Comparative Efficacy of Daratumumab Monotherapy and Pomalidomide Plus Low‐Dose Dexamethasone in the Treatment of Multiple Myeloma: A Matching Adjusted Indirect Comparison

Outcomes remain poor in patients with relapsed and refractory multiple myeloma. This article compares the overall survival of daratumumab monotherapy with pomalidomide plus low‐dose dexamethasone in patients with heavily treated multiple myeloma using a matching adjusted indirect comparison technique.

Abstract

Background.

Daratumumab (a human CD38‐directed monoclonal antibody) and pomalidomide (an immunomodulatory drug) plus dexamethasone are both relatively new treatment options for patients with heavily pretreated multiple myeloma. A matching adjusted indirect comparison (MAIC) was used to compare absolute treatment effects of daratumumab versus pomalidomide + low‐dose dexamethasone (LoDex; 40 mg) on overall survival (OS), while adjusting for differences between the trial populations.

Materials and Methods.

The MAIC method reduces the risk of bias associated with naïve indirect comparisons. Data from 148 patients receiving daratumumab (16 mg/kg), pooled from the GEN501 and SIRIUS studies, were compared separately with data from patients receiving pomalidomide + LoDex in the MM‐003 and STRATUS studies.

Results.

The MAIC‐adjusted hazard ratio (HR) for OS of daratumumab versus pomalidomide + LoDex was 0.56 (95% confidence interval [CI], 0.38–0.83; p = .0041) for MM‐003 and 0.51 (95% CI, 0.37–0.69; p < .0001) for STRATUS. The treatment benefit was even more pronounced when the daratumumab population was restricted to pomalidomide‐naïve patients (MM‐003: HR, 0.33; 95% CI, 0.17–0.66; p = .0017; STRATUS: HR, 0.41; 95% CI, 0.21–0.79; p = .0082). An additional analysis indicated a consistent trend of the OS benefit across subgroups based on M‐protein level reduction (≥50%, ≥25%, and <25%).

Conclusion.

The MAIC results suggest that daratumumab improves OS compared with pomalidomide + LoDex in patients with heavily pretreated multiple myeloma.

Implications for Practice.

This matching adjusted indirect comparison of clinical trial data from four studies analyzes the survival outcomes of patients with heavily pretreated, relapsed/refractory multiple myeloma who received either daratumumab monotherapy or pomalidomide plus low‐dose dexamethasone. Using this method, daratumumab conferred a significant overall survival benefit compared with pomalidomide plus low‐dose dexamethasone. In the absence of head‐to‐head trials, these indirect comparisons provide useful insights to clinicians and reimbursement authorities around the relative efficacy of treatments.

Introduction

Despite the widespread use of proteasome inhibitors (PIs) and immunomodulatory drugs (IMiDs), outcomes remain poor in patients with relapsed and refractory multiple myeloma (MM) [1], [2], [3]. A recent International Myeloma Working Group study determined that patients who were refractory to both a PI and an IMiD had a median overall survival (OS) of 13 months [3]. Additionally, an analysis of real‐world survival outcomes from two U.S. patient databases found that the median OS was 7.9 months for patients who received ≥3 prior lines of therapy (LOTs), including a PI and an IMiD, or who were double refractory to a PI and an IMiD [2].

Daratumumab (DARA) is a human monoclonal antibody targeting CD38, a type 2 transmembrane glycoprotein that is highly and ubiquitously expressed on myeloma cells [4], [5], [6], [7]. The direct, on‐tumor activity of DARA results from several CD38 immune‐mediated actions, including complement‐dependent cytotoxicity, antibody‐dependent cell‐mediated cytotoxicity, and antibody‐dependent cellular phagocytosis, as well as modulation of CD38 enzymatic activity and apoptosis [4], [8], [9], [10]. The mechanism of action of DARA also includes an immunomodulatory component where treatment causes a reduction in immunosuppressive CD38⁺ myeloid‐derived suppressor cells, regulatory T cells, and regulatory B cells, and also an increase in T‐cell clonality [11].

