Comparative Efficacy of Belantamab Mafodotin in Combination With Bortezomib and Dexamethasone Versus Standards of Care in Patients With Third‐Line or Later Relapsed/Refractory Multiple Myeloma

To the Editor,

1.

Multiple myeloma (MM) typically requires multiple lines of therapy (LOT) over time due to relapse [1]. Treatments for MM include proteasome inhibitors (PIs), immunomodulatory drugs (IMDs), anti‐CD38 monoclonal antibodies (mAbs), and B‐cell maturation antigen (BCMA) targeting therapies including bispecific antibodies (BsAbs), chimeric antigen receptor T‐cell (CAR‐T) therapies, and an antibody–drug conjugate (ADC). These therapies are frequently used in combination regimens [1, 2].

BCMA‐targeting therapies have transformed the treatment landscape of MM by producing unprecedented survival outcomes versus standards of care; however, clinical challenges remain, including accessibility to sites of care, patient eligibility, manufacturing delays, and repeat dosing [3, 4, 5, 6]. Belantamab mafodotin is an ADC targeting BCMA with a unique mechanism of action, consisting of payload‐ and immune‐mediated effects that may contribute to an adaptive immune response [7].

In the United States (US), belantamab mafodotin with bortezomib plus dexamethasone (BVd) is approved for adults with RRMM who received ≥ 2 prior LOT, including a PI and an IMD [8]. The US Food and Drug Administration (FDA) approval for this population was supported by results from the DREAMM‐7 study (NCT04246047), which demonstrated superior progression‐free survival (PFS) and overall survival (OS) with BVd versus daratumumab plus bortezomib and dexamethasone (DVd) in patients with ≥ 1 prior LOT [5, 9].

To further understand BVd's place within the treatment landscape, this study aimed to compare the relative efficacy of BVd versus four comparator regimens. To align with the FDA‐approved indication for BVd [8], indirect treatment comparison (ITC) analyses of PFS and OS were conducted using the subgroup of eligible patients who received ≥ 2 prior LOT including a PI and an IMD in the DREAMM‐7 (BVd) [5, 9], KarMMa‐3 (idecabtagene vicleucel [Ide‐cel]; NCT03651128), ICARIA‐MM (isatuximab and pomalidomide plus dexamethasone [IsaPd]; NCT02990338), DREAMM‐8 (pomalidomide and bortezomib plus dexamethasone [PVd]; NCT04484623), and ELOQUENT‐3 (elotuzumab and pomalidomide plus dexamethasone [EPd]; NCT02654132) trials (Table 1 details the selected populations from the studies included in the ITC and Table S1 details approved indications for the regimens). Full individual patient data (IPD) were available for DREAMM‐7 and DREAMM‐8 studies, with reconstructed IPD and aggregated summary baseline data for KarMMa‐3, ICARIA‐MM, and ELOQUENT‐3. A feasibility assessment was conducted to evaluate the suitability of ITC between BVd and relevant comparators by comparing the studies across study design characteristics, eligibility criteria, patient baseline characteristics, and availability and definitions of outcomes. Population adjustment is described in the Supporting Information Methods; inverse probability of treatment weighting was used for population adjustment for comparison with PVd and unanchored matching adjusted indirect comparison (MAIC) was used for all other comparisons.

TABLE 1.

Patient populations from comparator studies included for comparison with BVd ^a .

Comparator study	Comparator treatment	Comparator population for ITC	BVd population for ITC
KarMMa‐3	Ide‐cel (CAR T‐cell therapy)	All patients who underwent randomization/ITT (N = 254)	All randomized 3 L–5 L patients who have previously received a PI and an IMD (N = 96)
ICARIA‐MM	IsaPd (anti‐CD38 mAb + IMD + steroid)	All patients who underwent randomization/ITT (N = 154)	All randomized 3 L+ patients who have previously received a PI and an IMD (and lenalidomide specifically), and have a known cytogenetic risk at baseline (N = 85)
DREAMM‐8	PVd (IMD + PI + steroid)	All randomized 3 L+ patients who have previously received a PI and an IMD (and lenalidomide specifically) (N = 69)	All randomized 3 L+ patients who have previously received a PI and an IMD (and lenalidomide specifically) (N = 86)
ELOQUENT‐3	EPd (anti SLAMF7 mAb + IMD + steroid)	All patients who underwent randomization/ITT (N = 60)	All randomized 3 L+ patients who have previously received a PI (and bortezomib specifically) and an IMD, and have a known cytogenetic risk at baseline (N = 103)

