.v.	NCT02092922, phase II	None
Other novel therapies
Edmonston strain of measles virus	Oncolytic virotherapy	i.v.	NCT02192775, phase II	None
CAR therapy	Anti-BCMA Anti-CD138	i.v.	NCT02215967, phase I NCT01886976, phase I/II	Cyclophosphamide, fludarabine None

Abbreviations: BAFF=B-cell activating factor; BCMA=B-cell maturation antigen; Btk=Bruton tyrosine kinase; BET=bromodomain and extraterminal; CAR=chimeric antigen receptor; i.v.=intravenous; KSP=kinesin spindle protein; PD-1=programmed death 1; PI=proteasome inhibitor; p.o.=per oral.

Proposed treatment guidelines for RRMM

Introduction of modern therapies, new and emerging therapies, and improvements in transplant outcomes have increased treatment options for patients with RRMM. The challenge for the physician is to choose the appropriate therapy or combination of therapies to improve outcome. Here we propose a road-map for treatment of patients with RRMM that takes into consideration disease- and patient-related factors, prior treatment response, and history of toxicity (Figure 3). When available, we highly recommend enrollment of patients in clinical trials designed to answer unresolved issues in the treatment of patients with RRMM.

Figure 3

Proposed treatment guidelines for management of relapsed/refractory multiple myeloma (RRMM). Treatment decisions are guided by previous therapeutic exposure, comorbidities, risk assessment and disease- and/or treatment-related symptoms. The availability of new therapies has increased the complexity of the treatment algorithm for RRMM. BTZ=bortezomib; C=cyclophosphamide; CFZ=carfilzomib; CTD=cyclophosphamide, thalidomide and dexamethasone; D=dexamethasone; GEP=gene expression profiling; Len=lenalidomide; PLD=pegylated liposomal doxorubicin; PAN=panobinostat; POM=pomalidomide; RIC=reduced-intensity or nonmyeloablative conditioning; T=thalidomide; TD=thalidomide and dexamethasone.

Expert commentary

Despite substantial progress, myeloma remains an incurable disease plagued by multiple relapses and increasing resistance to therapy. The past decade has been marked by the unraveling of the pathobiological process underlying myeloma pathogenesis, the emergence of new therapies and improvement in transplant technology, which paves the way for improved responses even in patients with double-refractory MM. While randomized trials have shown superiority of modern therapies over older regimens like mephalan-prednisone, very few have shown superiority of one modern therapy over the other in terms of OS and patient-reported quality-of-life. Approach to therapy is often dictated by regional availability of drugs, HSCT technology and varied regulatory frameworks that exist in different parts of the world. Thus, there remains marked heterogeneity in how NDMM or RRMM are treated around the world.

Use of new genomic and molecular prognostic tools throughout the disease course may give better insights into the clonal dynamics of the myeloma cell and pave the way for targeted and personalized therapy approaches. Validation of promising biomarkers will help to stratify patients based on risk and potential therapeutic benefit. With the increased complexity of treatment options for patients with RRMM, physicians must understand the guidelines for administration of new agents in the context of the patient's therapeutic and disease history. The strategy provided here may help facilitate a clear path through the complex treatment landscape in RRMM.