Prognostic Importance of Measurable Residual Disease (MRD) Kinetics and Progression-Free Survival (PFS) Benefit in MRD+ Patients (Pts) with Ixazomib Vs Placebo As Post-Induction Maintenance Therapy: Results from the Multicenter, Double-Blind, Phase 3 TOURMALINE-MM4 Trial in Non-Transplant Newly Diagnosed Multiple Myeloma (NDMM) Pts

Background

MRD status is one of the most relevant prognostic factors in transplant-eligible NDMM pts. However, it is perhaps in elderly NDMM pts – the most common pt subgroup, in whom an optimal balance between efficacy and toxicity is of utmost importance – that such sensitive response assessments could help to avoid under- or over-treatment. Although prospective data on MRD in this setting are limited, available findings from recent studies indicate the feasibility and benefit of achieving MRD negativity after induction. As a result, MRD status is increasingly included as an endpoint in clinical trials of transplant-ineligible NDMM pts. However, there are virtually no MRD data from the maintenance phase of treatment in elderly non-transplant NDMM pts.

Aim

Evaluate the impact of evolving MRD kinetics on PFS in non-transplant NDMM pts receiving the oral proteasome inhibitor (PI) ixazomib or placebo as post-induction maintenance therapy in the TOURMALINE-MM4 trial (NCT02312258).

Methods

Pts achieving a partial response (PR) or better as best response following 6–12 months of standard-of-care induction therapy were randomized 3:2 to ixazomib (n=425) or placebo (n=281) for up to 24 months (26 cycles). Stratification factors included prior PI exposure, pre-induction disease stage, age at randomization, and post-induction best response. The primary endpoint was PFS from randomization. Median follow-up was 21.1 months. Correlation of MRD status with outcomes was a prespecified secondary endpoint. Bone marrow aspirate samples were collected at screening (n=386), cycle 13 (n=137), and cycle 26/end of treatment (n=48) from eligible pts and used to evaluate MRD kinetics in those confirmed/suspected to have achieved complete response (CR) at each time point. MRD status was determined using 8-color flow cytometry (estimated sensitivity of 10^-5). At screening and post-screening, pts with <CR who were missing MRD data were imputed as MRD+, whereas ongoing pts in CR with missing MRD data were classified as ‘missing’. After imputation, 8% and 13% of pts were missing MRD data at screening and post-screening, respectively.

Results

At screening, MRD status was available in 650 pts: 70 were MRD– (ixazomib, n=44/383 [11%]; placebo, n=26/267 [10%]) and 580 MRD+ (ixazomib, n=339 [89%]; placebo, n=241 [90%]). With ixazomib vs placebo there was no difference in PFS from randomization among MRD– pts (24-month rate: 71.7% vs 65.8%), whereas among MRD+ pts a significant improvement was observed (24-month rate: 34.0% vs 16.7%; median 16.6 vs 8.7 months; hazard ratio [HR] 0.599, 95% confidence interval [CI] 0.486–0.739, p<0.001).

Overall, among 615 pts with known post-screening MRD status, failure to achieve or maintain MRD– status resulted in a >8-fold increased risk of progression and/or death (HR 8.77; 95% CI 4.13–18.9; p<0.001. In 57 pts with undetectable MRD post-screening (sustained MRD– or converted to MRD–), there was no significant difference in PFS with ixazomib vs placebo. Among 558 pts who were MRD+ post-screening (persistent MRD or converted to MRD+), PFS was significantly prolonged with ixazomib vs placebo (median 13.8 vs 8.5 months; HR 0.701 [95% CI 0.570–0.862], p=0.001).

Data on evolving MRD kinetics (screening and post-screening) were available for 577 pts; 30 were classified as having sustained undetectable MRD (MRD– to MRD–), 19 converted from MRD+ to MRD–, 23 converted from MRD– to MRD+, and 505 had persistent MRD (MRD+ to MRD+). Supporting the validity of MRD– status as a treatment endpoint during maintenance, PFS was longer in pts with sustained undetectable MRD and those converting from MRD+ to MRD– vs those converting from MRD– to MRD+ or with persistent MRD (Figure). Notwithstanding, pts converting from MRD– to MRD+ had significantly better PFS vs pts with persistent MRD (median 23.9 vs 10.2 months, HR 0.440, 95% CI 0.241–0.804, p=0.006).

Conclusions

This analysis of longitudinal MRD assessment during post-induction maintenance therapy in non-transplant NDMM pts demonstrates the prognostic importance of undetectable MRD in this setting, as well as the benefit of continuing treatment with ixazomib maintenance following best response to induction among MRD+ pts. The findings also highlight the need for additional treatment approaches in pts with persistent MRD, and the value of periodic MRD monitoring during maintenance to anticipate clinical relapse upon conversion from MRD– to MRD+ status.