Mye-llennial Dilemma: What Are the Optimal Endpoints for Multiple Myeloma Clinical Trials?

About two years ago, our institution’s multiple myeloma (MM) group was discussing evidence-based maintenance regimens for high-risk MM. While the doublet regimen with a proteasome inhibitor and immunomodulator is appealing to us, the data are primarily from nonrandomized, single-center experiences, and results from randomized trials are not yet available to support our claim. As we debated, the following became very clear: By the time we get results from a randomized trial, those results are already clinically irrelevant.¹ 

How do we better design our trials to answer the key clinical questions, and how do we pick our clinical trial endpoints?

Challenging the Standard of Care

We have witnessed a rapid transition of several MM drugs from bench to bedside within the past decade. The pace of development is often faster than the pace of our clinical trial development, deployment, and analysis. The SWOG-0777 trial is an example of the challenging standard of care dilemma — a well-intentioned trial that began in 2007, designed to compare a three-drug induction to a two-drug induction. Unfortunately, the final results were published years after the three-drug regimens had already been adapted as standard of care by the National Comprehensive Cancer Network guidelines and after three-drug regimens had crept into common clinical usage. The same is true for phase III trials in the relapsed MM setting. This vicious circle will continue to hound investigators unless we break away from traditional study designs, especially with the new chimeric T- or natural killer–cell therapies, antibody-drug conjugates, and bispecific antibodies.

Picking Surrogate Endpoints for Future Trials: Pros

A wealth of MM literature demonstrates that depth of response (i.e., ≥ complete response [CR]) correlates with improved median progression-free survival or overall survival (PFS/OS).¹  The postulated improvement of PFS/OS led clinicians to quote regimens by their depth of response.²  However, response kinetics are also important. Patients who achieved rapid response (< 3 months) and lost it early (< 24 months) have the worst outcomes.³ 

The International Myeloma Working Group (IMWG) recognizes CR plus negative minimal residual disease (MRD; by flow or next-generation sequencing at 10^–5) as the deepest measurable response. Clinical trials already linked the achievement of MRD negativity with improved survival in newly diagnosed MM.⁴  The group also defines sustained MRD as two MRD assessments deemed negative a year apart. This is an important concept to negate the risk of succumbing to an off-color surrogate endpoint. Sustained MRD negativity can be the right primary endpoint for treatment de-escalation questions in clinical trials for newly diagnosed MM, as well as smoldering MM.

Picking Surrogate Endpoints for Future Trials: Cons

Surrogate endpoints could be met and yet miss late-effects that may influence outcomes, such as second primary malignancies during lenalidomide maintenance, toxicities from carfilzomib, or long-term steroid use in older populations.^5-7 

For this reason, endpoints should still include an OS. Perhaps the best example was witnessed in the BELLINI Trial.⁸  Researchers randomized patients in a 2:1 fashion to receive bortezomib and dexamethasone plus venetoclax or placebo. The trial met its PFS clinical endpoint. The PFS was 22.4 months versus 11.5 months, favoring the venetoclax combination (p=0.01). However, sadly, there were 40 deaths in the venetoclax arm compared to 11 deaths in the placebo group. This worsening OS was mostly due to treatment-emergent infection and increased mortality after disease progression, suggesting the provocation of more aggressive biology upon relapse.

We should note that the BELLINI trial used a targeted agent for all comers. For patients with t(11;14) or overexpression of BCL2, overall response rate was impressively around 90 percent, and MRD negativity (10^–5) was between 17 and 25 percent. Perhaps, if the trial was designed to enroll only targetable patients, we would not encounter such a surprising discrepancy.⁹ 

The limitations of MRD as a surrogate endpoint include the following:

• MRD testing might not be achievable or meaningful for all patients. For example, frail patients with stable disease could achieve functional cure without MRD negativity.

• MRD negativity does not provide information about toxicity or death.

• The MRD timeline is not well defined.

• The depth of MRD testing (10^–5 vs.10^–6) is not well defined.

Finally, all those clinical endpoints, whether they focus on survival or on quality of life, still miss the important concept of the “financial cost of survival.”¹⁰ 

Conclusion

Investigators should not capitulate to the current ever-changing landscape when instead, we could subdue the dilemma of rapid synchronous drug development. We suggest using sustained MRD negativity as an endpoint for regimen development and approval. Simultaneously, we should continue to validate MRD as a tool, and only apply results from targeted therapy trials to targetable patients.