DARA monotherapy has been shown to provide marked clinical benefit for the treatment of heavily pretreated myeloma patients [12]. A pooled analysis included data for patients with heavily pretreated and highly refractory MM who received monotherapy with DARA 16 mg/kg in two open‐label, multicenter, phase I/II studies: GEN501 and SIRIUS [12]. Patients in GEN501 and SIRIUS received a median of 4 and 5 previous LOTs, respectively, and in the pooled analysis, the overall response rate (ORR) was 31%, the median OS was 20.1 months, and the median progression‐free survival (PFS) was 4.0 months. In addition to its single‐agent activity, DARA in combination with other established regimens significantly prolonged PFS, resulting in a >60% reduction in the risk of disease progression or death in two randomized phase III studies evaluating patients with relapsed or refractory MM who received ≥1 prior LOT [13], [14].

Based on data from the GEN501 and SIRIUS studies, the U.S. Food and Drug Administration (FDA) granted conditional accelerated approval in November 2015 to DARA monotherapy for patients with MM who have received ≥3 prior LOTs, including a PI and an IMiD, or who are double refractory to a PI and an IMiD [15], [16]. In May 2016, DARA received conditional approval by the European Medicines Agency (EMA) for the treatment of adult patients with relapsed and refractory MM, whose prior therapy included a PI and an IMiD and who have demonstrated disease progression on the last therapy [17], [18]. DARA also received FDA approval in November 2016 and EMA approval in April 2017 as combination therapy with bortezomib and dexamethasone, or lenalidomide and dexamethasone, for MM patients who have received at least 1 prior LOT based on data from the CASTOR and POLLUX phase III clinical trials [19], [20].

Pomalidomide (POM) is an analogue of thalidomide that stimulates T cells and NK cells, inhibits proinflammatory cytokine production, demonstrates antiangiogenic activity, and inhibits proliferation and induces apoptosis of tumor cells [21]. POM was approved, in combination with dexamethasone, by the FDA in February 2013 and by the EMA in August 2013 for the treatment of adult patients with relapsed and refractory MM who have received ≥2 prior treatment regimens, including both lenalidomide and bortezomib, and have demonstrated disease progression on the last therapy [22], [23], [24], [25], [26]. FDA approval of POM was based on the results of the phase II MM‐002 study [23], and EMA approval was primarily based on the findings of the phase III MM‐003 study [22]. In the MM‐003 study, patients received a median of 5 prior LOTs, and POM in combination with low‐dose dexamethasone (LoDex) achieved an ORR of 31%, a median PFS of 4.0 months, and a median OS of 13.1 months in MM patients who failed treatment with bortezomib and lenalidomide [27]. The open‐label, single‐arm, phase IIIb study (STRATUS) also assessed the safety and efficacy of POM + LoDex in patients who received a median of 5 prior LOTs and demonstrated an ORR of 33%, a median PFS of 4.6 months, and a median OS of 11.9 months [28].

In the absence of clinical trials directly comparing DARA monotherapy and POM + LoDex, the efficacy of both treatments was compared using a matching adjusted indirect comparison (MAIC). This approach provides a way of comparing absolute treatment effects while lowering the risk of bias associated with a naïve indirect comparison [29], [30]. The MAIC method was originally developed to adjust for potential biases related to treatment‐effect modifiers when performing a traditional indirect treatment comparison, based on examining relative treatment effects versus a common comparator between comparative trials. The MAIC method attempts to make trial populations comparable by matching the inclusion/exclusion criteria and baseline characteristics across the trials through the reweighting of patients from the trial population for which patient‐level data are available, to mimic the population of the other trial for which only aggregate results are available.

