Paper Details

ABSTRACT

Background: Sustained measurable residual disease (sMRD) negativity is a strong surrogate for survival in multiple myeloma, and prior analyses have linked sMRD duration with durable remission and potential cure. Curative modalities require optimally timed treatment discontinuation, and the ENDURANCE trial supports fixed-duration maintenance in selected patients without impacting survival. However, there are limited tools to prospectively guide individualized treatment cessation while preserving depth of response. Because durable remission may reflect both clonal eradication and enhanced immune surveillance, integrating clinical features with immunophenotyping of the tumor microenvironment may improve stratification of relapse risk and cure.

Methods: We reviewed patients at our center who achieved sustained MRD-negative, imaging-negative remission lasting ≥24 months (sMRD24), with MRD status defined by next-generation flow (NGF) or NGS. Relative survival was estimated with age- and sex-matched US population data. Relapse risk was estimated with death as a competing event using Gray test and Fine-Gray hazard models. We performed paired bone marrow scRNA-seq and scTCR-seq, and we used unsupervised compositional analysis by Bayesian beta-binomial modeling as well as pseudobulk differential expression to characterize emerging cell populations. T cell receptor (TCR) diversity was assessed by Shannon diversity index, with clonality quantified by the combined proportion of the 10 largest clonotypes.

Results: We identified 430 patients who achieved sMRD24 with median follow-up of 78 months, including 20% IMS/IMWG genomic high-risk and 24% ISS stage 3. MRD negativity was documented by NGF alone in 35% and confirmed by NGS in 65%. 5-year overall survival landmarked from sMRD24 was 92%, with 5-year relative survival of 1.03 (95% CI 0.99-1.08), suggesting survival similar to the general population. Risk factors for relapse after sMRD24 included genomic high-risk (aHR 3.44, p<0.001), ISS stage 3 (aHR 2.68, p=0.002), and MRD by NGF alone without NGS confirmation (aHR 8.02, p<0.001). Patients without these risk factors (n=172, 40%) formed an ultra-low-risk subgroup with an estimated 8-year relapse risk of 6%, compared to 21% and 55% with 1 and ≥2 risk factors.

A total of 303/430 patients discontinued therapy at varying timepoints, including 57 who remain MRD negative ≥5 years off treatment. The estimated 5-year relapse risk on surveillance was 16% overall, but just 2.6% in ultra-low-risk patients. For those who developed MRD resurgence after stopping treatment (n=71), 38% spontaneously converted back to MRD negative and 24% relapsed at median 39-month follow-up. Finally, among patients on surveillance who reinitiated therapy for progression (n=30), 29 remain in remission on subsequent therapy at median 17-month follow-up, and 1 high-risk patient died of disease.

We next profiled paired bone marrow samples from a subset of sMRD24 patients (n=11) both at early MRD negativity on maintenance (T1) and at a subsequent timepoint a median of 39 months later (T2), either with ongoing sMRD (n=5) or at MRD resurgence preceding relapse (n=6). At T1, patients with future relapse demonstrated expansion of CD8 T cells (FDR<0.001) with enrichment of activation and inflammatory markers such as HLA-DR (FDR=0.02) and IFNγ (FDR=0.04), as well as loss of antigen presenting cells such as cathepsin S-high CD14 monocytes (FDR=0.03); this pattern further deepened at T2. Differential gene expression across CD8 subsets showed upregulation of exhaustion and TNF/NF-κB programs with loss of IFN-response and proliferative signatures. Finally, relapsed patients had significantly diminished TCR diversity (p=0.045) and increased TCR clonality by top-10 clone fraction (p=0.03) compared to those with sustained remission.

Conclusions: Patients with sMRD24 had comparable survival to the general population and included an ultra-low-risk subgroup (genomic standard risk, ISS 1-2, MRD negative by NGS) with rare progression after stopping therapy and durable treatment sensitivity among the minority who relapsed. The subset of patients with sustained response who underwent longitudinal immunophenotyping demonstrated a resting immune profile characterized by reduced activated CD8 effector populations and preserved TCR diversity. Ultra-low-risk features and favorable T cell profiles after sMRD24 may guide future prospective strategies for early treatment de-escalation aimed at cure.