Glaxo’s myeloma Dreamm turns into a nightmare

Belantamab’s registrational Dreamm-2 study comes up short of an earlier trial and, more importantly, now seriously lags competitors’ data.

Glaxosmithkline failed to secure a late-breaker for belantamab mafodotin at this month’s Ash conference. Investors today found out the likely reason why: the data simply were not good enough.

Not only did the results of the Dreamm-2 trial underwhelm versus the earlier exploratory Dreamm-1 study, they fell well short of rival multiple myeloma agents with a similar mechanism of action – inhibiting the BCMA antigen. This approach is already hugely crowded, yet undeterred Glaxo today vowed to complete a regulatory filing for belantamab before 2019 was out.

Belantamab, an antibody-drug conjugate, had earlier yielded a 60% overall remission rate in Dreamm-1, a study in subjects who had failed at least three prior lines of therapy. Ash saw a multitude of presentations for rival BCMA players, some yielding ORRs of 90-100% (Ash 2019 – for Car-T the bispecific antibody threat is real, December 9, 2019).

From bad to worse

Today, however, things got even worse for Glaxo. The Dreamm-2 study, on which belantamab is to be filed, yielded an ORR of only 31%, including a 3% rate of complete remissions, the group said.

As if that were not bad enough, serious instances of keratopathy, a toxicity involving changes in the eye’s corneal epithelium, were seen in 27% of Dreamm-2 subjects. Johnson & Johnson’s Cartitude-1 trial of JNJ-4528, for instance, did not detail keratopathy as a side effect.

So what has gone so badly wrong for Glaxo? An immediate finger of blame can be pointed at the fact that in Dreamm-2 subjects had to have failed an anti-CD38 MAb like Darzalex; this represents a particularly hard-to-treat group, and today’s data suggest that in these patients belantamab simply does not work very well.

Such a view is confirmed by a more thorough reading of the earlier Dreamm-1 data, which included Darzalex-progressed and -naive subjects alike. This reveals an important dichotomy: those who had not previously been treated with the CD38 MAb had a 71% ORR to belantamab, but those who had yielded a result of just 39%.

Selected multiple myeloma trials of belantamab mafodotin (GSK2857916)
    Status as regards...  
Study Line Transplant IMID PI CD38 MAb Setting
Dreamm-1 ≥4th Post/ineligible Failed Failed NS Monotherapy
Dreamm-2 ≥4th Post/ineligible Failed Failed Failed Monotherapy
Dreamm-3* ≥3rd Post/ineligible Failed Revlimid Failed NS Vs Pomalyst
Dreamm-4 ≥4th Post/ineligible Failed Failed Failed Keytruda combo
Dreamm-5 ≥4th NS Failed Failed Failed Monother & combos w GSK3174998  or GSK3359609
Dreamm-6 ≥2nd Post/ineligible NS NS NS Revlimid or Velcade combo
Note: *Dreamm-3 is a phase III study, with PFS as primary and OS as key secondary endpoint; NS=not specified; IMID=Revlimid or Pomalyst; PI=proteasome inhibitor, ie, Velcade, Ninlaro or Kyprolis; GSK3174998=anti-Ox40 MAb; GSK3359609=anti-Icos MAb.

To illustrate what Glaxo is now up against, J&J’s 100% ORR result in Cartitude-1 was achieved despite the fact that 88% of this study’s enrolees were triple-refractory, meaning that they had failed a CD38 MAb.

JNJ-4528, a Car-T therapy, now looks like the most efficacious anti-BCMA asset in the industry pipeline, but outside cell therapy a more direct threat to Glaxo is Bristol-Myers Squibb’s anti-BCMA bispecific MAb CC-93269. This yielded an 89% ORR in a study where 90% of subjects were Darzalex-refractory, making it the big surprise of Ash.

Immediate competition for Glaxo will come from elsewhere: Bluebird/Bristol intend imminently to file their own anti-BCMA Car, ide-cel, on the strength of the Karmma study, 84% of whose subjects were CD38-refractory.

EvaluatePharma reveals 11 BCMA-targeting industry projects at phase II or above, while a recent analysis of the BCMA space by Bernstein had identified 25 Car-T projects, six bispecifics, two antibody-drug conjugates and two other assets. To say Glaxo has its work cut out would be putting it mildly.

Share This Article