A similar approach can be applied to single‐arm trials or trials without a common comparator. In such situations, which are common for rare diseases, and in oncology trials for patients with poor prognosis, the approach of reweighting of individual patient data (IPD) can be used to adjust comparisons of absolute outcomes between trials for cross‐trial differences and should be preferred over naïve unadjusted comparisons [29], [31]. This approach was recently described by Phillippo et al. as “unanchored” comparisons [32], in contrast to “anchored” comparisons of relative treatment effects between trials with a common treatment arm.

The primary objective of the present study was to compare OS of DARA monotherapy with POM + LoDex in patients with heavily treated MM using the MAIC technique. An additional analysis was performed that compared OS by depth of response (i.e., degree of M‐protein reduction) between DARA monotherapy and POM + LoDex to investigate whether any OS benefit of DARA is related to a particular response category, given that similar ORRs were observed across these studies [12], [28], [33]. Furthermore, comparison of PFS between DARA monotherapy and POM + LoDex was conducted using the MAIC technique.

Materials and Methods

IPD were pooled for patients receiving DARA 16 mg/kg in GEN501 (NCT00574288) [34] and SIRIUS (NCT01985126) [35]. Aggregate published results from phase III studies on POM + LoDex, MM‐003 (NCT01311687) [33] and STRATUS (NCT01712789) [28], were used as a comparator. No approvals were required from an institutional review board/ethics committee for this study.

MAIC Versus MM‐003 and STRATUS

In the absence of trials of DARA and POM + LoDex with a common comparator, an MAIC was performed, in which the absolute treatment effects between DARA and POM + LoDex were compared after matching the trial populations, using the algorithm of Signorovitch [29].

The MAIC method excluded cases according to inclusion/exclusion criteria in the comparator trial based on IPD, and a matching process was followed. IPD from the remaining patients in DARA trials were reweighted such that mean values or distributions of relevant baseline variables matched those reported in respective comparator studies.

Relevant matching parameters were identified based on published literature [36]. Parameters included (in order of relevance to outcome) refractory status to lenalidomide and/or bortezomib, number of prior LOTs, creatinine clearance, ECOG status, time since diagnosis, myeloma subtype (MM‐003 only), race (MM‐003 only), bone lesion presence (MM‐003 only), prior autologous stem cell transplantation (ASCT), and age. Parameter selection depended on whether or not they were measured and reported for the comparator trials.

Time‐to‐event outcomes were compared across matched populations using weighted statistical tests. IPD were simulated for the POM + LoDex arm based on published Kaplan‐Meier curves, using the Guyot et al. algorithm [37]. The simulated IPD were then pooled with reweighted IPD from the DARA studies to calculate the relative treatment effect on OS or PFS for DARA versus POM + LoDex. A hazard ratio (HR; including 95% confidence interval [CI]) was estimated using a weighted Cox proportional hazard model and a robust estimator for variance.

Whereas the POM studies excluded patients with prior exposure to POM, GEN501/SIRIUS did not have the same exclusion criterion. However, the base case analysis does not exclude the POM‐exposed patients from GEN501/SIRIUS due to sample size considerations. A sensitivity analysis was performed using only the POM‐naïve patients from GEN501/SIRIUS.

For the base case analysis, the MAIC was performed using all available relevant baseline characteristics. However, matching on a larger number of baseline characteristics requires more extreme weighting. The effective sample size (N_eff) is a measure of the impact of reweighting on the available statistical information in the IPD [29]. The more extreme the weights, the smaller the N_eff and the higher the uncertainty in the final statistical analysis of the weighted population. As a form of sensitivity analysis, the algorithm was repeated and excluded the least important baseline variables from the matching process. This was performed in a stepwise manner, excluding more baseline characteristics from the MAIC process at each step.

OS Analysis by Depth of Response

OS for DARA and POM + LoDex (MM‐003 only) was compared by depth of response, defined as level of M‐protein reduction (<25%, ≥25%, and ≥50%). MAIC adjustments were not possible here because the detailed baseline characteristics for each subgroup by M‐protein level reduction were not reported for MM‐003 [38].

