In the final DREAMM-3 analysis, belantamab mafodotin produced responses that lasted a median of 25.3 months versus 9.9 with pomalidomide plus dexamethasone (Pd), yet median progression-free survival was 11.2 versus 7.0 months and median overall survival was 24.0 versus 22.9 months. Among 325 patients with relapsed/refractory multiple myeloma, the PFS difference was not statistically significant. No new safety signals emerged, but ocular toxicity remained frequent.
Belantamab Mafodotin After the Negative Primary Analysis
B-cell maturation antigen (BCMA) is highly expressed on malignant plasma cells and is targeted by antibody-drug conjugates, bispecific antibodies, and CAR T cells. Belantamab mafodotin is a BCMA-directed antibody-drug conjugate that delivers monomethyl auristatin F. Its activity also involves immune-mediated mechanisms, including antibody-dependent cellular cytotoxicity. It has been associated with relatively limited immunosuppression, and no T-cell exhaustion has been demonstrated. Corneal toxicity is the main limitation, often requiring treatment interruptions or dose modifications.
The drug received initial conditional authorization as monotherapy in heavily pretreated disease. DREAMM-3 was the confirmatory trial. It missed its primary endpoint of superior PFS versus Pd, and the original authorization was withdrawn. Phase 3 trials of belantamab mafodotin with bortezomib-dexamethasone or Pd have since shown greater efficacy and led to approvals in several countries.
Study Design
DREAMM-3 was a multicenter, open-label, randomized phase 3 trial. Eligible patients had received at least two prior lines of therapy, including lenalidomide and a proteasome inhibitor. A total of 325 patients were randomized 2:1: 218 to belantamab mafodotin and 107 to Pd. Median age was 68 years, and about 22% were 75 years or older. Baseline characteristics were balanced.
The primary endpoint was investigator-assessed PFS. Secondary endpoints included OS, response rate, DOR, MRD negativity, and safety. The data cutoff was April 22, 2025, with median follow-up of about 22 months in both arms. Because the primary superiority endpoint had already failed, the final analysis was descriptive, with no further formal hypothesis testing.
Efficacy: PFS, OS, and Response Durability
The PFS hazard ratio moved toward belantamab mafodotin across analyses:
- Primary analysis: HR 1.03 (P = 0.56)
- Post-primary analysis: HR 0.90
- Final analysis: HR 0.87
The confidence interval at the final analysis included 1.0, so superiority was not established, and the trend does not overturn the negative primary result.
Median OS was 24.0 versus 22.9 months (HR 0.98), with no survival disadvantage and no survival benefit. Median PFS on the next line of therapy or death (PFS2) was 17.3 versus 13.0 months. This endpoint was exploratory. Median DOR was 25.3 months with belantamab mafodotin versus 9.9 months with Pd. Responses were also deeper:
- Stringent complete response: 9% versus 3%
- Complete response or better with MRD negativity: 8% versus fewer than 1%
DOR is calculated only among responders, so it doesn’t account for early progression or nonresponse. This explains how a large DOR difference coexists with a nonsignificant PFS difference.
Why PFS Improved With Longer Follow-Up
The PFS curves crossed early at the primary analysis, so the treatment effect was not constant over time. Patients who progressed within the first two months on belantamab mafodotin had a greater disease burden than early progressors on Pd. They had worse performance status, more prior treatment, more prior anti-CD38 exposure, and higher soluble BCMA and M-protein levels. The investigators suggest this imbalance contributed to the unfavorable primary result.
They also propose that immune-mediated activity produces delayed but durable responses, consistent with the longer time to best response and long DOR. DREAMM-3 wasn’t designed to test this mechanism. The investigators further suggest that a noninferiority design might have suited a single-agent versus two-drug comparison, though the trial tested superiority and didn’t meet it.
Safety
No new adverse-event patterns emerged with prolonged exposure. Grade 3-4 adverse events were similar between belantamab mafodotin and Pd (71% versus 66%), and neutropenia was less frequent (12% versus 40%).
Ocular toxicity was the main concern: blurred vision occurred in 48% versus 7%, and ophthalmic examination abnormalities in 79% of evaluable patients on belantamab mafodotin, most grade 2 or higher. Events were managed with dose delays and reductions.
Among patients with a first grade 2 or higher finding, 80% improved to grade 1 or better (median, 68 days), but 24 had an unresolved finding at last assessment. Earlier analyses showed responses could be maintained or deepened despite interruptions or dose reductions.
Exposure-adjusted rates of grade 3 or higher infection were lower with belantamab mafodotin (15.6 versus 38.8 events per 100 patient-years), consistent with limited immunosuppression. This was a descriptive comparison from an open-label study.
Limitations and Clinical Implications
The trial was open-label with investigator-assessed PFS. The final analysis was descriptive, and the PFS confidence interval remained compatible with both benefit and no benefit. The immune-mediated explanation is not established by the data, and DOR and response depth can’t substitute for the primary randomized comparison.
Monotherapy didn’t show superior PFS over Pd. Durable responses in a subset of patients didn’t produce a population-level benefit. Ocular monitoring and dose management remain necessary in the combination regimens now in use, which have shown greater efficacy than monotherapy.
Source: Dimopoulos MA, Hungria V, Meng H, et al. Final analysis results from DREAMM-3, a phase 3 study of belantamab mafodotin versus pomalidomide plus dexamethasone in patients with relapsed/refractory multiple myeloma. Blood Cancer Journal. 2026;16:184.
DREAMM-9 Final Analysis: Optimizing Belamaf Dosing and Scheduling Within BVRd for TI NDMM