Abbreviations: 3 L, third‐line therapy; 3 L+, third‐line therapy or more; 5 L, fifth‐line therapy; EPd, elotuzumab with pomalidomide plus dexamethasone; IA, interim analysis; Ide‐cel, idecabtagene vicleucel; IMD, immunomodulatory drug; IsaPd, isatuximab with pomalidomide and dexamethasone; ITC, indirect treatment comparison; ITT, intention to treat; mAb, monoclonal antibody; OS, overall survival; PFS, progression‐free survival; PI, proteasome inhibitor; PVd, pomalidomide and bortezomib plus dexamethasone.

^a From DREAMM‐7, IA1 was used for the comparison of PFS and IA2 was used for the comparison of OS.

A weighted Cox proportional hazards model (Cox regression) was used to compare clinical outcomes (PFS and OS) for BVd versus each respective comparator treatment. PFS is reported as an independent review committee–assessed outcome for comparisons of BVd versus PVd and Ide‐cel, and as an investigator‐assessed outcome for comparisons of BVd versus IsaPd and EPd, based on the data available for the respective comparator studies at mature data cuts.

DREAMM‐7 and DREAMM‐8 were conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines. All patients participating in the trials provided written informed consent before enrollment. The trial protocol and amendments were approved by the appropriate ethics body at each participating institution. Similarly, the KarMMa‐3, ICARIA‐MM, and ELOQUENT‐3 trial protocols were approved by local or independent IRBs or ethics committees at participating sites, and all patients provided written informed consent.

In the base case models, matching achieved good covariate balance across all matched covariates (Figure S1 and Tables S2–S5). Effective sample sizes (ESS) in the base case models were 35 for PVd and 50, 52, and 42 for BVd versus Ide‐cel, IsaPd, and EPd respectively. The sensitivity analysis that adjusted for all 10 covariates resulted in an ESS of 12 for the comparison of BVd versus PVd and did not achieve adequate balance. Consequently, sensitivity analysis weights were not used for BVd versus PVd comparisons.

In the comparison of BVd versus PVd, PFS was significantly longer for BVd (HR 0.517, 95% CI: 0.275–0.972; p = 0.041), while OS outcomes numerically favored BVd (HR 0.604, 95% CI: 0.331–1.103; p = 0.101; Figure 1). For comparisons of BVd with Ide‐cel, IsaPd, and EPd, HR values for both PFS and OS significantly favored BVd (Figures 1 and 2). HRs (95% CI; p‐value) reported for PFS for comparisons with BVd were: Ide‐cel, 0.468 (0.287–0.762; p = 0.002), IsaPd, 0.538 (0.346–0.836; p = 0.006), and EPd, 0.337 (0.173–0.654; p = 0.001), and HRs (95% CI; p‐value) reported for OS for comparisons with BVd were: Ide‐cel, 0.509 (0.274–0.945; p = 0.033), IsaPd, 0.348 (0.195–0.623; p < 0.001), and EPd, 0.405 (0.209–0.787; p = 0.008).

FIGURE 1.

(A) PFS and (B) OS for BVd versus PVd and (C) PFS and (D) OS for BVd versus Ide‐cel. PFS is reported as independent review committee assessed for PVd and Ide‐cel comparisons. BVd, belantamab mafodotin with bortezomib plus dexamethasone; CI, confidence interval; ESS, effective sample size; HR, hazard ratio; Ide‐cel, idecabtagene vicleucel; IPTW, inverse probability of treatment weighting; OS, overall survival; PFS, progression‐free survival; PH, proportional hazards; PVd, pomalidomide and bortezomib plus dexamethasone.