Results

A total of 148 patients received DARA 16 mg/kg in the GEN501 and SIRIUS trials combined [12], of whom 82 (55.4%) were POM‐exposed and 66 (44.6%) were POM‐naïve.

The observed median OS for patients treated with DARA was 20.1 months compared with 13.1 months for patients treated with POM + LoDex in MM‐003 and 11.9 months in STRATUS (Fig. 1) [12], [27], [28].

Figure 1.

Naïve unadjusted comparisons of overall survival among patients treated with DARA versus POM + LoDex [12], [27], [28]. Data for POM + LoDex are from the MM‐003 and STRATUS studies.

Abbreviations: DARA, daratumumab; ITT, intent‐to‐treat; LoDex, low‐dose dexamethasone; POM, pomalidomide.

Republished from [28], with permission from the American Society of Hematology.

The HR (95% CI) from a naïve, unadjusted comparison of patients treated with DARA versus POM + LoDex was 0.61 (0.46–0.81; p = .0005) for MM‐003 and 0.55 (0.43–0.71; p < .0001) for STRATUS. Among DARA‐treated, POM‐naïve patients, the median OS was not reached at the time of analysis (Fig. 1). The HR (95% CI) from a naïve unadjusted comparison of POM‐naïve patients treated with DARA versus POM + LoDex was 0.38 (0.25–0.60; p < .0001) for MM‐003 and 0.34 (0.22–0.52; p < .0001) for STRATUS.

Matching Adjusted Indirect Comparison

As inclusion/exclusion criteria of the DARA and POM + LoDex trials were similar with respect to relevant patient characteristics, no DARA patients were excluded in the main analysis. In the sensitivity analysis of POM‐naïve patients only, DARA patients previously exposed to POM were excluded. Twelve patients had missing values for one or more baseline characteristics and were excluded from the MAIC when any of these baseline characteristics were used in the matching process. Myeloma subtype, race, and bone lesion presence were not available for STRATUS and could not be used in the matching process. Table 1 (DARA and MM‐003) and Table 2 (DARA and STRATUS) report the values of the baseline characteristics for all relevant trial arms before and after matching.

Table 1. Baseline demographics and characteristics of DARA‐treated and POM + LoDex–treated patients (MM‐003).

Abbreviations: —, not reported; ASCT, autologous stem cell transplantation; CrCl, creatinine clearance; DARA, daratumumab; ECOG, Eastern Cooperative Oncology Group; Ig, immunoglobulin; LoDex, low‐dose dexamethasone; N_eff, effective sample size; POM, pomalidomide.

Table 2. Baseline demographics and characteristics of DARA‐treated and POM + LoDex–treated patients (STRATUS).

Abbreviations: ASCT, autologous stem cell transplantation; CrCl, creatinine clearance; DARA, daratumumab; ECOG, Eastern Cooperative Oncology Group; LoDex, low‐dose dexamethasone; N_eff, effective sample size; POM, pomalidomide.

After matching, all available baseline characteristics were well balanced across trial populations (Tables 1, 2). The MAIC‐adjusted HR for OS was 0.56 (95% CI, 0.38–0.83; p = .0041) for MM‐003 and 0.51 (95% CI, 0.37–0.69; p < .0001) for STRATUS when all available baseline characteristics were matched, representing a reduction in the risk of death by 44% and 49% for DARA compared with POM + LoDex, respectively (Fig. 2A).

Figure 2.

Matching adjusted indirect comparison of OS among patients treated with DARA versus POM + LoDex in the ITT population (A) and in the POM‐naïve population (B). Data for POM + LoDex are from the MM‐003 and STRATUS studies.

Abbreviations: CI, confidence interval; DARA, daratumumab; HR, hazard ratio; ITT, intent‐to‐treat; LoDex, low‐dose dexamethasone; OS, overall survival; POM, pomalidomide.

For the sensitivity analysis using POM‐naïve patients only from the DARA trials, it was not possible to match all baseline characteristics given the small number of remaining patients (n = 66). In general, the smaller the IPD dataset, the more difficult it is to match to another dataset as there are fewer patients for up‐ or down‐weighting to mimic the comparator aggregated data. Therefore, baseline characteristics that were judged of lower prognostic importance (e.g., race, bone lesion presence, prior ASCT, age) were not used in the matching process.

Among POM‐naïve patients, the MAIC‐adjusted OS HR was 0.33 (95% CI, 0.17–0.66; p = .0017) for MM‐003 and 0.41 (95% CI, 0.21–0.79; p = .0082) for STRATUS, which represented a reduction in the risk of death by 67% and 59% for DARA compared with POM + LoDex, respectively (Fig. 2B).

The sensitivity of results regarding the choice of baseline characteristics included in the matching process, and the influence of matching on many variables on uncertainty (N_eff), was limited. The OS HRs between DARA and POM + LoDex were fairly consistent across a range of results for both MM‐003 and STRATUS studies, matching on a different set of baseline characteristics (Fig. 3A). The HR varied between 0.61 and 0.55 versus MM‐003 and between 0.58 and 0.51 versus STRATUS. The exclusion of POM‐exposed patients had the biggest impact on the results (HR reduced from 0.61 to 0.38; Fig. 3B), whereas adding additional baseline characteristics in the matching algorithm only had limited impact. In the sensitivity analysis excluding the POM‐exposed patients, the HR ranged from 0.47 to 0.33 for MM‐003 and from 0.41 to 0.34 for STRATUS.

Figure 3.

HRs for overall survival with different numbers of baseline characteristics matched in the intent‐to‐treat population (A) and the POM‐naïve population (B). Data for POM + LoDex are from the MM‐003 and STRATUS studies. The covariate list used to determine the HRs is cumulative from the top to the bottom of the figures.

Abbreviations: ASCT, autologous stem cell transplantation; CI, confidence interval; CrCl, creatinine clearance; DARA, daratumumab; ECOG, Eastern Cooperative Oncology Group; HR, hazard ratio; N_eff, effective sample size; POM, pomalidomide.

The MAIC‐adjusted PFS HR for DARA versus POM + LoDex was 0.72 (95% CI, 0.50–1.05; p = .0848) for MM‐003 and 0.86 (95% CI, 0.65–1.15; p = .3208) for STRATUS when all available baseline characteristics were matched. The results of the sensitivity analyses varied between 0.88 and 0.72 versus MM‐003 and between 0.97 and 0.86 versus STRATUS (Fig. 4A). The exclusion of POM‐exposed patients impacted HR versus MM‐003 (range 0.78–0.51) and STRATUS (range 0.87–0.63; Fig. 4B), whereas the addition of baseline characteristics in the matching algorithm only had limited impact.

Figure 4.

HRs for progression‐free survival with different numbers of baseline characteristics matched in the intent‐to‐treat population (A) and the POM‐naïve population (B). Data for POM + LoDex are from the MM‐003 and STRATUS studies. The covariate list used to determine the HRs are cumulative from the top to the bottom of the figures.

Abbreviations: ASCT, autologous stem cell transplantation; CI, confidence interval; CrCl, creatinine clearance; DARA, daratumumab; ECOG, Eastern Cooperative Oncology Group; HR, hazard ratio; N_eff, effective sample size; POM, pomalidomide.

OS Analyses by Depth of Response

An analysis of OS in the intent‐to‐treat (ITT) population, as measured in subgroups based on reduction in M‐protein levels (≥50%, ≥25%, and <25%), revealed that OS is consistently longer in DARA‐treated patients across each response category (Fig. 5A) [38]. Among POM‐naïve patients, the OS benefit of DARA versus POM + LoDex was more pronounced than when the ITT populations were compared. Within this subgroup, the OS associated with the lowest response category for patients treated with DARA (reduction of M‐protein levels <25%) was comparable with the OS associated with patients at a deeper response category of ≥50% reduction in M‐protein levels after treatment with POM + LoDex (Fig. 5B).