FIGURE 2.

(A) PFS and (B) OS for BVd versus IsaPd and (C) PFS and (D) OS for BVd versus EPd. PFS is reported as investigator assessed for IsaPd and EPd comparisons. BVd, belantamab mafodotin with bortezomib plus dexamethasone; CI, confidence interval; EPd, elotuzumab with pomalidomide plus dexamethasone; ESS, effective sample size; HR, hazard ratio; IsaPd, isatuximab with pomalidomide and dexamethasone; OS, overall survival; PFS, progression‐free survival; PH, proportional hazards.

Treatment effects remained consistent, and in most cases, statistical significance was preserved after inclusion of additional matching factors in sensitivity analyses (Figures S2–S4).

The phase 3 DREAMM‐7 study demonstrated significant improvements in PFS and OS with BVd compared with DVd in patients with RRMM after ≥ 1 prior LOT, with PFS improvement maintained in subgroups of patients with ≥ 2 prior LOT [5, 9]. In a prior ITC, BVd improved PFS and OS versus comparators, including daratumumab plus carfilzomib and dexamethasone, isatuximab plus carfilzomib and dexamethasone, and DVd, in patients with RRMM who received ≥ 1 prior LOT [10]. The current ITC further supports the findings of DREAMM‐7 and builds upon the prior ITC, showing that for patients who received ≥ 2 prior LOT, BVd was associated with improved PFS and OS versus all standard of care regimens evaluated, including a numerical OS improvement for the comparison with PVd.

In RRMM, there remains an unmet need for novel, effective therapies that are accessible at all sites of care and to a broader patient population [2]. This population‐adjusted ITC study suggests a consistent, clinically meaningful PFS and OS benefit for BVd versus standard of care triplets in patients with RRMM receiving third‐line or later therapy, including those previously exposed or refractory to lenalidomide; this reinforces that the efficacy seen with BVd in prior analyses [5, 10] is maintained in patients in later therapy lines, where therapeutic options are limited and prognoses are typically poor [1, 2]. BVd is generally well‐tolerated, with ocular events associated with belantamab mafodotin being common and transient with high rates of resolution [9]. Infections, which are commonly associated with BCMA‐targeting therapies [11, 12], are generally manageable with BVd with high resolution rates, low rates of fatal events, low treatment discontinuation rates, and no requirement for prophylaxis or hospitalization for monitoring [5, 8]. Compared to other currently‐available BCMA‐targeting therapies, BVd's outpatient administration without a requirement for specialized treatment centers may improve accessibility for a broader patient population [13].

Population‐balancing methods such as propensity score matching require access to IPD from all studies to adjust for differences in the distribution of effect‐modifying covariates. The lack of IPD for some comparators in this ITC may therefore have constrained the analyses. The MAIC method was used to estimate treatment effects when access to patient‐level data was limited. However, bias from residual confounding may remain, particularly if studies without IPD differ systematically from those with IPD. Therefore, as with all ITCs, residual confounding cannot be fully ruled out [14]. The MAIC was also limited to use of comparator trial patient populations for comparisons rather than the BVd target population, which led to small sample sizes for some comparisons after covariate adjustment.

Additionally, heterogeneity in trial design, follow‐up duration, and efficacy response assessment methods among the comparator studies may have influenced effect estimates. Furthermore, many clinical trials also did not specify or control for post‐protocol treatment sequencing, dose modifications, or subsequent LOT, which can potentially bias overall survival analyses and interpretation [15].

Future research is needed, including head‐to‐head randomized trials to confirm efficacy and safety findings from indirect comparisons, real‐world comparative studies to assess effectiveness, safety, and clinical practice patterns in broader patient populations, and analysis of treatment sequencing strategies. Lastly, there is a need for treatments that are accessible in broader patient settings and have less treatment burden in terms of administration, travel, and monitoring requirements [6, 16]; further research on quality of life and real‐world use of belantamab mafodotin will provide valuable insights.