Figure 5.

Naïve comparison of overall survival by depth of response in both the total population (A) and the POM‐naïve subgroup (B; MM‐003 study). Data for POM + LoDex are from the MM‐003 study.

Abbreviations: DARA, daratumumab; ITT, intent‐to‐treat; LoDex, low‐dose dexamethasone; POM, pomalidomide.

Discussion

In the absence of head‐to‐head trials, indirect comparison studies can provide useful insights into the relative efficacy of various treatment options [29], [30]. When no comparative trials are available, the usual approach of indirect treatment comparisons based on the comparison of relative treatment effects across trials versus a common comparator is not feasible. In such cases, the MAIC is one of few acceptable alternatives to compare absolute treatment effects while minimizing the risk of bias due to differences in populations [29], [30]. This technique has recently been used to compare the efficacy of treatments for a wide range of disorders and endpoints [30], [31], [39], [40].

In the present study, MAIC results suggest that DARA improved OS compared with POM + LoDex in patients with heavily pretreated and refractory MM, with consistent comparative results versus the 2 POM trials. The primary analysis suggests that there was a 44% (MM‐003) and 49% (STRATUS) reduction in the risk of death (HR = 0.56 and 0.51, respectively) compared with POM + LoDex. As there was a high percentage of POM‐refractory patients (55%) treated with DARA in GEN501 and SIRIUS, and because the MM‐003 and STRATUS studies only included POM‐naïve patients, the OS advantage observed with DARA may be a conservative estimate. When including POM‐naïve patients only from the DARA studies in the comparison, results suggest a 67% and 59% reduction in the risk of death compared with POM + LoDex, respectively.

Interestingly, although there was a numerical PFS benefit with DARA compared with POM + LoDex, it was not statistically significant. DARA's mechanism of action may explain the differences observed in DARA‐derived PFS and OS benefit. For example, prolonged exposure to DARA may have provided the opportunity for DARA's immunomodulatory mechanism of action to exert its effect, increasing T‐cell clonality and inducing lysis of immune‐suppressive CD38⁺ myeloid‐derived suppressor cells, regulatory B cells, and regulatory T cells [11]. This would allow the opportunity for synergistic effects with subsequent therapies, such as IMiDs, that costimulate T cells. Among other hypotheses, Usmani et al. also propose that, because DARA is well tolerated, patients may have an opportunity to recover from the stress of prior therapies, which allows them to receive aggressive subsequent therapies [12]. Additionally, DARA was the only agent in development with significant clinical activity in advanced patients that would likely have had an impact after the POM studies were conducted. Further studies will be required to conclusively determine the reason for differences in DARA‐derived PFS and OS benefit.

Results of the unadjusted comparisons of OS by depth of response indicated a consistent trend across subgroups based on M‐protein‐level reduction (≥50%, ≥25%, and <25%), confirming that DARA demonstrates an OS advantage over POM + LoDex regardless of depth of response.

Indirect comparisons with available standard of care regimens have been used previously by Health Technology Assessment bodies to inform reimbursement decisions for myeloma therapies [41]. Specifically, the MAIC method was used to account for differences in patients’ baseline characteristics between available studies, and the result was a new set of postinduction response rates, stem cell transplant rates, and post‐transplant response rates for induction treatment of adults with previously untreated MM who are eligible for high‐dose chemotherapy with hematopoietic stem cell transplantation. The MAIC approach was also used in another study that assessed PFS and OS outcomes for different treatment regimens used in patients with relapsed and/or refractory MM who received ≥2 prior treatments [42]. Recently, the National Institute for Health and Care Excellence published a technical support document [32] on population‐adjusted indirect comparisons such as the MAIC method. While acknowledging their limitations, they recognize the value of such methods for Health Technology Assessment in the absence of a connected network of randomized evidence, or where single‐arm studies are involved.

Some limitations should be noted in considering the results of this MAIC study. This study was a retrospective analysis of prospective studies, and so caveats associated with these types of analyses (e.g., statistical limitations and confounding variables) should be taken into account when interpreting these results. Additionally, despite use of MAIC, some limitations remain due to certain differences between the study populations and how the data were collected from these trials. For example, the matching process could only take into account baseline characteristics that were commonly reported for the DARA studies on one hand and for MM‐003 or STRATUS on the other. Myeloma subtype, race, and bone lesion presence were not available for STRATUS and could therefore not be included in the matching process for that study. International Staging System (ISS) stage, beta‐2 microglobulin level, and cytogenetic risk were not captured in GEN501 and could therefore not be used in any match. Because ISS stage and cytogenetic risk could be prognostic factors, a sensitivity analysis was performed to assess the impact of these missing characteristics on the comparison with POM + LoDex. Multiple imputation was used to first impute the missing values for the 42 GEN501 patients (out of 148 patients in the pooled data), and the full dataset was used in the MAIC algorithm. However, the results of the sensitivity analysis were comparable to those of the main analysis, demonstrating little impact of including or excluding these 2 characteristics.

Although a wide range of prognostic factors were accounted for in the matching process, another limitation of this analysis is that residual confounding for unobserved patient characteristics cannot be excluded. Additionally, as the subgroup of POM‐naïve patients from the DARA‐treated cohort only included 66 patients, the sensitivity analysis of these patients led to more uncertainty. Nevertheless, the results of the MAIC indicated that POM‐naïve patients may receive additional benefit from DARA. Lastly, the OS analysis by depth of response was naïve and thus not based on a MAIC analysis.

Recently published data revealed that the combination of DARA with standard of care regimens (bortezomib/dexamethasone and lenalidomide/dexamethasone) significantly reduces the risk of disease progression and/or death in relapsed or refractory myeloma patients who received ≥1 prior treatment [13], [14], leading to the recent approval of this indication by the FDA [16]. Results of an ongoing network meta‐analysis of data from these recently published trials are expected to provide additional perspective on the efficacy of DARA in combination with standard of care regimens for the treatment of patients with relapsed or refractory MM.

Conclusion

The findings of this MAIC study showed that monotherapy with DARA improved clinical outcomes compared with POM + LoDex in patients with heavily pretreated and refractory MM. In the absence of head‐to‐head trials, indirect comparisons such as MAIC analyses can provide clinicians and reimbursement decision‐makers with useful insights on the relative efficacy of various treatments for patients with MM.

Author Contributions

Conception/design: Suzy Van Sanden, Tetsuro Ito, Joris Diels

Collection and/or assembly of data: Suzy Van Sanden, Tetsuro Ito, Joris Diels

Data analysis and interpretation: Suzy Van Sanden, Tetsuro Ito, Joris Diels, Martin Vogel, Andrew Belch, Albert Oriol

Manuscript writing: Suzy Van Sanden, Tetsuro Ito, Joris Diels, Martin Vogel, Andrew Belch, Albert Oriol

Final approval of manuscript: Suzy Van Sanden, Tetsuro Ito, Joris Diels, Martin Vogel, Andrew Belch, Albert Oriol

Disclosures

Suzy Van Sanden: Janssen (E), Johnson & Johnson (OI); Tetsuro Ito: Janssen (E), Johnson & Johnson (OI); Martin Vogel: Janssen (E), Johnson & Johnson (OI); Joris Diels: Janssen (E), Johnson & Johnson (OI); Albert Oriol: Celgene, Janssen, Amgen (C/A, SAB). Andrew Belch indicated no financial relationships.

(C/A) Consulting/advisory relationship; (RF) Research funding; (E) Employment; (ET) Expert testimony; (H) Honoraria received; (OI) Ownership interests; (IP) Intellectual property rights/inventor/patent holder; (SAB) Scientific advisory board