Although beyond the scope of this ITC, evaluating differences in infection rates associated with ADCs compared with bispecific antibodies and CAR T‐cell therapies could also provide important insights. CAR T‐cell and BsAb mechanisms of action, T‐cell exhaustion, prior therapies, use of lymphodepleting chemotherapy, and prolonged cytopenias can increase infection risk [17, 18]. Differences in safety outcomes, particularly infection rates, could meaningfully influence treatment selection in routine clinical practice when considering the balance between efficacy, tolerability, and healthcare resource utilization burden of managing AEs.

Finally, analyses of health‐related quality of life and economic outcomes in routine clinical practice would provide further insights into the utility of BVd relative to comparators. Together, these areas of focus would strengthen the evidence base to help guide both clinical and policy decision‐making for the treatment of RRMM.

Patients with RRMM who have received ≥ 2 prior LOT, including a PI and an IMD, have an unmet need for new, effective, and accessible treatments. This ITC indicates that BVd has the potential to deliver clinically meaningful improvements in PFS and OS compared with current standards of care in this patient population.

Author Contributions

Joshua Richter, Yvonne Efebera, and Habte Yimer contributed to data interpretation. Pragya Shukla, Venediktos Kapetanakis, Molly Purser, and Nick Ballew contributed to study concept or design, data acquisition, data analysis, and data interpretation. Natalie Boytsov contributed to study concept or design and data interpretation.

Funding

This work was supported by GSK (Study 307239).

Ethics Statement

DREAMM‐7 and DREAMM‐8 were conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines. The trial protocol and amendments were approved by the appropriate ethics body at each participating institution. Similarly, the KarMMa‐3, ICARIA‐MM, and ELOQUENT‐3 trial protocols were approved by local or independent IRBs or ethics committees at participating sites.

Consent

All patients participating in the trials provided written informed consent before enrollment.

Conflicts of Interest

Joshua Richter reports consultancy fees from Johnson & Johnson – Janssen, Bristol Myers Squibb, Pfizer, Karyopharm, Sanofi, Takeda, Genentech, AbbVie, and Regeneron and speaker bureau fees from Johnson & Johnson – Janssen, Bristol Myers Squibb, Sanofi, and Adaptive Biotechnologies. Yvonne Efebera reports honoraria from Janssen, Takeda, GSK, Oncopeptides, Sanofi, and Pfizer; advisory committee with Oncopeptides, Sanofi, GSK, Janssen, Pfizer, and Takeda; research funding from BMS; and independent adjudication committee with Takeda and Orca. Habte Yimer reports being on the speaker bureau for AbbVie, Genmab, Pfizer, Amgen, Janssen, Beigene, G1 Therapeutics, and Karyopharm. Pragya Shukla and Venediktos Kapetanakis are employees of Evidera, a business unit of PPD, part of Thermo Fisher Scientific, a consultancy that received funding from GSK for the work reported in this article. Molly Purser, Natalie Boytsov, and Nick Ballew are employees of, and hold financial equities in, GSK.

Supporting information

Table S1: Overview of comparator studies and approvals.

Table S2: Baseline characteristics before and after adjustment (BVd vs. PVd).

Table S3: Baseline characteristics before and after adjustment (BVd vs. Ide‐cel).

Table S4: Baseline characteristics before and after adjustment (BVd vs. IsaPd).

Table S5: Baseline characteristics before and after adjustment (BVd vs. EPd).

Figure S1: Covariate balance before and after matching adjustments (base case).

Figure S2: (A) PFS and (B) OS for BVd versus Ide‐cel (sensitivity analysis).

Figure S3: (A) PFS and (B) OS for BVd versus IsaPd (sensitivity analysis).

Figure S4: (A) PFS and (B) OS for BVd versus EPd (sensitivity analysis).

Data Availability Statement

Data used in the analysis were obtained from resources available in the public domain and previously published reports. Additionally, for their studies, GSK makes available anonymized individual participant data and associated documents from interventional clinical studies that evaluate medicines, upon approval of proposals submitted to: https://www.gsk‐studyregister.com/en.